ProductTex: friction study of this website · simulated participants
193 of 240 found something they'd use.
72–98% of them still decided the product was not for them. This report is about the distance between those two numbers: what people wanted, where the route to it broke, and what to change first.
Summary
We ran 360 simulated visits to ProductTex: Product people at software companies and Developers using AI coding agents, with unrelated visitors as a control. Nobody was given a task. Each participant arrived, looked around for up to 30 actions, and left when they chose to. We ran the study twice on different people and report only what appeared every time.
The appetite is there. 193 of 240 named something they would come back for, and 10 typed into the product rather than only clicking around it. 204 decided it was not for them, a judgement that moves between runs and which we report as a range in the results.
Three things cost the most, ranked by how many visits they ended rather than how often they were mentioned. In participants' own words: “Still tiny text: 'ProductTex: quick read · s', 'Participants', '60 simulated visits'” (on /sample/report.html; 50 raised it, 14 left there); “Small text and fiddly controls” (on /sample/report.html; 27 raised it, 6 left there); “Tiny 'Sample report' label” (on /sample; 28 raised it, 2 left there).
If you change one thing, change this: On /sample/report.html: raise the contrast and size of ‘60 simulated visits’. The audit records the measured ratio; 4.5:1 and 14px are the floor. We would then re-run the same 240 people on the changed page and expect the exits on that screen to fall; that re-run is included.
The groups did not decide the same way. Product people at software companies: 4 of 120 decided it was for them, 5 typed into it, median 6 actions. Developers using AI coding agents: 2 of 120, 5 typed, median 6 actions. That is a 2-point gap on the verdict between people who wanted the same thing (95 and 98 of them named something they would use). Whatever separates the two groups is doing more to the outcome than anything on the page they share.
The control group behaved differently, which is what makes the rest interpretable: 3 of 120 people with no reason to be here typed anything, against 10 of 240. Where a finding was raised by both groups at similar rates we say so; it is then a defect for anyone, not a problem with your audience.
What we set out to learn
- Whether Product people at software companies and Developers using AI coding agents, arriving with no instructions, find anything in the product worth coming back for.
- How far they get on their own, and where their visits end.
- What stops them, on which screen, in their own words — and whether each obstacle is specific to your audience or would stop anyone.
Before the study you told us where you believed visitors struggle. We ran blind against those guesses; they are graded in the results.
- visitors will not trust what 'simulated participants' means
- the price feels steep before seeing a sample report
- signing up asks for commitment before showing any value
Who took part
Participants are simulated: each is a persona drawn from a calibrated pool of ten million, given a real browser and the product, and left alone. We say this once here and once in the appendix, and then call them participants, because that is the role they play in the sessions. The sample sizes below are the number of visits we ran, not an estimate of your traffic.
| Group | Selected on | Visits | Pool |
|---|---|---|---|
| Product people at software companies arrived: you saw ProductTex mentioned in a newsletter for software founders and clicked the link | role: Product; reads English | 120 | 156,015 |
| Developers using AI coding agents arrived: you saw ProductTex mentioned in a newsletter for software founders and clicked the link | coding ai usage frequency: Daily; reads English | 120 | 1,947 |
| Control: unrelated visitors | age: 18-24 or 25-34 or 35-44 or 45-54 or 55-64 or 65-74 or 75-84 or 85+; seniority: Entry or Mid or Senior or Lead / Principal or Manager or Director or VP or C-suite or Founder; reads English | 120 | 2,253,290 |
The control group has no plausible use for the product. It is here because people click things regardless of interest; a finding is only about your audience if the control did not raise it too.
Everyone here reads the language the product is in.
How the study ran
Each participant opened the site in a fresh browser and was told nothing except that they had access to it. They could click, type, choose, scroll or leave, for up to 30 actions. At every step we recorded what they saw, what they did, why they said they did it, what irritated them, and a one-word mood. We ran the whole study 2 times on different people and kept only what appeared every time. Complaints were grouped by wording, counted by people rather than mentions, and each group was traced back to the screen it was raised on and to what happened next. The full method, with every parameter of this study, is in the appendix.
Results
By the numbers
| Group | Visits | Median actions | Typed | Named something they'd use | Decided not for them |
|---|---|---|---|---|---|
| Product people at software companies | 120 | 6 | 5 | 78–80% | 83–85% |
| Developers using AI coding agents | 120 | 6 | 5 | 78–85% | 85–87% |
| Control: unrelated visitors | 120 | 4 | 3 | 72–77% | 92–95% |
“Named something they would use” and “decided it was not for them” overlap on purpose: a person can find something worth using and still leave. Where a cell shows a range, it is the two runs of this study, on different people: the low and the high. A single number means both runs agreed to the point. The two are not equally solid, and this report does not present them as though they were.
How far they got
Of 360 people:
310
did something & got past the first screen
of 360
of 360
13
typed into it
of 360
of 360
You have not told us what finishing looks like in your product, so this stops at “typed into it”. Name a milestone and it gets a rung.
Where Product people at software companies actually went, and what they did
Screens (boxed) and what personas did on them (rounded), in order — where they landed, what they did there, where that took them. Ribbon width is how many took that path. These are shares of the simulated visitors we ran, not a forecast of your real conversion — a simulated rate converges without being accurate (a real landing-to-signup rate is 1–5%). Rust ribbons are people who left. The control cohort is excluded — a flow picture of visitors with no reason to convert is not a finding.
Where Developers using AI coding agents actually went, and what they did
Screens (boxed) and what personas did on them (rounded), in order — where they landed, what they did there, where that took them. Ribbon width is how many took that path. These are shares of the simulated visitors we ran, not a forecast of your real conversion — a simulated rate converges without being accurate (a real landing-to-signup rate is 1–5%). Rust ribbons are people who left. The control cohort is excluded — a flow picture of visitors with no reason to convert is not a finding.
What got in the way
Each problem is rated on the usual four-point scale — blocker, major, minor, cosmetic — from what the sessions show: how many people met it, and how many of them left there. Under each one: where it happened, what people had just done, what they did next, how it compares with the control, what we would change, and how you would know the change worked. Open a row for the evidence. Holds means it appeared in both runs, on different people.
Findings in this report were not checked against screenshots. Verification was switched off when it was built.
Product people at software companies
50MajorholdsStill tiny text: 'ProductTex: quick read · s', 'Participants', '60 simulated visits'.
Dense unbroken paragraphs on a phone, no subheadings, and the tiny grey meta text ('ProductTex: quick read · s', '60 simulated visits') is close to unreadable.
Some text is still tiny — 'ProductTex: quick read', '60 simulated visits'
Tiny text on labels like 'ProductTex: quick read' and '60 simulated visits'
per run: run 1: 24 · run 2: 26
- Where
- /sample/report.html (127) — after ‘Open the sample report’, ‘9 BLOCKER Still no sample report’
- Then
- scrolling (100), stop (14)
- Ends visits
- 14 of 50 left within a step of saying it; 1 more were cut off by the step budget there
- Who
- Control: unrelated visitors 2/120, Developers using AI coding agents 1/120, Product people at software companies 50/120 — essentially only your audience raised it (control 2 of 120)
- Do
- On /sample/report.html: raise the contrast and size of ‘60 simulated visits’. The audit records the measured ratio; 4.5:1 and 14px are the floor.
- Check
- Re-run with a CSS override on ‘60 simulated visits’; the contrast audit must report zero failures on it before the re-run counts.
50 people · cog patience: Moderate (26), academic field: Business (16), region: South Asia (13), region: Sub-Saharan Africa (13)
27MajorholdsSmall text and fiddly controls
Some text is tiny ('ProductTex: quick read · s', '60 simulated visits') and the BLOCKER controls look fiddly to hit.
Tiny controls and text make the page fiddly to read.
Tiny 'BLOCKER' labels and small controls are fiddly
per run: run 1: 11 · run 2: 16
- Where
- /sample/report.html (43); / (4) — after ‘Open the sample report’, ‘10 BLOCKER Raw JSON error on scr’
- Then
- scrolling (35), stop (6)
- Ends visits
- 6 of 27 left within a step of saying it
- Who
- Control: unrelated visitors 0/120, Developers using AI coding agents 0/120, Product people at software companies 27/120 — essentially only your audience raised it (control 0 of 120)
- Do
- On /sample/report.html: raise the contrast and size of ‘10’. The audit records the measured ratio; 4.5:1 and 14px are the floor.
- Check
- Re-run with a CSS override on ‘10’; the contrast audit must report zero failures on it before the re-run counts.
No capture of this moment was kept for this group.
27 people · cog patience: Moderate (15), academic field: Business (7), academic field: STEM (7), age bracket: 35-44 (6)
28MinorholdsTiny 'Sample report' label
Quotes repeat the label word-for-word: 'Sample report' — 2 people would use it. 'the sample report'. That reads like filler, not evidence.
Little text like 'SAMPLE REPORT' is very small, I have to squint
The 'SAMPLE REPORT' label at the top is tiny and faint — I nearly missed it.
per run: run 1: 16 · run 2: 12
- Where
- /sample (30); /sample/report.html (3) — after ‘Read a sample report’, ‘Sample report’
- Then
- Open the sample report (23), scrolling (7)
- Ends visits
- 2 of 28 left within a step of saying it
- Who
- Control: unrelated visitors 1/120, Developers using AI coding agents 0/120, Product people at software companies 28/120 — essentially only your audience raised it (control 1 of 120)
- Do
- On /sample: raise the contrast and size of ‘SAMPLE REPORT’. The audit records the measured ratio; 4.5:1 and 14px are the floor.
- Check
- Re-run with a CSS override on ‘SAMPLE REPORT’; the contrast audit must report zero failures on it before the re-run counts.
28 people · cog patience: Moderate (16), academic field: STEM (11), academic field: Business (7), region: South Asia (7)
33MajorshakySmall grey labels and tiny text are hard to read
Some of the text is very small — the kicker, the '60 simulated visits' label, the blocker titles are hard to read.
Small grey text is hard to read
The type is genuinely small for a page whose selling point is being a quick read
per run: run 1: 20 · run 2: 13
- Where
- /sample/report.html (37); / (14) — after ‘Open the sample report’, ‘10 BLOCKER Raw JSON error on scr’
- Then
- scrolling (36), stop (8)
- Ends visits
- 8 of 33 left within a step of saying it
- Who
- Control: unrelated visitors 1/120, Developers using AI coding agents 2/120, Product people at software companies 33/120 — essentially only your audience raised it (control 1 of 120)
- Do
- On /sample/report.html: raise the contrast and size of ‘60 simulated visits’. The audit records the measured ratio; 4.5:1 and 14px are the floor.
- Check
- Re-run with a CSS override on ‘60 simulated visits’; the contrast audit must report zero failures on it before the re-run counts.
33 people · cog patience: Moderate (20), academic field: Business (11), academic field: STEM (11), region: Sub-Saharan Africa (8)
23CosmeticholdsNo video anywhere, just text.
Still no video anywhere, and I read best by watching — a wall of methodology prose on a phone is hard work
No video anywhere, only text — again
Two walls of text and still no report on screen
per run: run 1: 13 · run 2: 10
- Where
- /sample/report.html (28); /sample (12) — after ‘Read a sample report’, ‘Open the sample report’
- Then
- scrolling (24), Open the sample report (11)
- Ends visits
- 9 of 23 left within a step of saying it
- Who
- Control: unrelated visitors 3/120, Developers using AI coding agents 0/120, Product people at software companies 23/120 — 7.7× the control
- Do
- No template fits this one — open the row and read what they said.
- Check
- Open the row.
23 people · cog patience: Moderate (13), academic field: Business (7), academic field: Interdisciplinary (7), region: Sub-Saharan Africa (7)
18CosmeticholdsStill no video, and the page says it continues a long way down
A single dense paragraph under SUMMARY with no break, and the page tells me it continues a long way further down.
No video anywhere after all this scrolling, and 'the page continues a long way further down' is a bad sign
Still no video, and the page says it continues a long way down
per run: run 1: 9 · run 2: 9
- Where
- /sample/report.html (21); / (4) — after ‘Open the sample report’, ‘Read a sample report’
- Then
- scrolling (12), stop (10)
- Ends visits
- 10 of 18 left within a step of saying it
- Who
- Control: unrelated visitors 0/120, Developers using AI coding agents 0/120, Product people at software companies 18/120 — essentially only your audience raised it (control 0 of 120)
- Do
- No template fits this one — open the row and read what they said.
- Check
- Open the row.
18 people · cog patience: Moderate (11), academic field: STEM (10), region: South Asia (7), region: Sub-Saharan Africa (4)
20MinorholdsSmall text labels like 'Usability studies with sim' and 'The gap' are hard to read on a phone.
The 'USABILITY STUDIES WITH SIMULATED PARTICIPANTS' label is tiny and easy to miss, yet it's the most important qualifier on the page.
Some text is very small ('USABILITY STUDIES WITH SIMULATED PARTICIPANTS', 'THE GAP') and I had to squint
The tiny label 'USABILITY STUDIES WITH SIMULATED PARTICIPANTS' is easy to miss.
per run: run 1: 9 · run 2: 11
- Where
- / (25)
- Then
- Read a sample report (13), scrolling (7)
- Ends visits
- 2 of 20 left within a step of saying it
- Who
- Control: unrelated visitors 0/120, Developers using AI coding agents 0/120, Product people at software companies 20/120 — essentially only your audience raised it (control 0 of 120)
- Do
- On /: raise the contrast and size of ‘The gap’. The audit records the measured ratio; 4.5:1 and 14px are the floor.
- Check
- Re-run with a CSS override on ‘The gap’; the contrast audit must report zero failures on it before the re-run counts.
20 people · cog patience: Moderate (10), academic field: Business (8), region: South Asia (6), age bracket: 25-34 (4)
18MinorholdsTiny labels in the nav ('Docs', 'Sample report') are fiddly
The 'SAMPLE REPORT' label and the nav items like 'Docs' are tiny and easy to miss
The sample report text is tiny and the nav links are fiddly little targets.
Some labels like 'Sample report' and 'Docs' are tiny and fiddly to hit
per run: run 1: 6 · run 2: 12
- Where
- /sample (13); / (8) — after ‘Read a sample report’, ‘Sample report’
- Then
- Open the sample report (10), scrolling (9)
- Ends visits
- 3 of 18 left within a step of saying it
- Who
- Control: unrelated visitors 0/120, Developers using AI coding agents 0/120, Product people at software companies 18/120 — essentially only your audience raised it (control 0 of 120)
- Do
- On /sample: raise the contrast and size of ‘Docs’. The audit records the measured ratio; 4.5:1 and 14px are the floor.
- Check
- Re-run with a CSS override on ‘Docs’; the contrast audit must report zero failures on it before the re-run counts.
18 people · cog patience: Moderate (9), region: South Asia (7), academic field: Business (5), academic field: STEM (5)
18Cosmeticholds'Findings in this report were not checked against screenshots. Verification was switched off when it was built.' — so how do I trust any of this?
'Findings in this report were not checked against screenshots. Verification was switched off when it was built' — that quietly guts the credibility of everything above it, and it's buried in small grey text.
The note that verification was switched off undercuts the whole point of showing evidence
The disclaimer that verification was switched off undercuts a report whose value is supposedly its evidence
per run: run 1: 8 · run 2: 10
- Where
- /sample/report.html (36) — after ‘10 BLOCKER Raw JSON error on scr’, ‘9 BLOCKER Still no sample report’
- Then
- 10 BLOCKER Raw JSON error on scr (18), 9 BLOCKER Still no sample report (9)
- Ends visits
- 4 of 18 left within a step of saying it
- Who
- Control: unrelated visitors 0/120, Developers using AI coding agents 0/120, Product people at software companies 18/120 — essentially only your audience raised it (control 0 of 120)
- Do
- No template fits this one — open the row and read what they said.
- Check
- Open the row.
18 people · region: South Asia (7), cog patience: Moderate (7), age bracket: 25-34 (5), academic field: Arts (5)
16CosmeticholdsFindings not checked against screenshots — why should I trust this?
That note saying findings were 'not checked against screenshots' and verification was switched off makes me trust the whole list less.
Warnings that findings weren't checked against screenshots don't build trust
A note saying findings weren't checked against screenshots — so why should I trust the severity ratings?
per run: run 1: 9 · run 2: 7
- Where
- /sample/report.html (38) — after ‘10 BLOCKER Raw JSON error on scr’, ‘6 BLOCKER Raw JSON error instead’
- Then
- 10 BLOCKER Raw JSON error on scr (21), scrolling (6)
- Ends visits
- 3 of 16 left within a step of saying it
- Who
- Control: unrelated visitors 1/120, Developers using AI coding agents 0/120, Product people at software companies 16/120 — essentially only your audience raised it (control 1 of 120)
- Do
- No template fits this one — open the row and read what they said.
- Check
- Open the row.
16 people · cog patience: Moderate (7), academic field: Business (6), age bracket: 45-54 (6), academic field: STEM (6)
21MinorshakyThe logo and menu are very small and fiddly to tap.
Still very text-heavy for a phone screen, and 'Sample report' plus the logo and Menu are tiny targets.
The menu button is tiny and fiddly on a phone
The logo and menu are very small and fiddly to tap.
per run: run 1: 15 · run 2: 6
- Where
- / (18); /sample (15) — after ‘Read a sample report’, ‘typing into a field’
- Then
- scrolling (14), Read a sample report (10)
- Ends visits
- 2 of 21 left within a step of saying it
- Who
- Control: unrelated visitors 0/120, Developers using AI coding agents 0/120, Product people at software companies 21/120 — essentially only your audience raised it (control 0 of 120)
- Do
- On /: raise the contrast and size of ‘The Menu button’. The audit records the measured ratio; 4.5:1 and 14px are the floor.
- Check
- Re-run with a CSS override on ‘The Menu button’; the contrast audit must report zero failures on it before the re-run counts.
21 people · cog patience: Moderate (13), academic field: Business (9), region: Sub-Saharan Africa (7), cog patience: High (6)
18CosmeticholdsThe headline numbers don't reconcile plainly: 24 of 30 found something they'd use, yet 29 of the same 30 decided it wasn't for them — I can't tell if that's two different runs of 30 and the page never says so.
The headline numbers don't reconcile plainly: 24 of 30 found something they'd use, yet 29 of the same 30 decided it wasn't for them — I can't tell if that's two different runs of 30 and the page never says so.
The headline says 24 of 30 would use it, then immediately says 29 decided it was not for them — I had to re-read twice to see those are different pools
The numbers don't reconcile: '60 simulated visits' in the summary, but every stat is 'of 30', and 24 found something while 29 decided it wasn't for them — that's 53 opinions from 30 people.
per run: run 1: 11 · run 2: 7
- Where
- /sample/report.html (20) — after ‘Open the sample report’
- Then
- scrolling (19), stop (1)
- Ends visits
- 1 of 18 left within a step of saying it
- Who
- Control: unrelated visitors 0/120, Developers using AI coding agents 0/120, Product people at software companies 18/120 — essentially only your audience raised it (control 0 of 120)
- Do
- No template fits this one — open the row and read what they said.
- Check
- Open the row.
18 people · cog patience: Moderate (9), region: Sub-Saharan Africa (6), academic field: Business (5), age bracket: 13-17 (5)
Developers using AI coding agents
50BlockershakyTiny text ('ProductTex: quick read · s', '60 simulated visits', 'Participants')
Some text renders very small ('ProductTex: quick read · s', '60 simulated visits'), which is painful given I'm reading through an accessibility tool.
Tiny type again: 'ProductTex: quick read · s', '60 simulated visits', 'Participants'.
Tiny type: 'ProductTex: quick read · s', '60 simulated visits', 'Participants'.
per run: run 1: 21 · run 2: 29
- Where
- /sample/report.html (162); / (1) — after ‘Open the sample report’, ‘10 BLOCKER Raw JSON error on scr’
- Then
- scrolling (128), stop (17)
- Ends visits
- 17 of 50 left within a step of saying it
- Who
- Control: unrelated visitors 0/120, Developers using AI coding agents 50/120, Product people at software companies 1/120 — essentially only your audience raised it (control 0 of 120)
- Do
- On /sample/report.html: raise the contrast and size of ‘60 simulated visits’. The audit records the measured ratio; 4.5:1 and 14px are the floor.
- Check
- Re-run with a CSS override on ‘60 simulated visits’; the contrast audit must report zero failures on it before the re-run counts.
50 people · cog patience: Moderate (24), age bracket: 25-34 (17), region: South Asia (17), academic field: Interdisciplinary (16)
24MajorholdsTiny text and fiddly tap targets on a phone
Tiny labels ('60 simulated visits', 'Participants') and fiddly tap targets like '10' and 'BLOCKER'
Tiny tap targets top-left and top-right — hard to hit one-handed
Tiny 'ProductTex' and 'Menu' are fiddly on mobile.
per run: run 1: 13 · run 2: 11
- Where
- /sample/report.html (38); / (13) — after ‘10 BLOCKER Raw JSON error on scr’, ‘9 BLOCKER Still no sample report’
- Then
- scrolling (22), 10 BLOCKER Raw JSON error on scr (10)
- Ends visits
- 5 of 24 left within a step of saying it
- Who
- Control: unrelated visitors 0/120, Developers using AI coding agents 24/120, Product people at software companies 2/120 — essentially only your audience raised it (control 0 of 120)
- Do
- On /sample/report.html: raise the contrast and size of ‘10’. The audit records the measured ratio; 4.5:1 and 14px are the floor.
- Check
- Re-run with a CSS override on ‘10’; the contrast audit must report zero failures on it before the re-run counts.
No capture of this moment was kept for this group.
24 people · cog patience: Moderate (14), region: South Asia (9), academic field: Interdisciplinary (8), age bracket: 18-24 (7)
35MajorshakyTiny text and fiddly small controls
Small print and tiny fiddly controls ('10', 'BLOCKER') — I'm on desktop, there's no reason for that.
Small, fiddly controls ('10', 'BLOCKER') — annoying when I'm reading closely, not tapping randomly.
Small, fiddly hit targets and very small type on the controls I did see.
per run: run 1: 21 · run 2: 14
- Where
- /sample/report.html (62); / (3) — after ‘Open the sample report’, ‘10 BLOCKER Raw JSON error on scr’
- Then
- scrolling (41), stop (10)
- Ends visits
- 10 of 35 left within a step of saying it
- Who
- Control: unrelated visitors 0/120, Developers using AI coding agents 35/120, Product people at software companies 0/120 — essentially only your audience raised it (control 0 of 120)
- Do
- On /sample/report.html: raise the contrast and size of ‘10’. The audit records the measured ratio; 4.5:1 and 14px are the floor.
- Check
- Re-run with a CSS override on ‘10’; the contrast audit must report zero failures on it before the re-run counts.
35 people · cog patience: Moderate (16), region: South Asia (13), academic field: Interdisciplinary (13), age bracket: 25-34 (11)
25MinorholdsTiny text like 'Usability studies with sim'
The little 'Usability studies with simulated participants' label above the headline is tiny and I nearly skipped it, but it's actually the most important context.
The tiny all-caps line 'USABILITY STUDIES WITH SIMULATED PARTICIPANTS' is fiddly to read
Some text is very small, like 'Usability studies with sim' and 'The gap'.
per run: run 1: 12 · run 2: 13
- Where
- / (27)
- Then
- Read a sample report (17), stop (4)
- Ends visits
- 4 of 25 left within a step of saying it
- Who
- Control: unrelated visitors 0/120, Developers using AI coding agents 25/120, Product people at software companies 0/120 — essentially only your audience raised it (control 0 of 120)
- Do
- On /: raise the contrast and size of ‘Usability studies with sim’. The audit records the measured ratio; 4.5:1 and 14px are the floor.
- Check
- Re-run with a CSS override on ‘Usability studies with sim’; the contrast audit must report zero failures on it before the re-run counts.
25 people · cog patience: Moderate (12), academic field: Interdisciplinary (9), age bracket: 25-34 (8), age bracket: 18-24 (8)
28MinorholdsThe 'Sample report' label is tiny
The 'SAMPLE REPORT' label is tiny, and the top controls (ProductTex, Menu) are fiddly to tap on a phone
The tiny 'SAMPLE REPORT' label is fiddly to read
The 'Sample report' text is very small
per run: run 1: 17 · run 2: 11
- Where
- /sample (29) — after ‘Read a sample report’, ‘See how this reads in a report’
- Then
- Open the sample report (24), scrolling (4)
- Ends visits
- 1 of 28 left within a step of saying it
- Who
- Control: unrelated visitors 0/120, Developers using AI coding agents 28/120, Product people at software companies 0/120 — essentially only your audience raised it (control 0 of 120)
- Do
- On /sample: raise the contrast and size of ‘Sample report’. The audit records the measured ratio; 4.5:1 and 14px are the floor.
- Check
- Re-run with a CSS override on ‘Sample report’; the contrast audit must report zero failures on it before the re-run counts.
28 people · cog patience: Moderate (11), cog patience: Low (9), academic field: Interdisciplinary (8), age bracket: 18-24 (7)
23MinorholdsSome nav links and small labels are tiny and fiddly.
Small fiddly nav items at the top (Docs, Method) that an accessibility tool would struggle with
Nav items like 'Docs' and the logo look small enough to miss
Nav links like Docs and the logo are tiny and fiddly to hit
per run: run 1: 12 · run 2: 11
- Where
- /sample (20); / (11) — after ‘Read a sample report’, ‘Sample report’
- Then
- Open the sample report (16), Read a sample report (8)
- Ends visits
- 2 of 23 left within a step of saying it
- Who
- Control: unrelated visitors 0/120, Developers using AI coding agents 23/120, Product people at software companies 0/120 — essentially only your audience raised it (control 0 of 120)
- Do
- On /sample: raise the contrast and size of ‘Docs’. The audit records the measured ratio; 4.5:1 and 14px are the floor.
- Check
- Re-run with a CSS override on ‘Docs’; the contrast audit must report zero failures on it before the re-run counts.
23 people · cog patience: Moderate (13), age bracket: 25-34 (7), age bracket: 18-24 (7), cog patience: Low (6)
18MinorholdsTiny grey labels are hard to read on a 390px phone.
The tiny grey labels ('60 simulated visits', 'Participants') are hard to read on a phone outdoors
Small grey text is hard to read
Tiny grey labels are hard to read on a 390px phone.
per run: run 1: 7 · run 2: 11
- Where
- /sample/report.html (14); / (13) — after ‘Open the sample report’, ‘An example.’
- Then
- scrolling (19), stop (4)
- Ends visits
- 4 of 18 left within a step of saying it; 1 more were cut off by the step budget there
- Who
- Control: unrelated visitors 0/120, Developers using AI coding agents 18/120, Product people at software companies 2/120 — essentially only your audience raised it (control 0 of 120)
- Do
- On /sample/report.html: raise the contrast and size of ‘60 simulated visits’. The audit records the measured ratio; 4.5:1 and 14px are the floor.
- Check
- Re-run with a CSS override on ‘60 simulated visits’; the contrast audit must report zero failures on it before the re-run counts.
18 people · cog patience: Moderate (8), academic field: Interdisciplinary (8), academic field: STEM (5), region: South Asia (5)
17CosmeticholdsNo screenshot, sample report, or video anywhere.
Still no screenshot, video or working example of the actual report anywhere in this section
Still no screenshot, image, or working sample report anywhere.
Still no price, no report length, no screenshot of a working deliverable
per run: run 1: 8 · run 2: 9
- Where
- /sample/report.html (28); / (10) — after ‘10 BLOCKER Raw JSON error on scr’, ‘See pricing’
- Then
- scrolling (26), stop (9)
- Ends visits
- 9 of 17 left within a step of saying it
- Who
- Control: unrelated visitors 1/120, Developers using AI coding agents 17/120, Product people at software companies 3/120 — essentially only your audience raised it (control 1 of 120)
- Do
- No template fits this one — open the row and read what they said.
- Check
- Open the row.
17 people · region: South Asia (7), cog patience: Moderate (7), academic field: STEM (6), age bracket: 25-34 (6)
15MinorholdsThe ProductTex logo and Menu are tiny on mobile — fiddly tap targets
The ProductTex and Menu links at the top are tiny and fiddly to hit on a phone
Tiny tap targets for the logo and Menu
Header links (ProductTex, Menu) are tiny and fiddly on mobile.
per run: run 1: 8 · run 2: 7
- Where
- / (18); /sample (8) — after ‘Read a sample report’, ‘Menu’
- Then
- scrolling (10), Read a sample report (7)
- Ends visits
- 1 of 15 left within a step of saying it
- Who
- Control: unrelated visitors 0/120, Developers using AI coding agents 15/120, Product people at software companies 0/120 — essentially only your audience raised it (control 0 of 120)
- Do
- On /: enlarge the tap targets — 44×44px minimum with space between them — starting with ‘Menu button’, and check the layout at 390px wide.
- Check
- Re-run the phone cohort after the change; 'small' and 'fiddly' should leave the themes. In your analytics: tap error rate or rage taps on / for mobile sessions.
15 people · cog patience: Moderate (7), age bracket: 25-34 (6), academic field: Business (4), cog patience: Low (4)
15MinorholdsTiny text and small, fiddly row hit targets
Blockers are labelled 'BLOCKER' in tiny all-caps and the hit targets look fiddly, so I keep almost missing them
Controls '10' and 'BLOCKER' are small, fiddly hit targets.
'10' and 'BLOCKER' are tiny fiddly hit targets before I even reach the row text.
per run: run 1: 5 · run 2: 10
- Where
- /sample/report.html (28) — after ‘10 BLOCKER Raw JSON error on scr’, ‘9 BLOCKER Still no sample report’
- Then
- scrolling (11), 10 BLOCKER Raw JSON error on scr (8)
- Ends visits
- 1 of 15 left within a step of saying it
- Who
- Control: unrelated visitors 2/120, Developers using AI coding agents 15/120, Product people at software companies 0/120 — 7.5× the control
- Do
- On /sample/report.html: enlarge the tap targets — 44×44px minimum with space between them — starting with ‘10’, and check the layout at 390px wide.
- Check
- Re-run the phone cohort after the change; 'small' and 'fiddly' should leave the themes. In your analytics: tap error rate or rage taps on /sample/report.html for mobile sessions.
15 people · cog patience: Moderate (8), age bracket: 25-34 (6), region: South Asia (6), academic field: Business (6)
19CosmeticholdsThe page is a wall of text and I'd rather watch a video.
Still zero visuals — I've read a definition of a four-point scale and now I'm reading a list, on a medium where I'd rather watch someone do this.
Big block of text, no video — I'd rather watch a demo
Walls of text with no video anywhere — I'd rather be shown this than read it on my phone
per run: run 1: 9 · run 2: 10
- Where
- / (16); /sample/report.html (7) — after ‘Read a sample report’, ‘Open the sample report’
- Then
- scrolling (12), Read a sample report (7)
- Ends visits
- 2 of 19 left within a step of saying it
- Who
- Control: unrelated visitors 0/120, Developers using AI coding agents 19/120, Product people at software companies 0/120 — essentially only your audience raised it (control 0 of 120)
- Do
- No template fits this one — open the row and read what they said.
- Check
- Open the row.
19 people · cog patience: Moderate (9), age bracket: 25-34 (7), academic field: STEM (6), academic field: Interdisciplinary (5)
15CosmeticholdsThe 'verification switched off' disclaimer undermines trust
'Verification was switched off when it was built' sits in small grey text under a report that otherwise implies rigor
The 'verification was switched off' note makes the evidence feel weak
The 'verification was switched off' note undermines trust in the numbers
per run: run 1: 5 · run 2: 10
- Where
- /sample/report.html (38) — after ‘10 BLOCKER Raw JSON error on scr’, ‘12 COSMETIC A 'Pretty-print' che’
- Then
- 10 BLOCKER Raw JSON error on scr (15), scrolling (12)
- Ends visits
- 1 of 15 left within a step of saying it
- Who
- Control: unrelated visitors 0/120, Developers using AI coding agents 15/120, Product people at software companies 0/120 — essentially only your audience raised it (control 0 of 120)
- Do
- No template fits this one — open the row and read what they said.
- Check
- Open the row.
15 people · cog patience: Moderate (11), region: South Asia (6), academic field: Business (5), age bracket: 25-34 (5)
Control: unrelated visitors
38MajorholdsTiny labels: 'ProductTex: quick read · s', '60 simulated visits', 'Participants'
Very small, low-contrast text ('ProductTex: quick read · s', '60 simulated visits', 'Participants') and tiny '10 BLOCKER' chips that are fiddly to read.
Tiny text and small controls: 'ProductTex: quick read · s', '60 simulated visits', the 'BLOCKER' chips.
Tiny type on 'ProductTex: quick read · s' and '60 simulated visits' — squinting again.
per run: run 1: 17 · run 2: 21
- Where
- /sample/report.html (115); /sample (1) — after ‘Open the sample report’, ‘10 BLOCKER Raw JSON error on scr’
- Then
- scrolling (81), 10 BLOCKER Raw JSON error on scr (12)
- Ends visits
- 10 of 38 left within a step of saying it
- Who
- Control: unrelated visitors 38/120, Developers using AI coding agents 0/120, Product people at software companies 1/120
- Do
- On /sample/report.html: raise the contrast and size of ‘60 simulated visits’. The audit records the measured ratio; 4.5:1 and 14px are the floor.
- Check
- Re-run with a CSS override on ‘60 simulated visits’; the contrast audit must report zero failures on it before the re-run counts.
38 people · cog patience: Moderate (21), region: South Asia (16), age bracket: 25-34 (12), academic field: Business (10)
31MajorholdsSome text and small controls are tiny and fiddly.
Small fiddly controls ('10', 'BLOCKER') that look tappable but I can't tell what they do.
Fiddly little controls labelled 'BLOCKER' and '10' that I can't tell apart from text.
Some controls below are tiny ('10', 'BLOCKER') and will be hard to tap.
per run: run 1: 18 · run 2: 13
- Where
- /sample/report.html (52); / (4) — after ‘Open the sample report’, ‘9 BLOCKER Still no sample report’
- Then
- scrolling (36), stop (8)
- Ends visits
- 8 of 31 left within a step of saying it
- Who
- Control: unrelated visitors 31/120, Developers using AI coding agents 0/120, Product people at software companies 1/120
- Do
- On /sample/report.html: raise the contrast and size of ‘BLOCKER’. The audit records the measured ratio; 4.5:1 and 14px are the floor.
- Check
- Re-run with a CSS override on ‘BLOCKER’; the contrast audit must report zero failures on it before the re-run counts.
31 people · cog patience: Moderate (15), academic field: STEM (12), region: South Asia (11), age bracket: 35-44 (10)
19MajorholdsSome of the small grey text labels are hard to read.
The small grey labels and 'BLOCKER' tags are hard to read and hit on my low-bandwidth connection.
Some of the small text labels are hard to read.
Some of the small grey text labels are hard to read.
per run: run 1: 8 · run 2: 11
- Where
- /sample/report.html (32); / (11) — after ‘Open the sample report’, ‘10 BLOCKER Raw JSON error on scr’
- Then
- scrolling (27), stop (6)
- Ends visits
- 6 of 19 left within a step of saying it
- Who
- Control: unrelated visitors 19/120, Developers using AI coding agents 0/120, Product people at software companies 1/120
- Do
- On /sample/report.html: raise the contrast and size of ‘BLOCKER’. The audit records the measured ratio; 4.5:1 and 14px are the floor.
- Check
- Re-run with a CSS override on ‘BLOCKER’; the contrast audit must report zero failures on it before the re-run counts.
No capture of this moment was kept for this group.
19 people · cog patience: High (9), academic field: STEM (8), age bracket: 35-44 (6), academic field: Business (6)
18BlockerholdsRaw JSON error instead of a real sample report
The report itself says the sample report page throws a raw JSON error — that's exactly what I got when I opened it, so the whole thing feels half-finished
Mentions a raw JSON error on the sample report page — not reassuring
The sample report link just spat out a raw code error instead of a report
per run: run 1: 6 · run 2: 12
- Where
- /sample/report.html (23) — after ‘Open the sample report’, ‘12 BLOCKER Sample report link le’
- Then
- scrolling (14), stop (6)
- Ends visits
- 6 of 18 left within a step of saying it
- Who
- Control: unrelated visitors 18/120, Developers using AI coding agents 0/120, Product people at software companies 0/120
- Do
- On /sample/report.html: replace the error with a real page — your site's header, a sentence saying what happened, and a link back to where they came from — then fix what sent them there (‘Open the sample report’).
- Check
- Open /sample/report.html yourself: it must return a page, not an error. Then re-run the same people (Fix check); nobody should end their visit there. In your analytics: error-page views and exits from /sample/report.html.
No capture of this moment was kept for this group.
18 people · cog patience: Moderate (10), academic field: STEM (7), age bracket: 35-44 (6), region: Sub-Saharan Africa (6)
14BlockerholdsSome of the small text is hard to read on mobile
The page is a wall of text with no video, and some of the smaller text is hard to read on mobile.
Wall of text on mobile
Wall of text on mobile with tiny type
per run: run 1: 8 · run 2: 6
- Where
- /sample/report.html (15); / (8) — after ‘Open the sample report’, ‘Read a sample report’
- Then
- scrolling (10), stop (8)
- Ends visits
- 8 of 14 left within a step of saying it
- Who
- Control: unrelated visitors 14/120, Developers using AI coding agents 2/120, Product people at software companies 2/120
- Do
- On /sample/report.html: raise the contrast and size of the text they found hard to read. The audit records the measured ratio; 4.5:1 and 14px are the floor.
- Check
- Re-run with a CSS override on the text they found hard to read; the contrast audit must report zero failures on it before the re-run counts.
14 people · age bracket: 35-44 (7), cog patience: Moderate (7), academic field: STEM (4), region: Sub-Saharan Africa (4)
27MinorholdsThe 'Sample report' label is very small
'Sample report' in the nav and 'Open the sample report' in the body seem like the same destination — I keep having to guess which is which.
Small labels again ('Sample report' in the top bar is tiny)
The tiny 'SAMPLE REPORT' label above the headline is almost unreadable.
per run: run 1: 13 · run 2: 14
- Where
- /sample (26); /sample/report.html (3) — after ‘Read a sample report’, ‘Sample report’
- Then
- Open the sample report (24), scrolling (4)
- Ends visits
- 1 of 27 left within a step of saying it
- Who
- Control: unrelated visitors 27/120, Developers using AI coding agents 0/120, Product people at software companies 1/120
- Do
- On /sample: raise the contrast and size of ‘Sample report’. The audit records the measured ratio; 4.5:1 and 14px are the floor.
- Check
- Re-run with a CSS override on ‘Sample report’; the contrast audit must report zero failures on it before the re-run counts.
27 people · cog patience: Moderate (14), academic field: STEM (10), region: South Asia (9), age bracket: 35-44 (8)
14CosmeticholdsPage says it continues a long way further down
The page says it 'continues a long way further down' — that's a warning, not an invitation, and it makes me dread scrolling
The line telling me the page 'continues a long way further down' still reads like a threat, not an invitation
It says there's more above and a long way further down — I have no idea how long this report is
per run: run 1: 4 · run 2: 10
- Where
- /sample/report.html (18); / (1) — after ‘Open the sample report’, ‘Get a free Quick read’
- Then
- stop (10), scrolling (8)
- Ends visits
- 10 of 14 left within a step of saying it
- Who
- Control: unrelated visitors 14/120, Developers using AI coding agents 0/120, Product people at software companies 0/120
- Do
- No template fits this one — open the row and read what they said.
- Check
- Open the row.
14 people · cog patience: Moderate (7), region: South Asia (6), age bracket: 45-54 (4), academic field: STEM (4)
29CosmeticholdsNo video, just a wall of text
Another wall of text with no video anywhere, when I was promised something to watch
No video, just a wall of text
Another wall of text before I get to the actual report
per run: run 1: 14 · run 2: 15
- Where
- /sample/report.html (19); / (15) — after ‘Read a sample report’, ‘Open the sample report’
- Then
- scrolling (16), stop (13)
- Ends visits
- 13 of 29 left within a step of saying it
- Who
- Control: unrelated visitors 29/120, Developers using AI coding agents 4/120, Product people at software companies 4/120
- Do
- An impression, not a defect. Worth knowing; not worth a sprint unless it survives the fixes above.
- Check
- Not something to verify. Re-check after the fixes above; if it persists, then decide.
29 people · cog patience: Moderate (13), academic field: STEM (13), age bracket: 35-44 (9), region: South Asia (8)
14MinorholdsThe small nav links up top are fiddly
The small grey 'SAMPLE REPORT' label and the tiny nav words are fiddly at this text size.
Some of the nav text is tiny
Some nav text and controls are tiny and fiddly.
per run: run 1: 8 · run 2: 6
- Where
- / (11); /sample (8) — after ‘Read a sample report’, ‘Open the sample report’
- Then
- Open the sample report (8), scrolling (6)
- Ends visits
- 2 of 14 left within a step of saying it
- Who
- Control: unrelated visitors 14/120, Developers using AI coding agents 0/120, Product people at software companies 0/120
- Do
- On /: raise the contrast and size of ‘SAMPLE REPORT’. The audit records the measured ratio; 4.5:1 and 14px are the floor.
- Check
- Re-run with a CSS override on ‘SAMPLE REPORT’; the contrast audit must report zero failures on it before the re-run counts.
14 people · academic field: STEM (6), region: South Asia (5), cog patience: Moderate (5), age bracket: 35-44 (4)
13MinorholdsRows are small and fiddly to hit
The clickable finding rows are small targets for a list I'm meant to work through.
The blocker rows look fiddly to hit
The blocker rows look small and fiddly to hit
per run: run 1: 7 · run 2: 6
- Where
- /sample/report.html (37) — after ‘10 BLOCKER Raw JSON error on scr’, ‘9 BLOCKER Still no sample report’
- Then
- 10 BLOCKER Raw JSON error on scr (19), scrolling (12)
- Ends visits
- 2 of 13 left within a step of saying it
- Who
- Control: unrelated visitors 13/120, Developers using AI coding agents 1/120, Product people at software companies 0/120
- Do
- On /sample/report.html: raise the contrast and size of the text they found hard to read. The audit records the measured ratio; 4.5:1 and 14px are the floor.
- Check
- Re-run with a CSS override on the text they found hard to read; the contrast audit must report zero failures on it before the re-run counts.
13 people · cog patience: Moderate (6), academic field: STEM (5), age bracket: 35-44 (5), cog patience: High (5)
17CosmeticholdsThe disclaimer that findings were not checked against screenshots
Verification was switched off and findings weren't checked against screenshots — I only noticed this line now, and it quietly undercuts every rating above it.
Verification was switched off, so findings were not checked against screenshots.
Told the findings weren't checked against screenshots, which undercuts them
per run: run 1: 7 · run 2: 10
- Where
- /sample/report.html (38) — after ‘10 BLOCKER Raw JSON error on scr’, ‘9 BLOCKER Still no sample report’
- Then
- 10 BLOCKER Raw JSON error on scr (24), scrolling (7)
- Ends visits
- 1 of 17 left within a step of saying it
- Who
- Control: unrelated visitors 17/120, Developers using AI coding agents 0/120, Product people at software companies 1/120
- Do
- No template fits this one — open the row and read what they said.
- Check
- Open the row.
17 people · cog patience: Moderate (9), region: South Asia (7), cog patience: High (7), academic field: STEM (6)
13CosmeticholdsNo video anywhere, and I'd rather watch
Still a wall of text on a phone, still no video — this is the third screen in a row where I'd rather watch thirty seconds than read six paragraphs.
It's a wall of text and I'd rather watch a short video.
Still no video anywhere — I'm skimming past paragraphs I'd rather watch.
per run: run 1: 6 · run 2: 7
- Where
- / (10); /sample/report.html (10) — after ‘Open the sample report’, ‘Sample report’
- Then
- scrolling (10), Open the sample report (6)
- Ends visits
- 3 of 13 left within a step of saying it
- Who
- Control: unrelated visitors 13/120, Developers using AI coding agents 0/120, Product people at software companies 0/120
- Do
- No template fits this one — open the row and read what they said.
- Check
- Open the row.
13 people · region: South Asia (6), cog patience: Moderate (6), age bracket: 25-34 (5), cog patience: High (5)
What is working — do not break these
Nobody pointed participants at any of this. It is what they found by themselves and said they would come back for, and what caught their eye. Any change to the screens above should leave these intact.
Product people at software companies
- A sample report — 51 people would use it.
The free sample report, once I can actually open it.
- The free Quick read on my site — 15 people would use it.
A free quick read of my own site, if it opens with a blunt number like this instead of a table of demographics.
- The sample report, to judge whether the output is worth anything — 14 people would use it.
The sample report itself, to judge whether ProductTex's simulated feedback is worth paying attention to.
- The sample report, to judge the output quality before committing — 9 people would use it.
Read a sample report — if they ever fix the link, it is the only way to judge quality before handing over a URL
- A sample report, to judge whether the findings are specific enough to act on. — 9 people would use it.
The ranked blocker list with exit counts, to see whether the simulated complaints are specific enough to act on.
- The free Quick read on a site I control — caught 30 people's eye
- The sample report — caught 30 people's eye
- That a re-run of the same 30 people is included after the fix. — caught 28 people's eye
Developers using AI coding agents
- the sample report — 40 people would use it.
Reading a sample report to judge if the output is trustworthy
- The free Quick read on my site — 17 people would use it.
A free Quick read on my own site, if it runs with no setup and I can see a sample of its output first
- Open the sample report — 16 people would use it.
Open the sample report and read the summary section — but I still want to see the actual findings, not the intro to them
- A sample report, just to judge whether the simulated feedback is worth anything — 15 people would use it.
A sample report I can read before paying anything, to judge whether the simulated findings are actually actionable
- The sample report, just to see what the output looks like — 11 people would use it.
A quick read of one page of my site — if the sample ever actually showed me what that looks like.
- The promise to re-run the same 30 people after a fix — caught 37 people's eye
- The sample report — I want to see what the output actually looks like. — caught 37 people's eye
- The promise that their pre-study guesses are graded against the blind results — caught 28 people's eye
Getting there works: 310 of 360 did something. The loss is after that point, not before it.
Three visits, step by step
Counts say how often; a visit says what it was like. These three are chosen for what they illustrate, and every line is what the participant actually did and said, in order.
Someone who gave up where most people did
Product people at software companies · STEM, Under 5, Latin America, Very high, Balanced · left deciding it was not for them · would use: Read a sample report — I want to see the actual output before giving them any si
Left on /sample/report.html after raising the top finding — the most common way a visit ended.
- / — clicked ‘Read a sample report’ curious
Before I trust anything, I want to see what a report actually looks like — that's the fastest way to tell if this is real or hand-waving.
⚠ The small text ('Usability studies with sim', 'The gap') is hard to read on a phone
⚠ 'ProductTex' and 'Menu' are tiny tap targets on mobile - /sample — clicked ‘Open the sample report’ curious
I came to see the actual report, and this big blue button is exactly that — no guessing needed.
⚠ Three paragraphs of story before I can reach anything I actually came for
⚠ 'Quick read' still never explained — is that a report, a scan, a call? - /sample/report.html — scrolled skeptical
24 of 30 found something they'd use but 29 of 30 decided it wasn't for them — that math doesn't work, and I want one more swipe to see if there's an actual report or price down there.
⚠ 24 of 30 and 29 of 30 can't both be true of 30 people — I stopped trusting the numbers
⚠ 60 simulated visits vs 30 participants — which is it? - /sample/report.html — scrolled skeptical
If the sample report is the selling point, I need to see whether it gets less vague lower down.
⚠ 60 simulated visits vs 30 participants — which is it?
⚠ The headline '24 of 30 found something they'd use' immediately collides with 'Nobody got the payoff', so I can't tell if this result is good or bad - /sample/report.html — scrolled restless
I still have not seen a single chart or example of the actual report format, so one more scroll before I give up.
⚠ Endless paragraphs with no charts, screenshots or sample report layout on mobile
⚠ Contradictory signal between '24 of 30 found something they'd use' and 'nobody got the payoff' - /sample/report.html — scrolled restless
Another wall of prose — I'll keep going a bit in case a table, chart or actual report layout finally shows up.
⚠ Endless paragraphs on a phone, no visual break anywhere
⚠ 'We say this once here and once in the appendix' — defensive phrasing that made me doubt the rest - /sample/report.html — scrolled restless
Six scrolls in and I still haven't seen an actual finding — just table fragments with unexplained numbers, so I'll push a bit further for the payoff.
⚠ 18,676 and 2,253,290 still have no units — visitors? sessions? over what period?
⚠ The two numbers sit in columns with no header on this slice, so I can't tell which is which - /sample/report.html — scrolled restless
I still can't tell what these numbers mean; I need to see if headers appear lower down.
⚠ Table columns still lack headers on this slice
⚠ Control row bigger than audience row makes the comparison feel backwards - /sample/report.html — scrolled restless
The heading promised where they went — the content better actually be under it this time.
⚠ Heading 'Where Your customers actually went, and what they did' appears with nothing under it yet — just more scrolling
⚠ Tiny grey text like '60 simulated visits' and 'Participants' is below readable size on a phone - /sample/report.html — scrolled restless
This paragraph is describing the format of the ratings, not showing me one — I want to actually see a rated problem with its evidence.
⚠ Heading 'Where Your customers actually went, and what they did' has a diagram I can't make sense of, and no stated milestone
⚠ Tiny grey text ('60 simulated visits', 'Participants') is below readable size on a phone - /sample/report.html — clicked ‘10 BLOCKER Raw JSON error on screen…’ impatient
The page keeps saying 'open a row for the evidence' — fine, I'll open the worst one and see if there's actually anything behind it.
⚠ 'Findings in this report were not checked against screenshots. Verification was switched off when it was built.' — so how do I trust any of this?
⚠ The numbers in front of each severity (10, 12, 9, 6) are never explained — is that people, percentage, score? - /sample/report.html — clicked ‘10 BLOCKER Raw JSON error on screen…’ restless
I've read a whole screen of format promises — I'm opening this blocker row to finally see actual evidence.
⚠ The numbers before each severity (10, 12, 9, 6) are still never explained — people? percent? score?
⚠ No stated goal or success criteria, so calling something a BLOCKER means nothing - /sample/report.html — clicked ‘8 MAJOR The nav links are tiny and …’ restless
I want to see whether opening a non-blocker row actually shows the evidence format they promised.
⚠ No stated product or goal, so 'BLOCKER' has no meaning
⚠ Findings not checked against screenshots — why should I trust this? - /sample/report.html — clicked ‘6 BLOCKER Raw JSON error instead of…’ skeptical
Let’s open the 6 BLOCKER row and see if the evidence actually exists.
⚠ No stated product or goal, so 'BLOCKER' has no meaning
⚠ Findings not checked against screenshots — why should I trust this? - /sample/report.html — clicked ‘9 BLOCKER Still no sample report on…’ skeptical
I'll tap the 'Still no sample report' row to see if it finally opens some actual evidence.
⚠ Findings not checked against screenshots — why should I trust this?
⚠ Tiny grey labels and controls are hard to read on a phone - /sample/report.html — scrolled impatient
I keep getting the same complaint list — let me push down and find one row that's actually open with its evidence filled in.
⚠ None of the rows are open by default even though the page tells me to open one — I shouldn't have to hunt
⚠ The 'not checked against screenshots' note sits right next to findings that are clearly the same JSON bug counted twice - /sample/report.html — clicked ‘6 BLOCKER Raw JSON error instead of…’ impatient
Maybe this row finally opens into the fix and verification I've been looking for.
⚠ Tiny grey labels and severity chips are hard to read and tap on a phone
⚠ Rows aren't open by default, so I have to hunt for the detail - /sample/report.html — scrolled impatient
I’ll keep going down; maybe there’s one complete example further on.
⚠ 'No capture of this moment was kept for this group' sounds like an excuse for missing evidence
⚠ The 6 BLOCKER row is repeating and I already tried it - /sample/report.html — clicked ‘9 MAJOR small text and fiddly top m…’ hopeful-cautious
I still haven't seen one full row actually filled in, and this 9 MAJOR one is right here — let me open it and see if the evidence is real this time.
⚠ The severity chips and grey labels (MAJOR, COSMETIC, MINOR) are tiny on a phone
⚠ Rows are collapsed by default, so I have to tap each one to find out if the detail is even there - /sample/report.html — scrolled weary
Only thing on screen to click is the same blocked row I already opened, so I'll keep going down to find a row that actually has its evidence filled in.
⚠ Row 9 is the only thing I can tap on this screen and it's the same one that gave me nothing
⚠ Body text and severity chips are uncomfortably small on a phone - /sample/report.html — clicked ‘5 MINOR 'Sample report' label is ti…’ weary
That one's about the sample report label I came here for — maybe this row actually has the full evidence block.
⚠ Body text and severity chips are uncomfortably small on a phone
⚠ Controls like '10' and 'BLOCKER' are fiddly to hit - /sample/report.html — clicked ‘5 MINOR 'Sample report' label is ti…’ weary
Row 5 actually shows quoted evidence about the sample report — the first row that looks like it might be filled in.
⚠ Tiny type and fiddly controls make rows hard to hit on a phone
⚠ Long list with no way to jump straight to a complete example - /sample/report.html — clicked ‘4 COSMETIC 'Pretty-print' checkbox …’ weary
Let me tap the next finding to see if row 4 also opens a full evidence block.
⚠ The complete evidence block is detached from its row and tiny to read; the list is exhausting. - /sample/report.html — scrolled weary
I still haven't found the sample report I actually came for, so I'll keep going down even though my eyes are tired.
⚠ Tiny text in the header ('ProductTex: quick read · s', '60 simulated visits', 'Participants') is barely legible on mobile
⚠ Fiddly small controls ('10', 'BLOCKER') are hard to hit while walking - /sample/report.html — clicked ‘4 COSMETIC No screenshot or video o…’ weary
This row actually mentions the report I came looking for — maybe opening it gets me closer.
⚠ Tiny header text ('ProductTex: quick read · s', '60 simulated visits') is barely readable on mobile
⚠ Header looks cut off — 'quick read · s' seems broken - /sample/report.html — scrolled weary
Still haven't seen the actual sample report, so let me keep going down and look for it.
⚠ Tiny header text ('ProductTex: quick read · s', '60 simulated visits') is barely readable on mobile
⚠ Header looks cut off — 'quick read · s' seems broken - /sample/report.html — clicked ‘12 BLOCKER Sample report link led t…’ weary
This one says the sample report link is dead — that's the exact thing I came here to see, so I want to read the details.
⚠ Still no sample report after 39% of the page — and now a finding says the link is broken
⚠ Tiny header text and fiddly severity chips are hard to tap on the go - /sample/report.html — clicked ‘12 BLOCKER Sample report link led t…’ weary
This is the blocker finding I already hit — let me open it and see if they admit the sample report itself is broken.
⚠ Tiny header text and fiddly severity chips
⚠ Long unbroken stack of findings with no contents/jump list - /sample/report.html — scrolled weary
One more flick — if the sample report is anywhere it should have shown up by now, but I'll give it one last look before I give up.
⚠ Tiny grey caption text and small severity chips are fiddly on mobile
⚠ No contents or jump list, just a long unbroken stack of findings - /sample/report.html — left done
Forty-something percent in and it is still the same unbroken stack of findings — the one thing I came for 404'd, so why keep scrolling?
⚠ No table of contents or jump list — just a long unbroken stack of items
⚠ Everything here reads the same after a while; I stopped being able to tell findings apart
Someone who tried hardest
Developers using AI coding agents · Business, 25-34, East Asia, High, Snap decisions · left still unsure · would use: The free Quick read on one of my own pages, The free Quick read on one of my own
The most typing and the longest visit in the cohort — what a motivated visitor meets.
- / — clicked ‘Get a free Quick read’ intrigued
Big blue button, first thing that looks like it does something — I'll just try the free read.
⚠ The block of text under the headline is a bit dense for a first glance, though I'll read it
⚠ Not clear how long a 'Quick read' takes or what it needs from me - /start — typed ‘alex.chen@fastmail.com’ intrigued
The page promises email is all they need, so I'll just drop mine in and see the free read.
⚠ The paragraph under the headline is still dense — three ideas crammed into two lines
⚠ No mention anywhere of how long the free Quick read actually takes or what I have to hand it - /start — clicked ‘Copy key’ pressured
Snap: copy the key before it disappears forever.
⚠ API key is shown once and cannot be recovered — one slip and it's gone
⚠ I still don't know if I need this key before registering a site or if the dashboard is enough - /start — clicked ‘Sign in on the dashboard’ pressured
I've got the key, so let's just get to the dashboard and register my site.
⚠ API key is shown once and cannot be recovered — one slip and it's gone
⚠ I still don't know if I need this key before registering a site or if the dashboard is enough - /dashboard — scrolled pressured
I want to see the rest of the new-study form before I type anything, so I don't miss a step.
⚠ The dropdown says 'Quick read — $49' while the banner says my first Quick read is free — mismatch makes me pause.
⚠ 'Start free' is still in the top nav even though I'm signed in, which is confusing. - /dashboard — clicked ‘Register site’ pressured
I'll hit Register site and get my free Quick read going.
⚠ The 'Quick read — $49' dropdown clashes with the 'first Quick read is free' banner
⚠ Start free is still in the top nav even though I'm signed in - /dashboard — typed ‘https://alexchen.dev/pricing’ nudged
The field came pre-filled with example.com and threw an error, so I just need to put my actual page in and hit Register site again.
⚠ The URL box was pre-filled with https://example.com/ and then scolded me for it — the placeholder should be a placeholder, not a value
⚠ Two error messages stacked on the same field saying basically the same thing - /dashboard — typed ‘https://alexchen.dev/pricing’ exasperated
The URL box is showing that example value again, so I'll paste my real page in full and try to get registered.
⚠ The URL field has reset to https://example.com/ again after I already typed my address — it should keep what I entered
⚠ Four credit packs are still being pushed at me before I've successfully registered a single site - /dashboard — typed ‘https://alexchen.dev/pricing’ exasperated
The address box wiped itself back to example.com again, so I'm retyping mine before anything else moves.
⚠ The URL field keeps resetting to https://example.com/ after I've typed my address twice already
⚠ Four credit packs are still being pushed at me before I've registered a single site - /dashboard — left exasperated
The URL box wipes my address every time and I'm not retyping it a fourth time — this thing is broken for me.
⚠ The URL field resets to https://example.com/ every time — three attempts, three wipes
⚠ Credit packs are pitched at me before I've registered anything
A control participant, for contrast
Control: unrelated visitors · Business, 25-34, South Asia, Moderate, Agonizes · left still unsure · would use: Reading a sample report to judge whether the findings are genuinely useful, Open
No reason to be here. What they did shows how much of the behaviour above is about your audience rather than about anyone.
- / — scrolled intrigued
I want to read the rest of the page before deciding anything, since I prefer to understand the full picture first.
⚠ Some text like 'Usability studies with sim' and 'The gap' is very small and hard to read.
⚠ Controls like 'ProductTex' and 'Docs' are small and fiddly to hit. - / — clicked ‘Sample report’ intrigued
The claim about simulated participants is bold — I want to see what an actual report from this looks like before I believe any of it.
⚠ Small, low-contrast labels like 'THE GAP' and 'HOW A STUDY RUNS' strain my eyes
⚠ The four-step section repeats each heading twice (once bold, once as a label) — sloppy - /sample — clicked ‘Open the sample report’ intrigued
I came here specifically to judge a real sample report, and this is the button that actually opens one — no reason to keep reading the preamble first.
⚠ Tiny, low-contrast 'SAMPLE REPORT' label above the headline
⚠ The intro block is a dense wall of text with the bug anecdote buried mid-paragraph - /sample/report.html — scrolled skeptical
I want to see the rest of the report before deciding if this is useful.
⚠ The summary contradicts itself: 24 of 30 found something they'd use, yet 29 decided it was not for them.
⚠ The text is dense and the important bug details are buried in a long paragraph. - /sample/report.html — scrolled skeptical
I came here to judge the report's quality, and so far it's all about a broken error page — I need to see what's further down before I decide.
⚠ The report's headline finding is a bug I just hit myself — clicking 'Open the sample report' — which is either honest or a bad joke.
⚠ Dense paragraphs where the important count details are buried mid-sentence. - /sample/report.html — scrolled skeptical
The methodology section just ended and the actual findings are further down — I want to read those before judging anything.
⚠ Numbers like 18,676 and 2,253,290 sit far right of dense wrapped text, so I had to hunt for what they even refer to.
⚠ Small, low-contrast labels ('60 simulated visits', 'Participants') are genuinely hard to read. - /sample/report.html — scrolled restless
The method text is readable but the RESULTS table is just starting — the actual findings, and hopefully some sample output, are further down.
⚠ A 'BLOCKER' label and 'Raw JSON er' text are showing through as raw artifacts — that's broken rendering on what should be a polished study page.
⚠ The 2,253,290 number is stranded alone above the paragraph with no header, so I don't know if it's the control pool size or total actions. - /sample/report.html — scrolled restless
The numbers are here but the actual chart is half cut off, and I still haven't seen a single line of the report this whole page is supposedly selling.
⚠ The '10 BLOCKER Raw JSON er' artifacts are still bleeding through and look broken.
⚠ Labels like '60 simulated visits', 'Participants' and the 'ProductTex: quick read · s' string are painfully small for this bandwidth. - /sample/report.html — clicked ‘10 BLOCKER Raw JSON error on screen…’ restless
The page keeps promising evidence under each row — time to open the worst-rated one and see if the findings are real or just filler.
⚠ The report admits findings were never checked against screenshots — that undercuts every number on this page.
⚠ Text artifacts still reading 'Raw JSON er' look like broken copy-paste, not a product. - /sample/report.html — clicked ‘10 BLOCKER Raw JSON error on screen…’ skeptical
This is the one thing I said I'd do — open a single finding end-to-end and see if the write-up is actually usable.
⚠ The 'not checked against screenshots' admission quietly invalidates every number above it.
⚠ Half-visible text artifacts like 'Raw JSON er' look like broken copy-paste. - /sample/report.html — clicked ‘9 BLOCKER Still no sample report on…’ skeptical
I want to open the finding about no sample report, since that is exactly what I am missing.
⚠ Dense caveat prose above the findings makes me hold too much at once.
⚠ The 'not checked against screenshots' admission quietly undermines every number above it. - /sample/report.html — clicked ‘10 BLOCKER Raw JSON error on screen…’ skeptical
If these write-ups are the actual deliverable, I need to open one blocker and see whether the evidence holds up.
⚠ Tiny labels like '60 simulated visits' and 'Participants' are painful to read on this connection.
⚠ Dense caveat prose above the findings makes me hold too much at once. - /sample/report.html — scrolled restless
I've already poked that blocker row and got nothing new, so let me push further down and see if an actual report or evidence view ever shows up.
⚠ The 'not checked against screenshots' line sits right above the numbers and quietly poisons them.
⚠ Tiny grey labels ('60 simulated visits', 'Participants') are genuinely hard to read on this connection. - /sample/report.html — scrolled restless
This finding actually has a full WHERE/THEN/DO/CHECK structure — closest thing to a real write-up I've seen, so let me keep reading down before I judge it.
⚠ The raw JSON error screenshot inside the finding is a tiny unreadable blob — I can't actually tell what the participant saw.
⚠ Tiny grey labels ('ProductTex: quick read · s', 'Participants') are still hard to read on this connection. - /sample/report.html — clicked ‘6 BLOCKER Raw JSON error instead of…’ restless
I have already opened the 9 and 10 blockers, so let me pull open the other blocker to see if these write-ups hold up as a real report artifact.
⚠ Still no sample report anywhere on this page, only findings about it.
⚠ Tiny grey labels ('ProductTex: quick read · s') are hard to read on this connection. - /sample/report.html — scrolled restless
I've read finding 9's whole block twice now and there's still no actual sample report — let me see what the remaining twenty-odd findings down there are hiding.
⚠ Finding 9's DO line tells me to open /sample/report.html, but that link is dead text — nothing clickable.
⚠ The only clickable things are the finding rows themselves. - /sample/report.html — scrolled restless
I still haven't seen the actual sample report I came for — maybe it's further down past these findings.
⚠ Finding 6's DO line tells me to open /sample/report.html, but that is dead grey text, not a link.
⚠ The only clickable things on this whole page are the finding headers themselves. - /sample/report.html — scrolled restless
I'm a third of the way down and I still haven't seen the actual sample report, so let me keep going.
⚠ 36% down a very long page and the deliverable still hasn't shown up.
⚠ Tiny 'ProductTex: quick read · s' labels are hard to read and slow to load. - /sample/report.html — scrolled restless
Forty percent down and still no deliverable — one more push before I decide this page is just an endless list.
⚠ 40% down a very long page and the deliverable has still not appeared.
⚠ Every finding repeats the same two themes — tiny text, fiddly mobile nav. - /sample/report.html — clicked ‘12 BLOCKER Sample report link led t…’ restless
This is the one finding about the sample report link I actually came here for — clicking it might finally show me the evidence trail or the real deliverable.
⚠ Only finding headers are clickable — zero navigation, no way back to the top.
⚠ Every finding repeats the same two themes: tiny text and fiddly mobile nav. - /sample/report.html — scrolled restless
I came here for the sample report and I am still not seeing it — one more push down before I give up on this.
⚠ 44% down the page and still no deliverable — the sample report.
⚠ Identical themes repeating (tiny text, fiddly mobile targets) across findings. - /sample/report.html — clicked ‘4 MINOR 'SAMPLE REPORT' label is ti…’ restless
That 'SAMPLE REPORT' label finding is the closest thing to the deliverable I've been chasing — maybe the evidence image actually shows the report.
⚠ Every finding row is a tiny, cramped label/badge cluster that's hard to parse at this size.
⚠ Repeated themes (small text, fiddly mobile targets) across findings make the list feel padded. - /sample/report.html — scrolled restless
I want to see the last two findings before I commit to opening one.
⚠ Every finding row is a tiny, cramped label/badge cluster that's hard to parse.
⚠ Repeated themes (small text, fiddly mobile targets) make the list feel padded. - /sample/report.html — scrolled restless
I still haven't seen one finding fully worked through, and this 'what is working' section isn't it.
⚠ Severity labels render as cramped micro-controls that are hard to parse.
⚠ Still no sign of the full worked example after 49% of the page. - /sample/report.html — scrolled restless
Still no filled-in example anywhere, and these visit walkthroughs look like the closest thing to one — I'll keep going.
⚠ 53% down and the worked example I came for still hasn't appeared.
⚠ The case study I'm reading is about a link that returned raw JSON — so the 'most common way a visit ended' is a dead end I already hit myself. - /sample/report.html — scrolled restless
The four steps and any real screenshot of the report have to be further down — I'll finish the page before I decide anything.
⚠ Tiny type on 'ProductTex: quick read · s', '60 simulated visits', 'Participants' — squinting on a phone.
⚠ Small fiddly tap targets on '10 BLOCKER Raw JSON er' and '10'. - /sample/report.html — scrolled restless
There's still more page below and I haven't found a single visual of the report yet — let me see what else is down here.
⚠ Small type on 'ProductTex: quick read', '60 simulated visits', 'Participants' — squinting on a phone.
⚠ Fiddly tap targets on the '10 BLOCKER' rows. - /sample/report.html — scrolled restless
I've come this far — I'll check what's left below before I write the whole thing off.
⚠ Still no image, video, or filled-in example anywhere on this page.
⚠ Sample report link is a dead end that dumps raw JSON. - /sample/report.html — scrolled restless
The blocker fix table is the closest thing yet to the template I actually came for — let me see if the rest of the page shows a filled-in finding.
⚠ The 'sample report' — the one artefact I came for — is a raw {"error": "not_found"} page with a Pretty-print checkbox.
⚠ The table's small text and cramped BLOCKER cell are hard to read on a phone. - /sample/report.html — left resigned
Three near-identical BLOCKER rows in a row — the page is repeating itself and the one example I came for is still a 404, so there's nothing new down there for me.
⚠ Identical BLOCKER rows repeated three times — reads like a rendering bug, not a fix list.
⚠ The 'sample report' the whole product points to is a raw JSON error with a Pretty-print checkbox.
What to change, in order
Ordered by severity, then by how many visits the problem ended. Effort is our estimate of the change as described — S is a copy or style change, M a screen change, L a flow change. Each is verifiable: we re-run the same participants on the changed page, and the check names the number in your own analytics that should move.
| # | Severity | Change | Screen | Effort | Raised / left | How to verify |
|---|---|---|---|---|---|---|
| 1 | Blocker | On /sample/report.html: raise the contrast and size of ‘60 simulated visits’. The audit records the measured ratio; 4.5:1 and 14px are the floor. Tiny text ('ProductTex: quick read · s', '60 simulated visits', 'Participants') | /sample/report.html | S | 50 / 17 | Re-run with a CSS override on ‘60 simulated visits’; the contrast audit must report zero failures on it before the re-run counts. |
| 2 | Major | On /sample/report.html: raise the contrast and size of ‘60 simulated visits’. The audit records the measured ratio; 4.5:1 and 14px are the floor. Still tiny text: 'ProductTex: quick read · s', 'Participants', '60 simulated visits'. | /sample/report.html | S | 50 / 14 | Re-run with a CSS override on ‘60 simulated visits’; the contrast audit must report zero failures on it before the re-run counts. |
| 3 | Major | On /sample/report.html: raise the contrast and size of ‘10’. The audit records the measured ratio; 4.5:1 and 14px are the floor. Small text and fiddly controls | /sample/report.html | S | 27 / 6 | Re-run with a CSS override on ‘10’; the contrast audit must report zero failures on it before the re-run counts. |
| 4 | Major | On /sample/report.html: raise the contrast and size of ‘10’. The audit records the measured ratio; 4.5:1 and 14px are the floor. Tiny text and fiddly tap targets on a phone | /sample/report.html | S | 24 / 5 | Re-run with a CSS override on ‘10’; the contrast audit must report zero failures on it before the re-run counts. |
| 5 | Major | On /sample/report.html: raise the contrast and size of ‘10’. The audit records the measured ratio; 4.5:1 and 14px are the floor. Tiny text and fiddly small controls | /sample/report.html | S | 35 / 10 | Re-run with a CSS override on ‘10’; the contrast audit must report zero failures on it before the re-run counts. |
| 6 | Major | On /sample/report.html: raise the contrast and size of ‘60 simulated visits’. The audit records the measured ratio; 4.5:1 and 14px are the floor. Small grey labels and tiny text are hard to read | /sample/report.html | S | 33 / 8 | Re-run with a CSS override on ‘60 simulated visits’; the contrast audit must report zero failures on it before the re-run counts. |
| 7 | Minor | On /: raise the contrast and size of ‘Usability studies with sim’. The audit records the measured ratio; 4.5:1 and 14px are the floor. Tiny text like 'Usability studies with sim' | / | S | 25 / 4 | Re-run with a CSS override on ‘Usability studies with sim’; the contrast audit must report zero failures on it before the re-run counts. |
| 8 | Minor | On /sample: raise the contrast and size of ‘SAMPLE REPORT’. The audit records the measured ratio; 4.5:1 and 14px are the floor. Tiny 'Sample report' label | /sample | S | 28 / 2 | Re-run with a CSS override on ‘SAMPLE REPORT’; the contrast audit must report zero failures on it before the re-run counts. |
| 9 | Minor | On /sample: raise the contrast and size of ‘Sample report’. The audit records the measured ratio; 4.5:1 and 14px are the floor. The 'Sample report' label is tiny | /sample | S | 28 / 1 | Re-run with a CSS override on ‘Sample report’; the contrast audit must report zero failures on it before the re-run counts. |
| 10 | Minor | On /sample: raise the contrast and size of ‘Docs’. The audit records the measured ratio; 4.5:1 and 14px are the floor. Some nav links and small labels are tiny and fiddly. | /sample | S | 23 / 2 | Re-run with a CSS override on ‘Docs’; the contrast audit must report zero failures on it before the re-run counts. |
What this study cannot tell you
Read these at the same weight as everything above
- Participants are simulated. Rates are for comparing versions of a page on the same participants, never a forecast of your conversion.
- Load time and layout shift are invisible to participants; page speed is measured separately.
- Every participant is a first-time visitor.
- The milestone we set (the sign-up confirmation) also appears as text inside the sample report these participants read, so it matched there; the funnel stops at typed into it.
- These are simulated visits. They reproduce the ordering of problems and the words people use for them; they do not predict what share of your real visitors will do anything. Any rate here is for comparing one version of a page against another on the same participants, never for forecasting.
Instrument health
Whether our participants could actually operate your site. A study where they could not reads exactly like a study of a broken product, so this is graded before the report is built and printed here. clean
0
of 360 sessions errored
4
of 3,350 actions threw
0%
of clicks needed forcing
9
sessions cut off by the step budget
2
sessions met a control we could not drive
1
sessions saw a change with no structural trace
360 of 360
sessions could measure page speed
Controls this instrument could not drive, by kind: input ×2. These are our blindness and never appear as findings.
Controls that changed on screen with no change to the accessibility tree: button ×1, input ×1. A screen-reader user cannot see these respond either.
Appendix — how this was made
Everything below is read from the study's own record. Where a value was not recorded, it says so.
Who was tested
People are drawn from a pool of 10,000,000 simulated personas we generated ourselves (seed 20260813) and calibrated on age, region, urbanicity and gender, against UN targets — max deviation 0.02 pp. Not calibrated: the other 1,286 attributes, including the cognitive traits that drive browsing behaviour (patience, attention span, decision speed). The pool is not a representative sample of any population and this report never claims one.
A cohort is a plain-English description resolved once into explicit filters on the pool's attributes, frozen, and printed under Who we tested, so a re-run months later selects the same way. Draws are deterministic for the study seed: the same people can be re-run on a changed page. A cohort below 400 matching people is refused rather than drawn with repeats.
- Product people at software companies — pool 156,015, effective sample 156,015; arrival:
you saw ProductTex mentioned in a newsletter for software founders and clicked the link
- Developers using AI coding agents — pool 1,947, effective sample 1,947; arrival:
you saw ProductTex mentioned in a newsletter for software founders and clicked the link
- Control: unrelated visitors — pool 2,253,290, effective sample 2,253,290; arrival: cold (no stated reason to be here — this understates engagement)
A negative-control cohort with no plausible use for the product ran in the same study, same seed, same harness. A third of people with no use for a product still click its primary button, so a rate against zero is uninterpretable; the finding is the difference.
Everyone drawn reads the product's language at least at an intermediate level (site language: en). Audience: free — which tier of the product they were allowed to meet.
What each persona saw and could do
A real headless browser at 1280×800, on the live site, one fresh browser per person. Each step the persona receives: the page's interactive controls as the accessibility tree names them, each checked to be actually clickable at its own centre; a spatial description of where things sit and what is below the fold; a measured audit of contrast, text size and tap targets; and a screenshot of the screen, with no instruction to judge its appearance (asking for a verdict turns a user into a reviewer). They can click, type, choose from a dropdown, scroll, or stop.
Mode: roam. In roam nobody is given a task: they are told they have access to a product and left to decide whether any of it is for them. A goal, when one is set, states what the person wants and never what they will find.
Reading. Each participant could read only the text on their screen and was told how many lines lay below it; scrolling was how they read more, as it is for a person.
Devices. Each participant's browser was sized from their own device context: a 390×844 phone with a mobile user agent and touch for people who browse on the go, 1280×800 otherwise. Voice assistants, accessibility tools and low bandwidth are not modelled beyond the viewport.
Step budget: 30 actions per visit. A session that used every step without choosing to leave is counted as cut off by us, not as finished, and appears under instrument health.
What is carried between steps: what they now think the product is, whether it is for them, what has irritated them, what they would use, and a one-word mood. It is fed back so they can change their mind or lose patience. Nothing is fed back that reads as a verdict on the product.
Models
Reasoning and action: deepseek-flash. Seeing the screen: deepseek-flash. Study seed: 20261101. Different eyes produce different sessions, so a number here is comparable only to a study stamped with the same pair.
Model calls recorded: 3,351.
How things were counted
Two runs, different people. 2 run(s) of 60 people per cohort. A finding that appears in every run within 10 points holds; one that appears in every run but moves more is shaky; one absent from a run is marked so. One run is an anecdote with a sample size.
Themes are clustered arithmetically, not by a model: TF-IDF over every phrasing, average-linkage merging at threshold 0.24, the same answer every time. A model may merge finished cluster labels by meaning; it never forms the clusters. Counts are of people who raised a thing, not of mentions — one person saying it six times is one.
Rates. Below 60 people per run, counts are printed and percentages are not, because rates do not converge below that. Below 120, judgement-shaped rates ("not for me", "stopped early") are printed as ranges: they moved up to 27 points between identical runs at n=30 and 2 points at n=120.
Ranking. severity = (people who raised it + 2 × people who left within a step of raising it) × 1.0 if it held across runs else 0.6 × 1.0 for a defect, 0.4 for a taste remark. "Left" means the simulated visit ended on the screen where the complaint was made, or on the next one with nothing further said; it is a count of sessions, not a prediction about your visitors.
Severity scale. Decided from the sessions, not by judgement. Blocker: a defect, and at least a third of the people who met it (three or more) left there. Major: a defect, and either five or more left there or a quarter of the group raised it. Minor: a defect raised by fewer, with few or no exits. Cosmetic: an impression or taste remark with no exits attributable. Effort (S / M / L) is our estimate of the change as described: copy or style, one screen, or a flow.
Verification. Findings were not checked against screenshots for this report.
Definitions. Every derived number travels with a hash of the rule that produced it, so two studies are compared only on the same rule: funnel stages b9f4387a; appetite 2cb41a8a.
Instrument
360 sessions, 3,350 steps. 0 sessions errored; 4 actions threw; 0% of clicks needed forcing past something covering them; 9 sessions were cut off by the step budget; 2 met a control this instrument could not drive. The full table and verdict are under Instrument health above.
Limits of the method
- Simulated visitors click roughly ten times more than real ones and act rather than read. Ordering and themes transfer; absolute rates do not.
- Every persona is a first-time visitor. Returning-user behaviour is not modelled.
- Personas receive a settled page. Load time, layout shift and animation are invisible to them; speed is measured separately and reported, never fed to the persona.
- Colour, contrast and size come from a measured audit of the page, not from the persona's impression. Impression, attention and meaning come from the screenshot.
- A control this instrument cannot drive — a canvas, a drag handle, a custom widget — is recorded as our blindness, never as a finding about the product.
Every rule above was set by a study we ran on our own instrument before it was used on a customer's product; the measurement log behind them is available on request.
Appendix — every participant
360 visits. Who they were, where they went in order, what they typed, how they left, what they would use, and the first thing that got in their way. Sorted by group, then run, then participant.
Open the table
| Group | # | Device | Who | Path | Acts | Typed | Left | Would use | First friction |
|---|---|---|---|---|---|---|---|---|---|
| Control: unrelated visitors | r1·0 | desktop | Social science, 18-24, None, Snap decisions | / → /sample → /sample/report.html | 11 | 0 | left · not for me | Read a sample report, Reading an actual sample report end to end | No video, just a wall of text |
| Control: unrelated visitors | r1·1 | desktop | Business, 18-24, Moderate, Quick | / → /start → /sample → /sample/report.html | 12 | 0 | left · not for me | Get a free Quick read (if I had a site to test), Sample report, to see | Lots of small text and a wall of copy — the headline is clear but everything after it is d |
| Control: unrelated visitors | r1·2 | desktop | STEM, 65-74, Moderate, Quick | / → /sample | 2 | 0 | left · not for me | No video, just a lot of text. | |
| Control: unrelated visitors | r1·3 | mobile | Arts, 35-44, Moderate, Quick | / → /start → /dashboard | 7 | 1 | left · not for me | The free Quick read on my own site | That tiny 'USABILITY STUDIES WITH SIMULATED PARTICIPANTS' line is hard to read on a phone |
| Control: unrelated visitors | r1·4 | desktop | Humanities, 35-44, Moderate, Balanced | / → /sample → /sample/report.html | 11 | 0 | left · not for me | Read a sample report, Open the sample report | Two near-identical calls to action — “Start free” and “Get a free Quick read” — I can't te |
| Control: unrelated visitors | r1·5 | mobile | Interdisciplinary, 35-44, Moderate, Quick | / → /start | 2 | 0 | left · not for me | Read a sample report | The top nav text is tiny. |
| Control: unrelated visitors | r1·6 | mobile | STEM, 35-44, Moderate, Quick | / → /sample → /sample/report.html | 12 | 0 | left · not for me | Read a sample report, Open the sample report to see the actual output | The text is small and there's a lot of it, which is hard on a phone. |
| Control: unrelated visitors | r1·7 | desktop | Social science, 25-34, High, Agonizes | / | 2 | 0 | left · not for me | It's a wall of text and I'd rather watch a short video. | |
| Control: unrelated visitors | r1·8 | desktop | STEM, 35-44, High, Quick | / → /sample → /sample/report.html | 22 | 0 | left · not for me | The free sample report, to judge whether the simulated feedback is act | It says "Watch" but there is nothing to watch — no video, just a wall of text, and I stron |
| Control: unrelated visitors | r1·9 | mobile | STEM, 35-44, Low, Agonizes | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Read a sample report, The sample report to judge the output. | Wall of text on mobile |
| Control: unrelated visitors | r1·10 | mobile | Business, 35-44, High, Snap decisions | / → /start | 2 | 0 | left · not for me | the free quick read if I had a site to test | No video, just a wall of text which puts me off |
| Control: unrelated visitors | r1·11 | desktop | STEM, 45-54, Moderate, Quick | / → /sample → /sample/report.html | 4 | 0 | left · not for me | the sample report — to see if the output is actually useful or just ge | Big wall of text with no video — for something about watching people, there's nothing to w |
| Control: unrelated visitors | r1·12 | desktop | Humanities, 45-54, Moderate, Balanced | / → /sample → /sample/report.html | 4 | 0 | left · not for me | The sample report itself, just to see what one of these actually looks | No video, only paragraphs — I prefer watching. |
| Control: unrelated visitors | r1·13 | desktop | Interdisciplinary, 45-54, Moderate, Quick | / → /sample → /sample/report.html | 3 | 0 | left · not for me | The free Quick read on one of my own pages, just to see if the report | No video anywhere — and video is how I actually absorb a new tool |
| Control: unrelated visitors | r1·14 | desktop | Social science, 55-64, Moderate, Deliberate | / → /sample → /sample/report.html | 19 | 0 | left · not for me | Read a sample report, Reading the sample report to judge whether the o | A wall of text with no video up front; that puts me off. |
| Control: unrelated visitors | r1·15 | desktop | STEM, 35-44, High, Quick | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Read a sample report, Open the sample report | No video, just a wall of text and jargon like 'simulated participants'. |
| Control: unrelated visitors | r1·16 | desktop | Business, 45-54, None, Snap decisions | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Read a sample report, Read the sample report | The long paragraph of text slows me down; I nearly skipped it. |
| Control: unrelated visitors | r1·17 | desktop | Humanities, 55-64, Low, Agonizes | / → /sample → /sample/report.html | 3 | 0 | left · not for me | Reading a sample report, to understand what the output looks like, Rea | The opening paragraph is a heavy block of text for someone who skims |
| Control: unrelated visitors | r1·18 | desktop | Social science, 55-64, High, Snap decisions | / → /sample → /sample/report.html | 9 | 0 | left · not for me | the sample report, to see if the simulated feedback is believable, the | 'Quick read' is used twice before anyone explains what it is |
| Control: unrelated visitors | r1·19 | desktop | Humanities, 45-54, Low, Deliberate | / → /sample → /sample/report.html | 5 | 0 | left · not for me | the sample report, to judge whether the output is specific or generic, | The top navigation is crowded and 'Docs' is easy to miss. |
| Control: unrelated visitors | r1·20 | desktop | Interdisciplinary, 25-34, None, Quick | / → /sample | 2 | 0 | left · not for me | Read a sample report | no video, long paragraphs, and small top navigation text |
| Control: unrelated visitors | r1·21 | desktop | Interdisciplinary, 18-24, Moderate, Quick | / → /sample → /sample/report.html | 4 | 0 | left · unsure | Read a sample report, Read the sample report | The tiny all-caps kicker is hard to read, and the page looks like it goes on forever. |
| Control: unrelated visitors | r1·22 | desktop | Arts, 25-34, Low, Deliberate | / → /sample → /sample/report.html | 17 | 0 | left · not for me | Read a sample report, Opening an actual sample report to see if the ou | The tiny eyebrow text 'USABILITY STUDIES WITH SIMULATED PARTICIPANTS' is nearly unreadable |
| Control: unrelated visitors | r1·23 | mobile | STEM, 45-54, Moderate, Balanced | / → /sample → /sample/report.html | 8 | 0 | left · not for me | Read a sample report, Open the sample report | No video on the first screen despite saying 'Watch'. |
| Control: unrelated visitors | r1·24 | desktop | Arts, 25-34, Low, Snap decisions | / | 1 | 0 | left · not for me | No video, just a wall of text | |
| Control: unrelated visitors | r1·25 | mobile | Interdisciplinary, 35-44, High, Balanced | / → /sample → /sample/report.html | 4 | 0 | left · not for me | The page is text-heavy and still doesn’t say what it costs. | |
| Control: unrelated visitors | r1·26 | mobile | Business, 25-34, Moderate, Snap decisions | / → /start → /dashboard | 8 | 1 | left · not for me | Get a free Quick read, The free Quick read on my own site | Way too much text on a mobile screen — give me a video or a demo instead |
| Control: unrelated visitors | r1·27 | mobile | Social science, 35-44, Low, Agonizes | / | 1 | 0 | left · not for me | Párrafo demasiado largo para leer en el móvil | |
| Control: unrelated visitors | r1·28 | mobile | Humanities, 25-34, Low, Balanced | / → /sample → /sample/report.html | 3 | 0 | left · not for me | Wall of text on mobile, no video on the first screen | |
| Control: unrelated visitors | r1·29 | desktop | Business, 18-24, Low, Balanced | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Read a sample report, Open the sample report to see what the output ac | Lots of text to read |
| Control: unrelated visitors | r1·30 | mobile | Business, 45-54, Moderate, Balanced | / → /pricing | 8 | 0 | left · not for me | A free Quick read, if I had a site to test | The tiny top-left and top-right controls are fiddly on mobile. |
| Control: unrelated visitors | r1·31 | mobile | STEM, 45-54, Moderate, Snap decisions | / → /start | 2 | 0 | left · not for me | The small text at the top is hard to read on mobile. | |
| Control: unrelated visitors | r1·32 | mobile | STEM, 35-44, Very high, Snap decisions | / → /sample → /sample/report.html | 7 | 0 | left · not for me | The page is a wall of text with no video, and some of the smaller text is hard to read on | |
| Control: unrelated visitors | r1·33 | desktop | Business, 55-64, Low, Agonizes | / | 1 | 0 | left · not for me | The top menu text is small and fiddly. | |
| Control: unrelated visitors | r1·34 | desktop | STEM, 25-34, High, Agonizes | / → /sample → /sample/report.html | 25 | 0 | left · not for me | Reading a sample report to sanity-check the output quality, Reading th | 'Read in their own words why' sits oddly next to 'participants are simulated, not real peo |
| Control: unrelated visitors | r1·35 | desktop | Interdisciplinary, 35-44, Low, Agonizes | / | 1 | 0 | left · not for me | The page is a wall of text with no video. | |
| Control: unrelated visitors | r1·36 | desktop | STEM, 25-34, Moderate, Snap decisions | / → /sample → /sample/report.html | 23 | 0 | left · not for me | The sample report, The sample report itself, to judge whether the outp | The paragraph under the headline is dense — a lot of text packed into small type for a her |
| Control: unrelated visitors | r1·37 | desktop | Arts, 45-54, Moderate, Balanced | / | 1 | 0 | left · not for me | The simulated participants feel untrustworthy | |
| Control: unrelated visitors | r1·38 | desktop | Business, 35-44, Moderate, Deliberate | / → /sample → /sample/report.html | 10 | 0 | left · not for me | Read a sample report | Some text renders very small — the section labels like 'THE GAP' and 'HOW A STUDY RUNS' ar |
| Control: unrelated visitors | r1·39 | desktop | Humanities, 55-64, Low, Balanced | / → /sample → /sample/report.html | 4 | 0 | left · not for me | The small all-caps labels and tiny nav links are hard to read. | |
| Control: unrelated visitors | r1·40 | mobile | Interdisciplinary, 18-24, Moderate, Snap decisions | / → /start | 2 | 0 | left · not for me | Read a sample report, just to see what the output is | Headline says 'Watch' but there's no video anywhere — just a wall of text on a phone |
| Control: unrelated visitors | r1·41 | desktop | STEM, 35-44, Moderate, Agonizes | / → /sample → /sample/report.html | 18 | 0 | left · not for me | Probably the free Quick read if I had a site to test., Possibly the Sa | The phrase 'simulated participants' is repeated, but I still wonder how reliable the simul |
| Control: unrelated visitors | r1·42 | mobile | Humanities, 35-44, Moderate, Quick | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Read a sample report, Open the sample report | No video on the first screen despite the word 'Watch' |
| Control: unrelated visitors | r1·43 | mobile | STEM, 65-74, Moderate, Agonizes | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Read a sample report, Open the sample report | No video or visual example on the first screen |
| Control: unrelated visitors | r1·44 | mobile | Business, 18-24, Moderate, Quick | / | 1 | 0 | left · not for me | Wall of text on the first screen | |
| Control: unrelated visitors | r1·45 | mobile | Arts, 25-34, Moderate, Quick | / → /sample → /sample/report.html | 27 | 0 | left · not for me | Read a sample report, The full real-result write-up once I can read de | “Simulated participants” in caps up top but the truth-bomb that they aren't real people on |
| Control: unrelated visitors | r1·46 | mobile | STEM, 25-34, Moderate, Deliberate | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Read a sample report, Read a sample report to judge the output quality | The paragraph is a wall of text on mobile; I only skimmed the headline and buttons. |
| Control: unrelated visitors | r1·47 | mobile | STEM, 55-64, Moderate, Deliberate | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Reading a sample report to judge whether the simulated feedback is cre | Wall of text with no video anywhere — that is exactly what puts me off |
| Control: unrelated visitors | r1·48 | desktop | STEM, 65-74, High, Snap decisions | / → /start → /sample → /sample/report.html | 8 | 0 | left · not for me | Read a sample report, Open and read the sample report. | The line about participants being simulated makes me doubt whether the findings mean anyth |
| Control: unrelated visitors | r1·49 | desktop | Business, 35-44, High, Deliberate | / → /sample → /sample/report.html | 28 | 0 | left · not for me | Reading a sample report to judge whether simulated feedback is credibl | The free thing requires 'a site you control' — that quietly excludes anyone who does not o |
| Control: unrelated visitors | r1·50 | desktop | STEM, 35-44, Low, Snap decisions | / | 1 | 0 | left · not for me | Long paragraphs of text make me tune out fast. | |
| Control: unrelated visitors | r1·51 | desktop | Business, 35-44, Low, Deliberate | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Read a sample report, Open the sample report | The page is a wall of text and there is no video, which puts me off. |
| Control: unrelated visitors | r1·52 | desktop | Social science, 18-24, High, Snap decisions | / → /sample → /sample/report.html | 14 | 0 | left · unsure | Read a sample report, Open the sample report | The tiny label 'Usability studies with simulated participants' is easy to miss, and 'simul |
| Control: unrelated visitors | r1·53 | mobile | Social science, 55-64, Moderate, Quick | / → /sample → /sample/report.html | 26 | 0 | left · not for me | The sample report, to decide whether the simulated-visitor output is w | Tiny 'ProductTex' and 'Menu' links are fiddly to tap on a phone |
| Control: unrelated visitors | r1·54 | desktop | Interdisciplinary, 25-34, High, Snap decisions | / → /start → /dashboard → /pricing | 17 | 6 | left · not for me | The free Quick read on a site I control, One free Quick read on my own | No video anywhere — pure text walls, and I'd rather watch than read. |
| Control: unrelated visitors | r1·55 | desktop | STEM, 45-54, High, Quick | / → /sample → /sample/report.html | 16 | 0 | left · not for me | Read a sample report, Open the sample report button, to judge the outp | The small eyebrow text 'Usability studies with simulated participants' is tiny and easy to |
| Control: unrelated visitors | r1·56 | mobile | Arts, 25-34, High, Deliberate | / → /sample → /sample/report.html | 8 | 0 | left · not for me | Read a sample report, Open the sample report | Two buttons side by side with no clue which one to pick first |
| Control: unrelated visitors | r1·57 | desktop | STEM, 55-64, None, Agonizes | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Sample report, The sample report, to see if it points out real problem | The headline says 'Watch simulated visitors' but there is nothing to watch, which is a bit |
| Control: unrelated visitors | r1·58 | desktop | Business, 25-34, Moderate, Agonizes | / → /sample → /sample/report.html | 30 | 0 | left · unsure | Reading a sample report to judge whether the findings are genuinely us | Some text like 'Usability studies with sim' and 'The gap' is very small and hard to read. |
| Control: unrelated visitors | r1·59 | desktop | Arts, 35-44, Low, Quick | / → /sample → /sample/report.html | 4 | 0 | left · not for me | The sample report itself, to judge whether the output is readable or j | The headline is long and took me a moment to parse |
| Control: unrelated visitors | r2·0 | mobile | Interdisciplinary, 25-34, Moderate, Balanced | / → /sample → /sample/report.html | 5 | 0 | left · not for me | Wall of text, no video, some text is very small. | |
| Control: unrelated visitors | r2·1 | mobile | Interdisciplinary, 45-54, None, Snap decisions | / → /sample → /sample/report.html | 5 | 0 | left · not for me | Read a sample report, Reading the sample report to judge whether the o | It's a wall of text with no video or visuals. |
| Control: unrelated visitors | r2·2 | desktop | Business, 25-34, Moderate, Snap decisions | / → /sample → /sample/report.html | 21 | 0 | left · not for me | Read a sample report, Reading a full sample report before deciding | Some text is very small, like the top label and 'THE GAP'. |
| Control: unrelated visitors | r2·3 | desktop | Humanities, 65-74, Moderate, Quick | / | 2 | 0 | left · not for me | Too much text and no video on the first screen. | |
| Control: unrelated visitors | r2·4 | desktop | Business, 55-64, High, Quick | / → /sample → /sample/report.html | 24 | 0 | left · not for me | Open the sample report to see what kind of findings this tool actually | The small uppercase labels like 'USABILITY STUDIES WITH SIMULATED PARTICIPANTS' and 'THE G |
| Control: unrelated visitors | r2·5 | mobile | Humanities, 35-44, High, Quick | / → /sample → /sample/report.html | 4 | 0 | left · not for me | No video, and I prefer watching over reading long text. | |
| Control: unrelated visitors | r2·6 | mobile | Social science, 25-34, Moderate, Quick | / → /start | 2 | 0 | left · not for me | The free Quick read, if I can try it without owning a site | No video, just a wall of text |
| Control: unrelated visitors | r2·7 | desktop | Business, 55-64, Low, Balanced | / | 2 | 0 | left · not for me | Wall of text, no video, small navigation links | |
| Control: unrelated visitors | r2·8 | desktop | Humanities, 35-44, Moderate, Balanced | / → /sample → /sample/report.html | 16 | 0 | left · not for me | The sample report, to judge quality before trusting the free read., Th | The 'Usability studies with simulated participants' line above the headline is almost unre |
| Control: unrelated visitors | r2·9 | mobile | Medicine, 25-34, Moderate, Deliberate | / → /sample → /sample/report.html | 10 | 0 | left · not for me | Opening a sample report to see what the actual output looks like, A vi | Lots of text on a phone, no video, and some text is very small. |
| Control: unrelated visitors | r2·10 | mobile | Interdisciplinary, 35-44, Very high, Snap decisions | / → /start | 2 | 0 | left · not for me | Some of the small text is hard to read on mobile | |
| Control: unrelated visitors | r2·11 | mobile | Business, 35-44, Moderate, Quick | / → /sample → /sample/report.html | 4 | 0 | left · not for me | The single 'change this one thing' recommendation per report, if every | A solid block of text and no video anywhere — reading all that on my phone on the go is a |
| Control: unrelated visitors | r2·12 | desktop | Interdisciplinary, 65-74, Moderate, Quick | / → /sample → /sample/report.html | 18 | 0 | left · not for me | Reading a sample report to judge whether the simulated quotes sound be | It's all text — no video or demo of the product actually working |
| Control: unrelated visitors | r2·13 | mobile | STEM, 18-24, Low, Deliberate | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Reading a sample report to see how the findings are written., Reading | The tiny top-left logo and Menu are fiddly on mobile. |
| Control: unrelated visitors | r2·14 | mobile | STEM, 55-64, High, Snap decisions | / → /sample → /sample/report.html | 15 | 0 | left · not for me | Read a sample report, Open the sample report | Some text is very small, and the Menu button is tiny on mobile. |
| Control: unrelated visitors | r2·15 | desktop | Social science, 35-44, None, Quick | / | 1 | 0 | left · not for me | Too much text with no clear first step for someone with no goal. | |
| Control: unrelated visitors | r2·16 | desktop | Law, 35-44, Moderate, Snap decisions | / → /sample → /sample/report.html | 5 | 0 | left · not for me | Opening the sample report to see what one of these studies actually lo | It is a wall of text — no video, and I prefer watching to reading. The headline is very lo |
| Control: unrelated visitors | r2·17 | mobile | STEM, 35-44, High, Snap decisions | / → /start | 2 | 0 | left · not for me | The free Quick read on one of my pages, just to see the output | The headline reads a bit strange — 'get stuck on your live site' took me a second |
| Control: unrelated visitors | r2·18 | desktop | Business, 45-54, Moderate, Balanced | / → /sample → /sample/report.html | 30 | 0 | left · not for me | Read a sample report, Reading the full sample report end to end | The small line 'USABILITY STUDIES WITH SIMULATED PARTICIPANTS' is easy to miss and a bit h |
| Control: unrelated visitors | r2·19 | mobile | STEM, 65-74, None, Balanced | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Read a sample report | No video anywhere — just a wall of text, and the small paragraph is hard work on a phone |
| Control: unrelated visitors | r2·20 | desktop | Medicine, 25-34, Moderate, Quick | / → /sample → /sample/report.html | 3 | 0 | left · not for me | Read a sample report, Reading an actual sample report to see what the | The page is a wall of text; I’d rather watch a video. |
| Control: unrelated visitors | r2·21 | desktop | STEM, 25-34, High, Agonizes | / → /sample → /sample/report.html | 24 | 0 | left · unsure | Read a sample report, Reading a full sample report to judge whether th | The nav links are small and fiddly to hit. |
| Control: unrelated visitors | r2·22 | mobile | STEM, 45-54, Moderate, Snap decisions | / → /sample → /sample/report.html | 8 | 0 | left · not for me | Open the sample report to see if the output is easy to understand | No video on the first screen |
| Control: unrelated visitors | r2·23 | desktop | STEM, 55-64, High, Snap decisions | / → /sample → /sample/report.html | 28 | 0 | left · unsure | Read a sample report, Read the sample report | The tiny uppercase line at the top is hard to read. |
| Control: unrelated visitors | r2·24 | desktop | Business, 45-54, Low, Quick | / | 2 | 0 | left · not for me | The simulated visitors idea makes me question how useful the findings really are. | |
| Control: unrelated visitors | r2·25 | desktop | STEM, 45-54, Moderate, Agonizes | / → /sample → /sample/report.html | 11 | 0 | left · not for me | Read a sample report, Open the sample report to judge whether the outp | No video, and the page is mostly dense text. |
| Control: unrelated visitors | r2·26 | desktop | Humanities, 25-34, Moderate, Snap decisions | / → /sample → /sample/report.html | 15 | 0 | left · not for me | The sample report, to judge whether the simulated feedback reads like | 'Simulated participants' in giant uppercase at the top, but the disclaimers only come in s |
| Control: unrelated visitors | r2·27 | desktop | Interdisciplinary, 35-44, High, Agonizes | / | 1 | 0 | left · not for me | Wall of text with no video to explain quickly. | |
| Control: unrelated visitors | r2·28 | desktop | Humanities, 55-64, Moderate, Deliberate | / | 1 | 0 | left · not for me | The idea of 'simulated participants' is odd and I don’t know how much to trust it. | |
| Control: unrelated visitors | r2·29 | desktop | Social science, 35-44, Very high, Agonizes | / → /sample → /sample/report.html | 21 | 0 | left · not for me | Maybe the sample report if I needed to evaluate the approach., The sam | The line about '10 million simulated personas' and 'a control group of people with no use |
| Control: unrelated visitors | r2·30 | desktop | Interdisciplinary, 45-54, Moderate, Quick | / | 1 | 0 | left · not for me | The page is a wall of text with no video demo. | |
| Control: unrelated visitors | r2·31 | desktop | Business, 18-24, Moderate, Agonizes | / → /sample → /sample/report.html | 5 | 0 | left · not for me | The sample report, to judge whether the output is useful., The sample | The explanation paragraph is a wall of small text — I nearly skipped it. |
| Control: unrelated visitors | r2·32 | desktop | Interdisciplinary, 35-44, Low, Balanced | / → /sample → /sample/report.html | 10 | 0 | left · not for me | Read a sample report, Read the sample report end to end | The tiny labels like “Usability studies with simulated participants” and “The gap” are har |
| Control: unrelated visitors | r2·33 | desktop | Social science, 25-34, Moderate, Quick | / → /sample → /sample/report.html | 3 | 0 | left · not for me | Read a sample report | Calling them 'simulated participants' makes me doubt whether the results mean anything. |
| Control: unrelated visitors | r2·34 | mobile | Social science, 25-34, Low, Deliberate | / → /sample → /sample/report.html | 4 | 0 | left · unsure | The 'if you change one thing, change this' line — it's the only part t | Too much text on a small screen, and no video to explain it faster. |
| Control: unrelated visitors | r2·35 | desktop | Social science, 35-44, Low, Balanced | / → /sample → /sample/report.html | 3 | 0 | left · not for me | The page is text-heavy and I prefer a video demo. | |
| Control: unrelated visitors | r2·36 | desktop | Humanities, 25-34, Moderate, Agonizes | / → /sample → /sample/report.html | 19 | 0 | left · not for me | Read a sample report, to judge whether the output is credible enough t | The phrase 'simulated participants' makes me wary — if I cited this in a whitepaper it wou |
| Control: unrelated visitors | r2·37 | desktop | STEM, 35-44, Moderate, Quick | / → /sample → /sample/report.html | 12 | 0 | left · not for me | Reading the sample report itself, out of curiosity about what a simula | The page is quite text-heavy and I still don't know pricing without leaving it. |
| Control: unrelated visitors | r2·38 | mobile | Interdisciplinary, 45-54, Very high, Deliberate | / | 2 | 0 | left · not for me | No video, just a wall of text on a phone. | |
| Control: unrelated visitors | r2·39 | mobile | Humanities, 35-44, Low, Deliberate | / | 1 | 0 | left · not for me | The main CTA assumes I have a site I control, which I don't. | |
| Control: unrelated visitors | r2·40 | desktop | STEM, 35-44, Moderate, Snap decisions | / → /sample → /sample/report.html | 5 | 0 | left · not for me | Running a check on my own site once I have one live, Running a check o | The page is dense and text-heavy; I skimmed the body paragraph. |
| Control: unrelated visitors | r2·41 | desktop | STEM, 45-54, Moderate, Deliberate | / | 1 | 0 | left · not for me | Wall of text | |
| Control: unrelated visitors | r2·42 | desktop | Business, 45-54, High, Deliberate | / → /sample → /sample/report.html | 14 | 0 | left · not for me | The sample report, The sample report, if I ever get to see it | A wall of text with no video anywhere — that's how I'd rather take this in over lunch |
| Control: unrelated visitors | r2·43 | desktop | Business, 35-44, High, Balanced | / → /sample → /sample/report.html | 19 | 0 | left · not for me | Reading a sample report to judge whether the findings are concrete or | The 'simulated, not real people' caveat is buried in small grey text under the buttons — t |
| Control: unrelated visitors | r2·44 | desktop | Business, 25-34, Low, Agonizes | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Read a sample report, Open the sample report — to see if the output is | No video or interactive demo — the headline says 'Watch' but there's nothing to watch. |
| Control: unrelated visitors | r2·45 | desktop | Humanities, 25-34, Moderate, Quick | / → /sample → /sample/report.html | 28 | 0 | left · not for me | A free Quick read on one of my own pages, if the sample report looks c | No video or demo — just text, and I'd rather watch a 60-second run-through |
| Control: unrelated visitors | r2·46 | mobile | Humanities, 65-74, Moderate, Balanced | / | 1 | 0 | left · not for me | No video | |
| Control: unrelated visitors | r2·47 | mobile | Business, 35-44, Moderate, Deliberate | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Read a sample report to see if the findings are actually specific and | The small line at the top ('Usability studies with simulated participants') is tiny on a p |
| Control: unrelated visitors | r2·48 | desktop | STEM, 35-44, High, Agonizes | / → /sample → /sample/report.html | 30 | 0 | cut off · unsure | A sample report to assess whether the simulated feedback is specific e | The disclaimer about simulated participants is buried in small grey text under the buttons |
| Control: unrelated visitors | r2·49 | mobile | Interdisciplinary, 25-34, Moderate, Snap decisions | / → /sample → /sample/report.html | 18 | 0 | left · not for me | Read a sample report, The sample report | The body copy is a light gray that's harder to read on a phone than the headline |
| Control: unrelated visitors | r2·50 | desktop | Business, 25-34, High, Quick | / → /sample → /sample/report.html | 4 | 0 | left · unsure | Read a sample report, The sample report to judge whether the output is | The top nav feels cluttered for a first screen. |
| Control: unrelated visitors | r2·51 | desktop | STEM, 55-64, High, Agonizes | / → /sample → /sample/report.html | 8 | 0 | left · not for me | Read a sample report, Opening the full sample report to see whether th | No video, and the page is text-heavy |
| Control: unrelated visitors | r2·52 | desktop | Interdisciplinary, 25-34, Moderate, Deliberate | / → /sample → /sample/report.html | 30 | 0 | left · not for me | Read a sample report, Reading a full sample report end to end to judge | Some tiny labels like 'THE GAP' and 'HOW A STUDY RUNS' are easy to miss. |
| Control: unrelated visitors | r2·53 | desktop | Business, 18-24, High, Agonizes | / → /sample → /sample/report.html | 13 | 0 | left · not for me | Read a sample report, Open the sample report | Wall of text and no video — I was promised 'watch' but there's nothing to watch. |
| Control: unrelated visitors | r2·54 | mobile | Arts, 18-24, Moderate, Balanced | / → /sample → /sample/report.html | 3 | 0 | left · not for me | It's a wall of text on a phone and nothing to watch, even though the headline says 'watch' | |
| Control: unrelated visitors | r2·55 | desktop | STEM, 45-54, Low, Quick | / → /sample → /sample/report.html | 3 | 0 | left · not for me | Read a sample report, The sample report | Wall of text, no video anywhere — I'd rather watch 60 seconds than read this |
| Control: unrelated visitors | r2·56 | desktop | Interdisciplinary, 35-44, Low, Balanced | / | 1 | 0 | left · not for me | The tiny all-caps label is easy to miss and the page assumes I already have a site. | |
| Control: unrelated visitors | r2·57 | desktop | Interdisciplinary, 45-54, Moderate, Agonizes | / → /sample → /sample/report.html | 4 | 0 | left · not for me | the sample report, just to see what the write-up looks like | That paragraph under the headline is a wall of text — I skipped most of it |
| Control: unrelated visitors | r2·58 | desktop | STEM, 45-54, High, Deliberate | / → /sample → /sample/report.html | 16 | 0 | left · not for me | The sample report, to see if the output is credible, The sample report | The small 'Usability studies with sim...' label at the top is cramped and easy to miss |
| Control: unrelated visitors | r2·59 | mobile | Business, 35-44, Low, Agonizes | / → /sample | 5 | 0 | left · not for me | Read a sample report, The sample report — to decide if the tool is wor | Walls of text, no video |
| Developers using AI coding agents | r1·0 | desktop | Business, 18-24, Moderate, Balanced | / | 1 | 0 | left · not for me | No video anywhere on the first screen | |
| Developers using AI coding agents | r1·1 | desktop | Arts, 25-34, Moderate, Agonizes | / → /sample → /sample/report.html | 15 | 0 | left · not for me | The sample report, right now, to judge whether simulated feedback is s | The one thing I'd question most — how believable these simulated participants actually are |
| Developers using AI coding agents | r1·2 | mobile | Business, 45-54, Low, Snap decisions | / → /start | 4 | 1 | left · not for me | The free Quick read, if it doesn't need much setup, The free Quick rea | Big block of text, no video — I'd rather watch a demo |
| Developers using AI coding agents | r1·3 | desktop | Arts, 25-34, Low, Deliberate | / → /sample → /sample/report.html | 3 | 0 | left · not for me | Read a sample report, The sample report itself, to judge whether the f | The tiny all-caps line 'USABILITY STUDIES WITH SIMULATED PARTICIPANTS' is fiddly to read |
| Developers using AI coding agents | r1·4 | desktop | STEM, 13-17, High, Quick | / → /sample | 4 | 0 | left · not for me | Nav links like Docs and the logo are tiny and fiddly to hit | |
| Developers using AI coding agents | r1·5 | mobile | Interdisciplinary, 25-34, Low, Deliberate | / → /sample → /sample/report.html | 5 | 0 | left · unsure | Read a sample report (to see if the simulated feedback is actually spe | Walls of text with no video anywhere — I'd rather be shown this than read it on my phone |
| Developers using AI coding agents | r1·6 | desktop | Business, 25-34, Moderate, Balanced | / → /sample → /sample/report.html | 13 | 0 | left · not for me | Read a sample report, Open the sample report and read the summary sect | The whole thing hinges on 'simulated participants' and nothing on this screen proves they |
| Developers using AI coding agents | r1·7 | desktop | Business, 25-34, Moderate, Balanced | / → /sample → /sample/report.html | 25 | 0 | left · not for me | A free Quick read on my own site, if the sample report looks believabl | Some of the labels are set very small, which makes the nav and section markers hard to sca |
| Developers using AI coding agents | r1·8 | desktop | Business, 35-44, Low, Balanced | / → /sample → /sample/report.html | 3 | 0 | left · not for me | Read a sample report, Open the sample report | Long text, no video, tiny labels in the header |
| Developers using AI coding agents | r1·9 | mobile | Interdisciplinary, 65-74, Low, Deliberate | / → /sample → /sample/report.html | 15 | 0 | left · not for me | Read a sample report, to see whether the findings are specific enough | Some of the text is very small and the top controls are fiddly to tap on a phone |
| Developers using AI coding agents | r1·10 | desktop | STEM, 35-44, Moderate, Balanced | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Read a sample report | Too much text on first screen and no video, even though the headline says 'watch'. |
| Developers using AI coding agents | r1·11 | desktop | Interdisciplinary, 25-34, Moderate, Agonizes | / → /sample → /sample/report.html | 3 | 0 | left · not for me | Read a sample report, Open the sample report | No video, and the page is very text-heavy. |
| Developers using AI coding agents | r1·12 | desktop | STEM, 13-17, Very high, Snap decisions | / → /sample | 2 | 0 | left · not for me | Reading the sample report | Way too much text, no video anywhere |
| Developers using AI coding agents | r1·13 | mobile | Social science, 25-34, Moderate, Snap decisions | / → /sample → /sample/report.html | 10 | 0 | left · not for me | Read a sample report, Read a sample report (still haven't found one) | No video, just text, and I strongly prefer video. |
| Developers using AI coding agents | r1·14 | mobile | Business, 18-24, High, Snap decisions | / → /start | 2 | 0 | left · not for me | Read a sample report | Tiny 'ProductTex' and 'Menu' are fiddly on mobile. |
| Developers using AI coding agents | r1·15 | desktop | Humanities, 18-24, Moderate, Deliberate | / | 1 | 0 | left · not for me | No video, just a long wall of text | |
| Developers using AI coding agents | r1·16 | desktop | STEM, 13-17, Moderate, Agonizes | / → /sample → /sample/report.html | 23 | 0 | left · not for me | Read a sample report, Read a sample report to understand what the prod | Some text is very small, like 'Usability studies with sim' and 'The gap'. |
| Developers using AI coding agents | r1·17 | desktop | STEM, 25-34, Low, Deliberate | / → /sample → /sample/report.html | 10 | 0 | left · not for me | The sample report — to judge whether the output is worth anything befo | The body copy is one dense grey block — for a tool about usability that's a bit ironic |
| Developers using AI coding agents | r1·18 | mobile | STEM, 35-44, Very high, Agonizes | / → /pricing | 30 | 0 | cut off · unsure | Read a sample report, The free first Quick read, to see the report for | No video anywhere, even though I’d rather watch a demo. |
| Developers using AI coding agents | r1·19 | mobile | Business, 75-84, Moderate, Agonizes | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Read a sample report, Open the sample report | Some text is very small, like the category label and 'THE GAP' |
| Developers using AI coding agents | r1·20 | desktop | Interdisciplinary, 13-17, Moderate, Snap decisions | / → /sample → /sample/report.html | 13 | 0 | left · not for me | The sample report, just to see what the output looks like, The sample | No video or demo — I have to read a wall of text to get it |
| Developers using AI coding agents | r1·21 | desktop | STEM, 35-44, Moderate, Deliberate | / → /sample → /sample/report.html | 12 | 0 | left · not for me | The sample report, to judge whether the output is actually useful., Th | The hero is a wall of text with no video, which puts me off. |
| Developers using AI coding agents | r1·22 | mobile | STEM, 13-17, Moderate, Balanced | / | 2 | 0 | left · not for me | The logo and menu are tiny and fiddly on mobile. | |
| Developers using AI coding agents | r1·23 | mobile | Interdisciplinary, 25-34, Low, Deliberate | / → /sample → /sample/report.html | 16 | 0 | left · not for me | Read a sample report, Reading a real report end-to-end to see if the n | The small menu and logo are fiddly on mobile. |
| Developers using AI coding agents | r1·24 | desktop | Interdisciplinary, 25-34, Moderate, Snap decisions | / | 2 | 0 | left · not for me | No video anywhere up top | |
| Developers using AI coding agents | r1·25 | desktop | Business, 25-34, Moderate, Balanced | / → /sample → /sample/report.html | 22 | 0 | left · for me | Opening the sample report to see the actual output format, The ranked | The paragraph about simulated people is fine but I still don't know how trustworthy simula |
| Developers using AI coding agents | r1·26 | mobile | Humanities, 18-24, Low, Deliberate | / | 5 | 0 | left · not for me | No video visible, just text on mobile | |
| Developers using AI coding agents | r1·27 | desktop | Interdisciplinary, 35-44, Low, Balanced | / | 1 | 0 | left · not for me | No video even though the headline says 'Watch' | |
| Developers using AI coding agents | r1·28 | desktop | Social science, 35-44, Moderate, Balanced | / → /sample → /sample/report.html | 4 | 0 | left · unsure | the sample report, to judge whether simulated feedback is worth anythi | The fine print about 'simulated, not real people' is buried under the buttons and tiny, ex |
| Developers using AI coding agents | r1·29 | desktop | Humanities, 18-24, Moderate, Snap decisions | / → /sample | 2 | 0 | left · not for me | The little 'Usability studies with simulated participants' label above the headline is tin | |
| Developers using AI coding agents | r1·30 | mobile | Arts, 18-24, None, Balanced | / → /sample → /sample/report.html | 7 | 0 | left · not for me | The free sample report, to see what one of these actually looks like, | Wall of text on a phone, no video or screenshot anywhere |
| Developers using AI coding agents | r1·31 | mobile | Business, 18-24, Moderate, Balanced | / → /sample → /sample/report.html | 16 | 0 | left · not for me | Reading a sample report to see if the output is specific enough to act | The ProductTex logo and Menu are tiny on mobile — fiddly tap targets |
| Developers using AI coding agents | r1·32 | mobile | STEM, 18-24, Moderate, Quick | / | 1 | 0 | left · not for me | Wall of text with no video | |
| Developers using AI coding agents | r1·33 | desktop | Interdisciplinary, 55-64, Moderate, Snap decisions | / → /sample → /sample/report.html | 5 | 0 | left · not for me | The free Quick read, if the sample report looks convincing, The free Q | Big wall of text — I prefer video, this is a lot to read at once |
| Developers using AI coding agents | r1·34 | mobile | STEM, 55-64, Very high, Snap decisions | / → /sample → /sample/report.html | 22 | 0 | left · not for me | Read a sample report to evaluate the output., Opening the sample repor | No video, and I prefer watching to reading. |
| Developers using AI coding agents | r1·35 | mobile | Humanities, 25-34, Moderate, Balanced | / → /sample → /sample/report.html | 14 | 0 | left · not for me | Reading a sample report to judge how useful the simulated feedback act | Some of the text is quite small on this phone screen — the 'Usability studies with sim...' |
| Developers using AI coding agents | r1·36 | desktop | STEM, 13-17, Moderate, Snap decisions | / → /start | 2 | 0 | left · not for me | The free quick read, just to see what fake visitors say about a page | That tiny all-caps line above the headline is basically invisible |
| Developers using AI coding agents | r1·37 | desktop | Arts, 18-24, Moderate, Quick | / → /sample → /sample/report.html | 13 | 0 | left · not for me | the free sample report, to judge whether the simulated feedback sounds | It's a wall of text and I'd rather watch a 60-second demo than read three paragraphs |
| Developers using AI coding agents | r1·38 | desktop | STEM, 65-74, Low, Quick | / → /sample → /sample/report.html | 11 | 0 | left · not for me | the sample report, just to see it, the sample report, just to see how | The small grey line about simulated people being free/not real is easy to miss and reads a |
| Developers using AI coding agents | r1·39 | desktop | STEM, 18-24, Moderate, Agonizes | / → /sample → /sample/report.html | 4 | 0 | left · unsure | Reading a sample report to judge whether the simulated feedback is act | No video or visual example — it's just paragraphs, which is tiring to get through |
| Developers using AI coding agents | r1·40 | desktop | Business, 18-24, Moderate, Agonizes | / → /pricing → /sample → /sample/report.html | 5 | 0 | left · not for me | The free Quick read on a site, just to see what the fake visitors say, | The big block of grey paragraph under the headline is way too much text for me |
| Developers using AI coding agents | r1·41 | desktop | Business, 13-17, High, Agonizes | / | 1 | 0 | left · not for me | The wall of text and tiny 'Usability studies with simulated participants' label put me off | |
| Developers using AI coding agents | r1·42 | mobile | STEM, 45-54, Moderate, Snap decisions | / → /start | 2 | 0 | left · not for me | Get a free Quick read | Wall of text to wade through |
| Developers using AI coding agents | r1·43 | desktop | Social science, 18-24, Moderate, Balanced | / → /sample → /sample/report.html | 12 | 0 | left · not for me | Reading a sample report to judge the quality of the simulated quotes., | No video anywhere — for something this visual, a 60-second walkthrough would beat this wal |
| Developers using AI coding agents | r1·44 | mobile | Arts, 35-44, Very high, Snap decisions | / → /sample → /sample/report.html | 3 | 0 | left · not for me | It assumes I have a site to test, which I don't. | |
| Developers using AI coding agents | r1·45 | desktop | Social science, 18-24, Very high, Quick | / → /sample → /sample/report.html | 4 | 0 | left · not for me | The sample report, to judge whether simulated feedback is credible bef | The headline is long and I had to read it twice |
| Developers using AI coding agents | r1·46 | mobile | Humanities, 18-24, Low, Balanced | / → /sample → /sample/report.html | 6 | 0 | left · not for me | Read a sample report, Reading the sample report itself | Assumes I have a site I control |
| Developers using AI coding agents | r1·47 | desktop | Interdisciplinary, 18-24, Moderate, Quick | / → /sample → /sample/report.html | 26 | 0 | left · not for me | Read a sample report, Read the actual sample report to judge whether t | The nav items 'Docs' and the logo look tiny and fiddly to hit |
| Developers using AI coding agents | r1·48 | mobile | Interdisciplinary, 18-24, Moderate, Quick | / → /sample → /sample/report.html | 4 | 0 | left · unsure | Read a sample report, Open the sample report | The long paragraph under the headline is a lot for a phone screen. |
| Developers using AI coding agents | r1·49 | mobile | Interdisciplinary, 45-54, Moderate, Agonizes | / → /sample → /sample/report.html | 19 | 0 | left · not for me | Read a sample report, Open the sample report | Some of the text is really tiny on my phone — the blue label up top and 'The gap' are hard |
| Developers using AI coding agents | r1·50 | mobile | Medicine, 35-44, Low, Quick | / → /sample → /sample/report.html | 5 | 0 | left · not for me | Open the sample report to see the actual output, The sample report out | The tiny all-caps kicker is hard to read on mobile. |
| Developers using AI coding agents | r1·51 | mobile | Humanities, 55-64, High, Deliberate | / → /sample → /sample/report.html | 6 | 0 | left · unsure | Read a sample report, Reading a sample report to judge if the output i | The page is text-heavy and there's no video, which I'd prefer |
| Developers using AI coding agents | r1·52 | desktop | Business, 55-64, High, Snap decisions | / → /start | 2 | 0 | left · not for me | The page is a long wall of text with no video, which puts me off. | |
| Developers using AI coding agents | r1·53 | desktop | Interdisciplinary, 18-24, High, Agonizes | / → /sample → /sample/report.html | 10 | 0 | left · not for me | Read a sample report, Open the sample report | Lots of small text and a long page below makes me want to bounce before reading much |
| Developers using AI coding agents | r1·54 | desktop | STEM, 25-34, Moderate, Snap decisions | / → /sample → /sample/report.html | 29 | 0 | cut off · for me | The sample report, to judge whether the simulated feedback is useful o | 'Quick read' is jargon I have to guess at — is it a report type or a speed promise? |
| Developers using AI coding agents | r1·55 | desktop | Interdisciplinary, 75-84, Moderate, Snap decisions | / → /sample → /sample/report.html | 4 | 0 | left · not for me | The sample report, just to see the output, The sample report, to judge | The top row is crowded and the tiny 'simulated participants' line is hard to read. |
| Developers using AI coding agents | r1·56 | mobile | Business, 25-34, High, Quick | / → /sample → /sample/report.html | 23 | 0 | left · not for me | Read a sample report, to judge whether the findings are useful or just | It's a wall of text with no video, which is how I actually want to see this explained |
| Developers using AI coding agents | r1·57 | desktop | Interdisciplinary, 35-44, High, Snap decisions | / → /sample → /sample/report.html | 9 | 0 | left · unsure | Read a sample report, Read the actual sample report | No video demo — for a product about watching simulated visitors, I expected a video right |
| Developers using AI coding agents | r1·58 | mobile | Interdisciplinary, 13-17, High, Quick | / → /sample → /sample/report.html | 17 | 0 | left · not for me | the free sample report, to judge if the simulated feedback is believab | 'Quick read' and 'sample report' — I can't tell the difference between them from the label |
| Developers using AI coding agents | r1·59 | mobile | Medicine, 25-34, Moderate, Agonizes | / | 1 | 0 | left · not for me | Too much text on mobile with no video | |
| Developers using AI coding agents | r2·0 | mobile | Business, 18-24, Moderate, Balanced | / → /sample → /sample/report.html | 6 | 0 | left · not for me | Read a sample report, Open the sample report | Too much text and no video — I prefer watching. |
| Developers using AI coding agents | r2·1 | desktop | Interdisciplinary, 35-44, Very high, Quick | / → /sample → /sample/report.html | 9 | 0 | left · not for me | The sample report, to judge if this is real or just AI guessing., The | No video anywhere, and I'd much rather watch a 90-second demo than read this wall of text |
| Developers using AI coding agents | r2·2 | desktop | Social science, 45-54, Moderate, Quick | / → /sample → /sample/report.html | 4 | 0 | left · not for me | the sample report | The paragraph under the headline is a lot to read with a short attention span. |
| Developers using AI coding agents | r2·3 | mobile | Humanities, 45-54, High, Snap decisions | / → /sample → /sample/report.html | 17 | 0 | left · not for me | Reading a sample report to judge the quality of the write-up, Reading | 'ProductTex' as a name tells me nothing about what it does |
| Developers using AI coding agents | r2·4 | mobile | STEM, 18-24, Low, Snap decisions | / → /sample → /sample/report.html | 7 | 0 | left · unsure | Read a sample report, Open the sample report | The small text like 'Usability studies with sim' and 'The gap' is hard to read on my phone |
| Developers using AI coding agents | r2·5 | mobile | STEM, 25-34, Very high, Balanced | / → /sample → /start → /dashboard → /pricing | 29 | 3 | cut off · unsure | The free Quick read on a site I control, The free Quick read on a site | The label 'USABILITY STUDIES WITH SIMULATED PARTICIPANTS' is tiny and grey — easy to skip |
| Developers using AI coding agents | r2·6 | desktop | Interdisciplinary, 13-17, Moderate, Agonizes | / | 2 | 0 | left · not for me | It's a wall of text with no video or visuals, which is exactly the format I avoid | |
| Developers using AI coding agents | r2·7 | mobile | Business, 25-34, Moderate, Deliberate | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Read a sample report, to decide whether this is credible enough to eve | The word 'simulated' hits me twice before I've seen any output — hard not to read it as a |
| Developers using AI coding agents | r2·8 | mobile | STEM, 25-34, Low, Agonizes | / → /method → /sample → /sample/report.html | 30 | 0 | left · not for me | Read a sample report, A sample report, if one shows up further down | No video on this screen, even though I prefer to learn that way |
| Developers using AI coding agents | r2·9 | mobile | Interdisciplinary, 45-54, High, Balanced | / → /sample → /sample/report.html | 21 | 0 | left · not for me | Read a sample report, Open the sample report | Mobile text is tiny — the little blue kicker and the 'THE GAP' label are a squint |
| Developers using AI coding agents | r2·10 | desktop | Arts, 25-34, Moderate, Deliberate | / → /sample → /sample/report.html | 3 | 0 | left · not for me | No video anywhere, even though the headline says 'Watch'. | |
| Developers using AI coding agents | r2·11 | mobile | Business, 13-17, High, Balanced | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Some text is really small on mobile, like 'ProductTex' and 'The gap'. | |
| Developers using AI coding agents | r2·12 | desktop | STEM, 35-44, Low, Balanced | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Sample report, to judge the product, The sample report itself, to deci | No video despite the headline saying 'Watch' |
| Developers using AI coding agents | r2·13 | mobile | Interdisciplinary, 18-24, Moderate, Snap decisions | / → /start | 2 | 0 | left · not for me | The free Quick read, if I had a website to test. | No video, and the page is pretty text-heavy for a phone. |
| Developers using AI coding agents | r2·14 | mobile | Interdisciplinary, 18-24, Moderate, Balanced | / → /sample → /sample/report.html | 29 | 0 | left · not for me | Read a sample report, Open the sample report to see what the output ac | “Simulated participants” and “real browser” feel a bit contradictory at first glance. |
| Developers using AI coding agents | r2·15 | desktop | Business, 13-17, Moderate, Deliberate | / → /sample → /sample/report.html | 10 | 0 | left · unsure | The free Quick read on a site I control, The free Quick read on a site | No screenshot or video of an actual report on the first screen — it's all description, I h |
| Developers using AI coding agents | r2·16 | mobile | STEM, 65-74, Moderate, Snap decisions | / → /sample → /sample/report.html | 16 | 0 | left · not for me | A sample report to evaluate before running one on my own site, Reading | "Quick read" is unexplained jargon — I don't know what I'd be getting |
| Developers using AI coding agents | r2·17 | mobile | Arts, 45-54, None, Balanced | / → /sample → /sample/report.html | 3 | 0 | left · not for me | Read a sample report | Some text is very small and the links at top are tiny, which is annoying on mobile. |
| Developers using AI coding agents | r2·18 | desktop | STEM, 13-17, None, Deliberate | / → /sample → /sample/report.html | 5 | 0 | left · not for me | The page is a wall of text and I'd rather watch a video. | |
| Developers using AI coding agents | r2·19 | mobile | Humanities, 65-74, Moderate, Snap decisions | / → /start → /dashboard | 9 | 1 | left · not for me | The free Quick read on a site I control, The free Quick read on a site | Too much text on a small screen |
| Developers using AI coding agents | r2·20 | mobile | Interdisciplinary, 18-24, Low, Deliberate | / → /sample → /sample/report.html | 10 | 0 | left · not for me | Read a sample report, Open the sample report | The page is a wall of text; I'd rather watch a quick demo. |
| Developers using AI coding agents | r2·21 | desktop | STEM, 45-54, High, Snap decisions | / → /sample → /sample/report.html | 25 | 0 | left · not for me | The sample report, to judge whether the simulated feedback is credible | No video anywhere — a wall of text with no demo, which is exactly how I don't like learnin |
| Developers using AI coding agents | r2·22 | desktop | Interdisciplinary, 25-34, High, Quick | / → /start | 6 | 1 | left · not for me | Get a free Quick read, Get a free Quick read. | It says 'Watch' but there is no video — just a wall of text. |
| Developers using AI coding agents | r2·23 | desktop | Humanities, 25-34, None, Deliberate | / → /sample → /sample/report.html | 3 | 0 | left · not for me | Read a sample report — to judge whether the output is useful or just g | A wall of text with zero visuals or video — for something about watching people, there's n |
| Developers using AI coding agents | r2·24 | desktop | Social science, 18-24, High, Quick | / → /sample → /sample/report.html | 3 | 0 | left · unsure | the sample report | Long page, lots to scroll, and the free offer needs a site I control — which I don't have. |
| Developers using AI coding agents | r2·25 | desktop | Business, 13-17, Low, Deliberate | / → /sample → /sample/report.html | 6 | 0 | left · not for me | Read a sample report, Open the sample report, to see what the actual d | No video anywhere, just paragraphs — for a tool about watching people get stuck, text is a |
| Developers using AI coding agents | r2·26 | mobile | Interdisciplinary, 18-24, Moderate, Balanced | / → /sample → /sample/report.html | 28 | 0 | left · not for me | The sample report, to judge whether the simulated feedback is actually | The line 'Usability studies with simulated participants' is set tiny and nearly unreadable |
| Developers using AI coding agents | r2·27 | desktop | Humanities, 25-34, Moderate, Snap decisions | / → /sample → /sample/report.html | 6 | 0 | left · not for me | Read a sample report, to judge whether the simulated feedback is worth | It's a wall of text — no video, no demo, nothing to watch, which is exactly how I don't li |
| Developers using AI coding agents | r2·28 | desktop | Business, 18-24, Low, Balanced | / → /sample → /sample/report.html | 11 | 0 | left · not for me | reading the sample report to judge if the output is real or fluff, Rea | 'Quick read' isn't explained — is it a report type or a plan? |
| Developers using AI coding agents | r2·29 | desktop | Social science, 35-44, High, Snap decisions | / → /sample → /sample/report.html | 14 | 0 | left · not for me | the sample report to judge output quality, the Fix check, to verify a | The small all-caps text and tiny nav links are annoying to read on a low-bandwidth connect |
| Developers using AI coding agents | r2·30 | desktop | Social science, 25-34, Moderate, Snap decisions | / → /sample → /sample/report.html | 4 | 0 | left · not for me | The sample report, to judge if the output is credible, The sample repo | No video anywhere — just a wall of text, and I'd rather watch 90 seconds |
| Developers using AI coding agents | r2·31 | desktop | Business, 25-34, Moderate, Quick | / → /sample → /sample/report.html | 11 | 0 | left · unsure | The free Quick read on one page of my site, but only after I trust the | No video or demo — just text, and it keeps going down the page |
| Developers using AI coding agents | r2·32 | desktop | Arts, 25-34, Moderate, Deliberate | / → /sample → /sample/report.html | 30 | 0 | left · unsure | The sample report, to see if the output is believable, The sample repo | The 'simulated, not real people' note makes me wonder how much the quotes can be trusted — |
| Developers using AI coding agents | r2·33 | mobile | Arts, 18-24, Moderate, Quick | / | 1 | 0 | left · not for me | The page assumes I have a live site to test, and I don't. | |
| Developers using AI coding agents | r2·34 | mobile | Business, 13-17, High, Snap decisions | / | 1 | 0 | left · not for me | The headline is a long sentence; I skimmed it but didn't want to read the paragraph below. | |
| Developers using AI coding agents | r2·35 | desktop | Business, 18-24, Moderate, Snap decisions | / → /start → /sample → /sample/report.html | 5 | 0 | left · not for me | Read a sample report, to see the output before committing anything, Re | No video anywhere — it's all headline and paragraphs, which is exactly the format I skip |
| Developers using AI coding agents | r2·36 | desktop | Business, 25-34, Moderate, Snap decisions | / → /sample → /sample/report.html | 13 | 0 | left · not for me | A sample report to judge if the feedback is legit, The sample report i | No video anywhere — I'd rather watch a 60-second demo than read this |
| Developers using AI coding agents | r2·37 | desktop | Interdisciplinary, 45-54, Moderate, Quick | / → /sample → /sample/report.html | 18 | 0 | left · unsure | The sample report, The sample report itself, to judge the output quali | The top label 'USABILITY STUDIES WITH SIMULATED PARTICIPANTS' is very small and easy to mi |
| Developers using AI coding agents | r2·38 | desktop | STEM, 13-17, Moderate, Deliberate | / → /sample → /sample/report.html | 5 | 0 | left · not for me | The sample report, just to satisfy curiosity about how simulated feedb | No video anywhere — I'd rather watch a 30-second demo than read paragraphs |
| Developers using AI coding agents | r2·39 | mobile | Interdisciplinary, 18-24, Low, Snap decisions | / → /start | 2 | 0 | left · not for me | A free Quick read if I had a website to test | The small 'Usability studies with sim' text is hard to read on mobile |
| Developers using AI coding agents | r2·40 | desktop | STEM, 35-44, High, Deliberate | / → /sample → /sample/report.html | 19 | 0 | left · not for me | The free Quick read on my own site, if the sample report looks credibl | 'Simulated participants' is the whole selling point, but there's no explanation of how the |
| Developers using AI coding agents | r2·41 | desktop | Business, 25-34, High, Snap decisions | / → /start → /dashboard | 10 | 4 | left · unsure | The free Quick read on one of my own pages, The free Quick read on one | The block of text under the headline is a bit dense for a first glance, though I'll read i |
| Developers using AI coding agents | r2·42 | desktop | Arts, 35-44, Low, Snap decisions | / → /start → /sample → /sample/report.html | 4 | 0 | left · not for me | The free Quick read, if I had a site to test right now., The free Quic | Wall of text, no video |
| Developers using AI coding agents | r2·43 | desktop | Social science, 45-54, Low, Quick | / | 1 | 0 | left · not for me | The page is text-heavy and assumes I already have a live site to analyze. | |
| Developers using AI coding agents | r2·44 | mobile | Interdisciplinary, 25-34, Very high, Snap decisions | / → /sample → /sample/report.html | 19 | 0 | left · not for me | Read a sample report, Open the sample report | Small top links and a lot of text for a first screen. |
| Developers using AI coding agents | r2·45 | desktop | Interdisciplinary, 45-54, Moderate, Agonizes | / → /sample → /sample/report.html | 4 | 0 | left · not for me | The sample report, to see what the output looks like., The sample repo | The page is very long and keeps going, which makes me want to leave rather than scroll. |
| Developers using AI coding agents | r2·46 | mobile | Interdisciplinary, 18-24, Moderate, Balanced | / → /sample → /sample/report.html | 20 | 0 | left · not for me | Read a sample report, Open and read the full sample report | A lot of text on mobile with no video, even though I prefer video. |
| Developers using AI coding agents | r2·47 | desktop | STEM, 55-64, High, Quick | / → /sample → /sample/report.html | 4 | 0 | left · not for me | A sample report, just to judge whether the simulated feedback is worth | Some of the wording up top is tiny and grey, I nearly skipped it |
| Developers using AI coding agents | r2·48 | desktop | STEM, 13-17, Moderate, Balanced | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Read a sample report, Open the sample report | The tiny 'THE GAP' label and small grey text are easy to miss or struggle to read. |
| Developers using AI coding agents | r2·49 | mobile | Humanities, 25-34, High, Balanced | / → /sample → /sample/report.html | 22 | 0 | left · not for me | The sample report, to judge whether the simulated findings are worth a | The small links top-left and top-right ('ProductTex', 'Menu') are fiddly on a phone. |
| Developers using AI coding agents | r2·50 | desktop | Business, 25-34, High, Quick | / → /sample → /sample/report.html | 20 | 0 | left · not for me | Sample report, to sanity-check the quality of the simulated quotes bef | The 'simulated, not real people' disclaimer is doing a lot of quiet work — it makes me dou |
| Developers using AI coding agents | r2·51 | desktop | Medicine, 25-34, Low, Quick | / → /sample → /sample/report.html | 5 | 0 | left · not for me | Read a sample report, Open the sample report | 'Usability studies with simulated participants' is rendered tiny and awkwardly truncated u |
| Developers using AI coding agents | r2·52 | mobile | Humanities, 25-34, High, Agonizes | / → /sample → /sample/report.html | 17 | 0 | left · not for me | The sample report, to see if the 'why' is actually believable or just | The little blue label 'Usability studies with simulated participants' is tiny and hard to |
| Developers using AI coding agents | r2·53 | desktop | Arts, 25-34, Moderate, Quick | / → /sample → /sample/report.html | 4 | 0 | left · not for me | the sample report, just to see the format, The sample report. | The line about 'simulated, not real people' shows up twice and makes me wonder how much th |
| Developers using AI coding agents | r2·54 | desktop | STEM, 35-44, Moderate, Snap decisions | / → /sample → /sample/report.html | 11 | 0 | left · not for me | The sample report, just to evaluate credibility, The sample report its | 'Simulated participants' in tiny letters up top, then the same caveat repeated lower down |
| Developers using AI coding agents | r2·55 | mobile | Business, 25-34, Low, Agonizes | / → /sample → /sample/report.html | 4 | 0 | left · not for me | The sample report, just to understand what this actually looks like., | Wall of text and no video, which put me off because I strongly prefer watching. |
| Developers using AI coding agents | r2·56 | mobile | Arts, 35-44, Moderate, Quick | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Small text and fiddly tap targets on mobile | |
| Developers using AI coding agents | r2·57 | mobile | Business, 25-34, None, Agonizes | / | 7 | 0 | left · not for me | Read a sample report, The result format to understand where mobile use | The ProductTex and Menu links at the top are tiny and fiddly to hit on a phone |
| Developers using AI coding agents | r2·58 | mobile | STEM, 13-17, High, Snap decisions | / | 1 | 0 | left · not for me | It assumes I have a website to test. | |
| Developers using AI coding agents | r2·59 | desktop | Interdisciplinary, 35-44, Moderate, Quick | / → /sample → /sample/report.html | 11 | 0 | left · not for me | Read a sample report, to judge if the simulated feedback is believable | A big block of text and no video anywhere near the top — I skim, I don't read paragraphs |
| Product people at software companies | r1·0 | desktop | Business, 75-84, None, Agonizes | / | 2 | 0 | left · not for me | No video visible even though the headline says 'Watch' | |
| Product people at software companies | r1·1 | desktop | Business, 13-17, Moderate, Snap decisions | / → /sample → /sample/report.html | 11 | 0 | left · not for me | The sample report to see what the output looks like, The sample report | The free Quick read offer assumes I control a site, which I don't. |
| Product people at software companies | r1·2 | mobile | Business, 5-12, Moderate, Quick | / → /sample → /sample/report.html | 6 | 0 | left · not for me | Read a sample report, Open the sample report | No video anywhere — I'd understand this in 30 seconds if it were shown instead of written |
| Product people at software companies | r1·3 | desktop | Social science, 55-64, High, Agonizes | / → /sample → /sample/report.html | 5 | 0 | left · not for me | The tiny all-caps label at the top is hard to read; I had to squint. | |
| Product people at software companies | r1·4 | desktop | Social science, 25-34, Moderate, Deliberate | / → /sample → /sample/report.html | 13 | 0 | left · unsure | Read a sample report, Opening the sample report to gauge what a real r | The top nav mixes 'Method' and 'Docs' in a way that feels a bit vague for a first-time vis |
| Product people at software companies | r1·5 | mobile | Interdisciplinary, 25-34, Low, Snap decisions | / → /sample → /sample/report.html | 5 | 0 | left · not for me | Read a sample report, Open the sample report | The small uppercase label and dense paragraph under the huge headline are a lot for a phon |
| Product people at software companies | r1·6 | desktop | Arts, 25-34, Low, Balanced | / → /sample → /sample/report.html | 9 | 0 | left · not for me | The sample report, to judge whether the simulated feedback is useful., | The tiny label 'USABILITY STUDIES WITH SIMULATED PARTICIPANTS' is easy to miss. |
| Product people at software companies | r1·7 | mobile | Social science, 25-34, High, Agonizes | / → /sample → /sample/report.html | 13 | 0 | left · not for me | Get a free Quick read, A free Quick read on a real page of mine, if I | Some text is tiny on my phone, like the category label at the top. |
| Product people at software companies | r1·8 | mobile | Arts, 5-12, Moderate, Deliberate | / | 1 | 0 | left · not for me | Too much small text on a phone screen | |
| Product people at software companies | r1·9 | mobile | Business, 13-17, Moderate, Agonizes | / → /sample → /method → /sample/report.html | 24 | 0 | left · not for me | Read a sample report, Reading the sample report to learn how these sim | The small text like 'Usability studies with simulated participants' and 'The gap' is hard |
| Product people at software companies | r1·10 | desktop | Interdisciplinary, 35-44, Moderate, Balanced | / → /sample → /sample/report.html | 11 | 0 | left · not for me | Read a sample report to judge the product, Read the sample report to j | No video demo, just a lot of text |
| Product people at software companies | r1·11 | desktop | Business, 5-12, Moderate, Quick | / | 1 | 0 | left · not for me | The page is a wall of text with no video, which puts me off. | |
| Product people at software companies | r1·12 | desktop | Social science, 25-34, Low, Balanced | / → /method → /sample → /sample/report.html | 14 | 0 | left · not for me | The 'What it cannot tell you' section, to decide before I ever run a s | The tiny all-caps labels ('USABILITY STUDIES WITH SIMULATED PARTICIPANTS', 'THE GAP') are |
| Product people at software companies | r1·13 | desktop | STEM, 35-44, Low, Deliberate | / → /sample → /sample/report.html | 5 | 0 | left · unsure | The sample report, to see if the output is useful before giving a URL. | Wall of text and no video — I prefer watching to reading. |
| Product people at software companies | r1·14 | mobile | Interdisciplinary, 85+, Low, Deliberate | / | 1 | 0 | left · not for me | Too many long paragraphs in small text | |
| Product people at software companies | r1·15 | mobile | STEM, 18-24, Moderate, Deliberate | / → /sample → /sample/report.html | 7 | 0 | left · not for me | Read a sample report, Seeing the actual sample report layout before pa | No video, just paragraphs — and I prefer watching. |
| Product people at software companies | r1·16 | desktop | Arts, 25-34, Moderate, Quick | / → /sample | 2 | 0 | left · not for me | The headline says 'Watch' but there's no video, just a wall of text. | |
| Product people at software companies | r1·17 | desktop | Interdisciplinary, 5-12, Low, Balanced | / | 1 | 0 | left · not for me | The page is very text-heavy, and it warns that it continues a long way further down. | |
| Product people at software companies | r1·18 | mobile | Business, 35-44, High, Balanced | / → /docs → /pricing | 29 | 0 | cut off · unsure | Read a sample report, The worked real-result write-up — it shows me ex | The top menu links are tiny on mobile. |
| Product people at software companies | r1·19 | mobile | Social science, 18-24, High, Balanced | / → /sample → /sample/report.html | 6 | 0 | left · not for me | Read a sample report | The page is a wall of text and the actual report is still further down or behind a button |
| Product people at software companies | r1·20 | desktop | Interdisciplinary, 45-54, Moderate, Agonizes | / → /sample → /sample/report.html | 6 | 0 | left · not for me | No video anywhere, just a wall of text. | |
| Product people at software companies | r1·21 | mobile | STEM, 45-54, Very high, Agonizes | / → /sample → /sample/report.html | 6 | 0 | left · unsure | Read a sample report, to judge whether the simulated feedback is speci | Without "USD" or "US dollars" next to "Get a free Quick read," I can't tell if it's free t |
| Product people at software companies | r1·22 | desktop | Arts, 35-44, Moderate, Deliberate | / → /sample → /sample/report.html | 20 | 0 | left · not for me | Read a sample report, Open the sample report | The phrase 'simulated participants' is a bit worrying; I can't tell if this is AI pretendi |
| Product people at software companies | r1·23 | mobile | Interdisciplinary, Under 5, Very high, Agonizes | / | 2 | 0 | left · not for me | Wall of text on mobile | |
| Product people at software companies | r1·24 | desktop | STEM, 18-24, Moderate, Deliberate | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Read a sample report, Open the sample report | The page is very wordy and takes a while to get to the point; I had to skim a lot. |
| Product people at software companies | r1·25 | desktop | Business, 13-17, Moderate, Balanced | / → /sample → /sample/report.html | 5 | 0 | left · unsure | Read a sample report, Open the sample report | No video on the first screen even though the headline says 'Watch' |
| Product people at software companies | r1·26 | mobile | Humanities, 35-44, Low, Deliberate | / → /sample → /sample/report.html | 6 | 0 | left · not for me | Read a sample report — to judge quality before submitting a URL, Open | Big block of grey body text on a phone with no visuals or example — heavy going |
| Product people at software companies | r1·27 | mobile | Humanities, 5-12, Low, Balanced | / | 1 | 0 | left · not for me | No video despite 'Watch' | |
| Product people at software companies | r1·28 | mobile | Business, 45-54, Moderate, Quick | / → /sample → /sample/report.html | 18 | 0 | left · not for me | Read a sample report, Open the sample report to see how findings are a | Long text-heavy page on mobile, and I prefer video. |
| Product people at software companies | r1·29 | mobile | Business, 13-17, Low, Deliberate | / → /sample → /sample/report.html | 16 | 0 | left · not for me | Read a sample report, just to judge whether the feedback is genuine or | The small all-caps line at the top is tiny on my phone and I nearly missed it |
| Product people at software companies | r1·30 | mobile | Interdisciplinary, 85+, Moderate, Balanced | / → /sample → /sample/report.html | 21 | 0 | left · not for me | A sample report, to judge whether the feedback is worth anything, The | The lettering is very small on my phone, I had to squint at the top lines |
| Product people at software companies | r1·31 | desktop | STEM, 35-44, Moderate, Deliberate | / → /sample → /sample/report.html | 11 | 0 | left · not for me | The sample report, to judge whether the simulated feedback is credible | The tiny eyebrow line 'Usability studies with simulated participants' is nearly unreadable |
| Product people at software companies | r1·32 | mobile | Business, 5-12, Moderate, Quick | / → /sample → /sample/report.html | 6 | 0 | left · unsure | Read a sample report, Open the sample report (now actually clicking it | Total wall of text, no video or screenshot anywhere — for something visual like 'watching |
| Product people at software companies | r1·33 | mobile | Humanities, 25-34, High, Quick | / → /sample → /sample/report.html | 6 | 0 | left · not for me | Read a sample report | No video on the first screen, and the page is a wall of text on mobile |
| Product people at software companies | r1·34 | mobile | Medicine, 25-34, Moderate, Balanced | / → /sample | 3 | 0 | left · not for me | Text-heavy for a very short attention span | |
| Product people at software companies | r1·35 | desktop | Social science, 25-34, Low, Balanced | / → /sample → /sample/report.html | 12 | 0 | left · for me | Reading a sample report to judge the quality, Opening the sample repor | The intro is two dense paragraphs plus a long 'THE GAP' block — a lot of reading before I |
| Product people at software companies | r1·36 | desktop | Business, 45-54, Very high, Quick | / → /sample → /sample/report.html | 16 | 0 | left · not for me | A sample report, to judge whether the output is usable., The sample re | The words 'simulated participants' make me doubt the whole value — I keep wondering how fa |
| Product people at software companies | r1·37 | mobile | STEM, 85+, Moderate, Balanced | / → /sample → /sample/report.html | 5 | 0 | left · unsure | Read a sample report, Opening the sample report to judge whether the f | The small uppercase line at the top is hard to read on a phone |
| Product people at software companies | r1·38 | desktop | Interdisciplinary, 55-64, Low, Agonizes | / | 1 | 0 | left · not for me | All text, no video despite the word 'Watch' | |
| Product people at software companies | r1·39 | mobile | Business, 25-34, Low, Snap decisions | / → /start | 4 | 1 | left · not for me | The free Quick read, if I had a site to test., The free Quick read, on | The small text and tiny top buttons are fiddly on mobile. |
| Product people at software companies | r1·40 | mobile | Business, 25-34, Moderate, Deliberate | / → /sample → /start → /dashboard | 12 | 1 | left · not for me | The free Quick read on a site I control., The free Quick read. | Very small text and tiny menu controls are hard to read and tap on a phone. |
| Product people at software companies | r1·41 | desktop | Humanities, 35-44, Moderate, Quick | / → /sample → /sample/report.html | 11 | 0 | left · not for me | the free Quick read, later, if I had a page of my own to point it at, | The 'simulated, not real people' disclaimer is buried in small grey text under the buttons |
| Product people at software companies | r1·42 | mobile | STEM, 18-24, Moderate, Quick | / → /sample → /sample/report.html | 30 | 0 | cut off · unsure | A sample report, The sample report itself, to see how a report reads b | The all-caps labels are tiny on mobile and easy to skip. |
| Product people at software companies | r1·43 | mobile | STEM, 55-64, Moderate, Deliberate | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Read a sample report, Open the sample report | No video, just a wall of text on my phone. |
| Product people at software companies | r1·44 | desktop | Humanities, 18-24, Moderate, Quick | / → /sample | 2 | 0 | left · not for me | Too much text and no video, so I skimmed instead of reading. | |
| Product people at software companies | r1·45 | mobile | STEM, 13-17, Moderate, Agonizes | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Read a sample report, Open the sample report | The bottom half of the page is a long paragraph I basically skipped |
| Product people at software companies | r1·46 | mobile | Arts, 18-24, High, Deliberate | / → /docs | 10 | 0 | left · not for me | Read a sample report to see what the output looks like., Reading a ful | The tiny text for the kicker and section labels is hard to read on my phone. |
| Product people at software companies | r1·47 | desktop | Arts, Under 5, High, Agonizes | / → /sample → /sample/report.html | 11 | 0 | left · not for me | Read a sample report, Open the sample report | A lot of small plain text and no picture of the actual report |
| Product people at software companies | r1·48 | mobile | Interdisciplinary, 18-24, Moderate, Quick | / → /sample → /sample/report.html | 14 | 0 | left · not for me | Read a sample report, The sample report — I want to see if the output | No video, just a lot of text |
| Product people at software companies | r1·49 | mobile | Business, 13-17, Moderate, Snap decisions | / → /start → /method | 5 | 0 | left · not for me | the free Quick read on a site I control | The small caps line 'USABILITY STUDIES WITH SIMULATED PARTICIPANTS' is tiny and hard to re |
| Product people at software companies | r1·50 | desktop | Business, 25-34, Moderate, Quick | / → /sample → /sample/report.html | 5 | 0 | left · not for me | Read a sample report, Open the sample report | The phrase 'simulated participants' is a little confusing and slightly off-putting at firs |
| Product people at software companies | r1·51 | mobile | Interdisciplinary, 45-54, Moderate, Balanced | / → /sample → /sample/report.html | 17 | 0 | left · not for me | Read a sample report, The sample report | The label text above the headline is tiny and faint on a phone |
| Product people at software companies | r1·52 | desktop | Business, 45-54, Low, Agonizes | / → /sample → /sample/report.html | 15 | 0 | left · not for me | A free sample report to judge the quality of the simulated feedback, T | 'Simulated participants' is doing a lot of work in one line — I had to read twice to see t |
| Product people at software companies | r1·53 | desktop | STEM, 18-24, Moderate, Balanced | / → /sample → /sample/report.html | 9 | 0 | left · not for me | Read a sample report, Open the sample report | No video on a page that is mostly text — I'd rather watch a short demo than read this much |
| Product people at software companies | r1·54 | mobile | STEM, 45-54, Low, Snap decisions | / → /start → /dashboard | 11 | 2 | left · not for me | The free Quick read, The free Quick read on my own site | Small top nav and tiny grey text are hard to read on mobile. |
| Product people at software companies | r1·55 | mobile | Interdisciplinary, 45-54, Moderate, Snap decisions | / → /start → /method | 5 | 0 | left · unsure | The free Quick read, if it works on a site I control., The free Quick | No video, even though the headline says 'Watch'. |
| Product people at software companies | r1·56 | desktop | Humanities, 13-17, Moderate, Deliberate | / → /sample → /sample/report.html | 8 | 0 | left · not for me | Read a sample report, to judge whether the simulated feedback is belie | It's a wall of text — a 30-second video or a screenshot of the report would have done more |
| Product people at software companies | r1·57 | mobile | Interdisciplinary, 55-64, Moderate, Agonizes | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Read a sample report, Open the sample report | The paragraph is long and small on my phone |
| Product people at software companies | r1·58 | desktop | Social science, 35-44, Moderate, Balanced | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Read a sample report, Open the sample report | The headline says ‘Watch’ but there is no video, just a wall of text on the first screen. |
| Product people at software companies | r1·59 | desktop | Business, 18-24, Low, Deliberate | / → /sample → /sample/report.html | 3 | 0 | left · not for me | Read a sample report, Open the sample report | Long block of text, no video or demo, so I do not want to read it all. |
| Product people at software companies | r2·0 | mobile | Business, 25-34, Moderate, Deliberate | / → /pricing → /start → /method | 30 | 1 | cut off · for me | Read a sample report, A full sample report with the same structure as | The first screen is text-heavy, and the 'simulated, not real people' caveat might make me |
| Product people at software companies | r2·1 | desktop | Arts, 18-24, Moderate, Deliberate | / → /sample | 2 | 0 | left · not for me | “Quick read” is jargon — I do not know what I am getting if it is free. | |
| Product people at software companies | r2·2 | desktop | STEM, 25-34, Moderate, Deliberate | / → /sample → /sample/report.html | 7 | 0 | left · unsure | Reading a sample report, Reading the sample report to judge whether th | The page feels text-heavy after the headline. |
| Product people at software companies | r2·3 | desktop | Business, 45-54, Low, Balanced | / → /sample → /sample/report.html | 16 | 0 | left · not for me | A sample report to judge the output., The sample report to judge outpu | The page is text-heavy and abstract; I had to hunt for the button. |
| Product people at software companies | r2·4 | desktop | Business, 65-74, Moderate, Deliberate | / → /sample → /sample/report.html | 5 | 0 | left · unsure | It is all text — no video anywhere, and a video would explain this in thirty seconds | |
| Product people at software companies | r2·5 | desktop | STEM, 5-12, High, Quick | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Opening a sample report to see what a finished one actually looks like | Words like 'funnel chart' and 'below the fold' mean nothing to me |
| Product people at software companies | r2·6 | desktop | Business, 75-84, Low, Snap decisions | / → /sample → /sample/report.html | 10 | 0 | left · not for me | Sample report, The sample report | The top menu items are small and fiddly to hit. |
| Product people at software companies | r2·7 | desktop | Social science, 18-24, Low, Snap decisions | / | 1 | 0 | left · not for me | Too much text at first glance | |
| Product people at software companies | r2·8 | desktop | Business, 35-44, High, Agonizes | / → /sample → /sample/report.html | 18 | 0 | left · not for me | The free Quick read — but only if the sample report convinces me first | No video anywhere — it's all text, and there's a lot of it below |
| Product people at software companies | r2·9 | desktop | Business, 45-54, Very high, Snap decisions | / → /sample → /sample/report.html | 22 | 0 | left · not for me | Read a sample report, to judge quality before signing up., Open the sa | The 'USABILITY STUDIES WITH SIMULATED PARTICIPANTS' label is tiny and easy to miss, yet it |
| Product people at software companies | r2·10 | mobile | Business, 75-84, High, Balanced | / → /sample → /sample/report.html | 13 | 0 | left · not for me | A sample report, to judge whether the simulated feedback is worth anyt | A solid wall of text on a phone with no video anywhere, which is how I'd rather take this |
| Product people at software companies | r2·11 | desktop | STEM, 5-12, Moderate, Agonizes | / → /sample → /sample/report.html | 5 | 0 | left · not for me | the sample report | The paragraph under the headline is a big block of text, I skipped most of it |
| Product people at software companies | r2·12 | mobile | STEM, 5-12, Very high, Snap decisions | / → /sample → /sample/report.html | 5 | 0 | left · not for me | Read a sample report, Reading a sample report | Small text and small buttons on mobile make it fiddly to read and tap. |
| Product people at software companies | r2·13 | desktop | Humanities, 35-44, Moderate, Snap decisions | / → /sample → /sample/report.html | 13 | 0 | left · not for me | the sample report, the sample report itself, to judge whether the find | The tiny text labels like 'Usability studies with sim' and 'The gap' are hard to read. |
| Product people at software companies | r2·14 | desktop | STEM, 45-54, Moderate, Balanced | / → /sample → /sample/report.html | 17 | 0 | left · not for me | the sample report, the sample report itself | The small uppercase label at the top is hard to read. |
| Product people at software companies | r2·15 | mobile | Business, 45-54, High, Balanced | / → /sample → /sample/report.html | 9 | 0 | left · not for me | Read a sample report, Open the sample report | It's a text-heavy page and I'd rather watch a short video. |
| Product people at software companies | r2·16 | mobile | Interdisciplinary, 5-12, Moderate, Snap decisions | / → /start | 2 | 0 | left · not for me | The small text in the top label and tiny header controls are fiddly on my phone. | |
| Product people at software companies | r2·17 | desktop | STEM, 13-17, Low, Deliberate | / → /sample → /sample/report.html | 12 | 0 | left · not for me | The sample report, to judge if the output is real insight or generic f | The big paragraph of explanation is dense and I had to read it twice to catch that the vis |
| Product people at software companies | r2·18 | desktop | Interdisciplinary, 18-24, Moderate, Balanced | / → /sample → /sample/report.html | 4 | 0 | left · unsure | Reading a sample report to judge if this kind of testing is worth anyt | 'Simulated participants' is a confusing phrase — I had to read twice to get that they aren |
| Product people at software companies | r2·19 | desktop | Interdisciplinary, 13-17, Moderate, Deliberate | / | 1 | 0 | left · not for me | The opening paragraph is dense for a first skim. | |
| Product people at software companies | r2·20 | desktop | Business, 35-44, Very high, Balanced | / → /sample → /sample/report.html | 3 | 0 | left · for me | The free Quick read on my own site, if the sample looks credible., The | No video, the page is a wall of text; the small text is easy to miss. |
| Product people at software companies | r2·21 | mobile | Business, 45-54, High, Balanced | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Read a sample report, Open the sample report | The tiny 'Menu' and logo are fiddly to tap on mobile |
| Product people at software companies | r2·22 | desktop | Business, 13-17, Very high, Snap decisions | / → /sample → /sample/report.html | 3 | 0 | left · not for me | Reading a sample report to decide if the reports are worth anything., | 'Quick read' and 'simulated participants' are never explained in plain words |
| Product people at software companies | r2·23 | desktop | Interdisciplinary, 18-24, High, Deliberate | / → /sample → /sample/report.html | 21 | 0 | left · not for me | A sample report, to judge whether the findings are specific enough to | 'Simulated participants' is doing a lot of work in the subheading — I had to read twice to |
| Product people at software companies | r2·24 | desktop | STEM, Under 5, High, Balanced | / → /sample → /sample/report.html | 20 | 0 | left · not for me | Read a sample report, to see if the output is useful, Read a sample re | No video on the first screen, just a wall of text |
| Product people at software companies | r2·25 | desktop | Social science, 25-34, High, Agonizes | / → /sample → /sample/report.html | 5 | 0 | left · not for me | Read a sample report | The page is a wall of text, and I nearly left before reaching the buttons. |
| Product people at software companies | r2·26 | mobile | STEM, 45-54, Moderate, Snap decisions | / → /docs | 10 | 0 | left · not for me | Read a sample report, Read a sample report like this one end to end | The top nav and labels are tiny on my phone. |
| Product people at software companies | r2·27 | desktop | STEM, 55-64, Moderate, Deliberate | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Read a sample report, Run a Quick read on a site if I had one | The page is text-heavy, no video, and the simulated-participants claim makes me pause. |
| Product people at software companies | r2·28 | desktop | Social science, 55-64, Moderate, Balanced | / → /sample → /sample/report.html | 30 | 0 | cut off · unsure | The sample report, to judge the quality before signing up for anything | The headline says 'Watch simulated visitors' but there's nothing to watch here — no video |
| Product people at software companies | r2·29 | mobile | Humanities, 45-54, Moderate, Snap decisions | / | 1 | 0 | left · not for me | Says 'Watch' but there is no video | |
| Product people at software companies | r2·30 | desktop | STEM, Under 5, Low, Snap decisions | / → /start → /method | 3 | 0 | left · not for me | The free Quick read, if I had a live site. | No video or demo — just a wall of text, which puts me off. |
| Product people at software companies | r2·31 | desktop | Arts, 45-54, High, Balanced | / → /sample → /sample/report.html | 15 | 0 | left · not for me | The free sample report, to judge the quality of the output before cons | The kicker line 'Usability studies with simulated participants' is set so small it reads l |
| Product people at software companies | r2·32 | desktop | Social science, 35-44, Moderate, Quick | / → /sample → /sample/report.html | 11 | 0 | left · not for me | The free sample report, to judge the quality of the output., The full | The subtitle line about 'simulated participants' is tiny, and the biggest claim is buried |
| Product people at software companies | r2·33 | mobile | Social science, 85+, Moderate, Deliberate | / | 2 | 0 | left · not for me | Too much text with no video | |
| Product people at software companies | r2·34 | mobile | Humanities, Under 5, None, Agonizes | / | 1 | 0 | left · not for me | Too much dense text on mobile | |
| Product people at software companies | r2·35 | desktop | STEM, Under 5, Very high, Agonizes | / → /sample | 2 | 0 | left · not for me | The phrase 'simulated participants' makes me doubt the results. | |
| Product people at software companies | r2·36 | desktop | Business, 13-17, Very high, Agonizes | / → /sample → /sample/report.html | 3 | 0 | left · unsure | A sample report, to see how the feedback is written, The sample report | It's a wall of text with no video, and I'd rather watch something |
| Product people at software companies | r2·37 | desktop | Business, 35-44, Very high, Quick | / → /sample → /sample/report.html | 20 | 0 | left · not for me | Read a sample report, Reading a sample report end to end to judge the | The small eyebrow text 'USABILITY STUDIES WITH SIMULATED PARTICIPANTS' is hard to read. |
| Product people at software companies | r2·38 | mobile | STEM, 65-74, Moderate, Agonizes | / → /start → /method → /sample → /sample/report.html | 22 | 1 | left · not for me | Read a sample report, Quick read, because it costs nothing to see whet | The text is small on my phone, and the two buttons are close together. |
| Product people at software companies | r2·39 | mobile | STEM, Under 5, Very high, Balanced | / → /sample → /sample/report.html | 30 | 0 | left · not for me | Read a sample report — I want to see the actual output before giving t | The small text ('Usability studies with sim', 'The gap') is hard to read on a phone |
| Product people at software companies | r2·40 | mobile | Humanities, 13-17, Moderate, Agonizes | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Read a sample report | The little blue label 'USABILITY STUDIES WITH SIMULATED PARTICIPANTS' is tiny and I almost |
| Product people at software companies | r2·41 | mobile | Arts, 55-64, High, Snap decisions | / → /sample → /sample/report.html | 17 | 0 | left · not for me | Read a sample report — to check the quality of the output before commi | The 'ProductTex' logo and 'Menu' are tiny and fiddly to hit on a phone |
| Product people at software companies | r2·42 | desktop | Business, Under 5, Moderate, Quick | / → /sample → /sample/report.html | 4 | 0 | left · not for me | The sample report, to judge whether the output is real or just generic | Way too much text, no video or visuals anywhere on this screen |
| Product people at software companies | r2·43 | desktop | Humanities, 35-44, Moderate, Snap decisions | / | 1 | 0 | left · not for me | The page is a wall of text with no video, which put me off immediately. | |
| Product people at software companies | r2·44 | desktop | Business, 85+, Moderate, Deliberate | / → /sample → /sample/report.html | 17 | 0 | left · not for me | Read a sample report, Reading a full sample report end to end to judge | The headline says 'Watch' — but I want to read, not watch, and there is no video in sight. |
| Product people at software companies | r2·45 | desktop | Arts, 45-54, Moderate, Deliberate | / → /sample → /sample/report.html | 13 | 0 | left · not for me | The sample report, if it shows real, specific findings rather than gen | A wall of text with no visual — no screenshot, no chart, no example output on the first sc |
| Product people at software companies | r2·46 | desktop | Humanities, 35-44, Low, Snap decisions | / → /sample → /sample/report.html | 6 | 0 | left · not for me | The sample report — to judge if the output is concrete enough to act o | The small subheading and footer text look hard to read on a low-bandwidth screen. |
| Product people at software companies | r2·47 | mobile | Business, 5-12, Moderate, Agonizes | / → /sample → /sample/report.html | 8 | 0 | left · not for me | Read a sample report, Opening a sample report to see what one actually | The small text at the top and the tiny labels are hard to read on a phone. |
| Product people at software companies | r2·48 | desktop | STEM, 35-44, High, Snap decisions | / → /sample → /sample/report.html | 21 | 0 | left · not for me | The sample report first; possibly the free Quick read afterwards, The | The little label above the headline is tiny and almost disappears |
| Product people at software companies | r2·49 | desktop | Humanities, 45-54, High, Quick | / → /sample → /sample/report.html | 5 | 0 | left · not for me | The page is very text-heavy with no video. | |
| Product people at software companies | r2·50 | desktop | STEM, 5-12, Moderate, Deliberate | / → /sample → /sample/report.html | 17 | 0 | left · not for me | Opening a real report and reading the summary section to see if the fi | 'Simulated participants' is doing a lot of work in that headline — I had to read twice to |
| Product people at software companies | r2·51 | desktop | Interdisciplinary, 55-64, Low, Balanced | / → /sample → /sample/report.html | 14 | 0 | left · not for me | A free read on my own site, if the sample report shows something I cou | Some of the top-nav links ('Docs', 'ProductTex') look tiny and fiddly to hit |
| Product people at software companies | r2·52 | mobile | Social science, 45-54, Low, Agonizes | / → /sample → /sample/report.html | 4 | 0 | left · not for me | Read a sample report, Open the sample report | All text and no video, which puts me off. |
| Product people at software companies | r2·53 | desktop | Business, 25-34, Moderate, Balanced | / → /sample → /sample/report.html | 30 | 0 | cut off · for me | Read a sample report, Read the full quick-read report to see if the fi | Some of the body text is small and greyed, which made me squint. |
| Product people at software companies | r2·54 | mobile | Social science, 13-17, Moderate, Quick | / → /sample → /sample/report.html | 3 | 0 | left · not for me | Lots of text and no video, which puts me off | |
| Product people at software companies | r2·55 | mobile | STEM, Under 5, High, Deliberate | / → /sample → /sample/report.html | 24 | 0 | left · not for me | Read a sample report, Open the sample report, to see the real output b | The small text at the top and tiny buttons are hard to read on a phone. |
| Product people at software companies | r2·56 | mobile | Interdisciplinary, 18-24, Moderate, Balanced | / → /sample → /sample/report.html | 4 | 0 | left · not for me | the 'if you change one thing, change this' recommendation format — it' | The page is quite text-heavy on mobile, and the logo and menu are tiny. |
| Product people at software companies | r2·57 | desktop | Interdisciplinary, 25-34, High, Balanced | / → /sample → /sample/report.html | 11 | 0 | left · not for me | The sample report, to see if the results are useful., The sample repor | The page is a wall of text, no video, and some text is very small. |
| Product people at software companies | r2·58 | mobile | Business, 35-44, High, Agonizes | / | 7 | 0 | left · not for me | the control-group comparison — running a test against people with no r | No video anywhere, just a wall of text |
| Product people at software companies | r2·59 | desktop | Interdisciplinary, 65-74, Moderate, Snap decisions | / → /sample → /sample/report.html | 4 | 0 | left · unsure | Read a sample report, Open the sample report | The page is text-heavy and the top navigation labels are small and fiddly. |