Buyer simulation: check copy and pages before they go live.
A buyer simulation has personas from your audience read a piece of copy or a page before real buyers see it. 99takes runs it in your conversation with Claude: each persona gives a score and its main objection. The verdict follows the people your offer is built for.
Score per persona
Shortlist, not a measurement
- Headline 3
- Headline 1
- Headline 5
Six ways to check.
Compare two versions
about 73 creditsA score and an objection per persona; the verdict follows your core. For headlines, ads, sections and subject lines.
“Which headline wins over our buyers, A or B?”Test a new version of a page
price before the run99takes reads the page by URL, writes a challenger version from your audience research and shows it to you before testing your current page against it. Just the opening or the whole page. This comes before an A/B test with real traffic; it does not replace one.
“Test our landing page against a new version.”Analyse a landing page
price before the runCopy, structure and reading flow, how the page comes across on desktop and mobile, craft and, where there are competitors, a comparison with them. It ends with a plan of what to change first. More on this: landing page analysis.
“Analyse our landing page https://[your-page].”See how a page comes across
up to 12 credits, plus 2 to 3 for redesign suggestionsThe personas look at your page on desktop and on mobile: first impression, whether it looks made for people like them, how trustworthy it seems, whether they find what they are looking for. With screenshots, so you can see which page was checked, and suggestions for a redesign.
“Run a design check for https://[your-page].”Shortlist drafts
3 to 5 creditsWith five drafts, not every one needs a full run. A smaller jury sorts out which one is worth it. A shortlist, not a measurement.
“Which of these five headlines should go into the full run?”Compare several pages
price before the runSeveral pages against one audience, such as product or course pages built from one template. You get one comparison table across all pages.
“Which of our product pages is the weakest?”It reads the copy where it sits on the page. Design has its own check.
With a URL it reads the page the way visitors see it in the browser, from the headline to the small print.
| The jury judges | The jury does not judge |
|---|---|
| Headline, body copy and small print, where they sit on the page | Colours, images, typography: in an A/B test both versions look the same. The design check covers how the page comes across |
| Whether copy only wins at the edge: then you get a warning, not a win | Your market numbers: with enough traffic for clean A/B tests, they are the better source |
| Whether everyone likes it a little and nobody picks it first | The position in political copy, only whether the message lands |
From the verdict to better copy.
Have a better version written
99takes writes challenger versions from the same research, without unsupported numbers, and tests them blind against yours.
Challenger versions →Was the jury right?
Report what happened after launch. 99takes publishes a hit rate per jury from ten decidable results, with their number and a range.
Hit rate →Questions about buyer simulation
What does a buyer simulation cost?
It is billed in credits. Comparing two versions with 12 personas costs about 73 credits, the same comparison 1:1 about 19 credits. 99takes names the price before every run. A new account starts with 350 credits, valid for 30 days, without payment details.
Does it replace an A/B test?
No. An A/B test with real traffic measures what visitors do. The simulation comes before it: it shows which version is worth testing and which objections a conversion rate does not explain. Pages with too little traffic for a clean A/B test also get a verdict before launch this way.
How reliable is the verdict?
It is a prediction. Whether it holds is counted by the hit rate: after launch you report which version won in the market. 99takes only publishes a rate per jury from ten decidable results, always with their number and a range.
What do I need to get started?
An account on run.99takes.com and Claude, either claude.ai or Claude Code, where 99takes is added as a connector. Plus an audience: your own, researched with 99takes, or one of the ready-made audiences.