For founders who pay creators to post
Not every view becomes a customer. WhichPost learns which ones do, and finds the buyers links never see.
Connect Stripe and add your posts. Every creator gets a clear call, rebook, test again or pause, with a range on every result.
On a mobile app? Connect RevenueCat instead
Simulated sampleThe sample campaign at a glance
194
paying customers found from posts
Likely 185 to 203
7
creators worth rebooking
Out of 25 creators
US$4.849
of budget freed by pausing 8 creators
Their views look bought
Further out, more buyers. Simulated sample
Each dot is a creator. Further out means more paying customers per 1,000 views; a bigger dot means more views. Grey dots had views that look bought.
Point at or tap a dot to see that creator's numbers.
Why links and views fall short
Two creators with the same views can bring very different numbers of customers. WhichPost finds the ones whose posts make people buy.
Paying for views buys reach, not buyers, and some views aren't people at all. Reach is easy to count. Buyers take more to see.
WhichPost measures which creators' posts bring paying customers, with a range on every result, so you rebook the ones who sell, not the ones with the most views.
Captions can't hold one, and a creator's single bio link is shared by every brand they post for. Viewers search for you on Google or the App Store instead. Links and codes see only the few who tap or type.
How it works
Connect once and WhichPost keeps itself up to date. Setup takes about two minutes, and nothing changes in your product.
Create a restricted key in Stripe with read access, then paste it here. WhichPost reads when people start a checkout or a trial, and when they pay.
On a mobile app? RevenueCat works the same way.
What it reads
Read-only access
Paste links to the posts you paid for, from Whop, Vyro, Clipping.io or a direct creator deal, or connect TikTok accounts you own. WhichPost checks their views every hour.
What it reads
Public data only
Every creator ranked by the paying customers they brought, each with its range, and a clear call on who to book again.
What you get
A range on every result
Prefer files? Upload exports instead
The report
Paying users with their likely range, what each one cost, and what to do next. This is the sample report, the same campaign as above.
Paying users from posts
194
Likely 185 to 203
The method
A spike after a post looks like proof. Often it isn't. Our AI models read every hour of every post against everything else that moves your sign-ups, so credit lands where it was earned.
The hour, the weekday, your market's rhythm, holidays, your own growth. All of it shapes your sign-ups, and the models account for every bit, so a good Monday is never mistaken for a great post.
Posts go live together. Views turn into customers after different delays. Some views aren't people at all. The models sort out who caused what, and give each creator only the customers they brought, with a range.
Each creator's cost per paying customer is weighed against what a customer is worth to you. Rebook, test again or pause, made clear.
Spots bought views. When a post's views don't behave like real people watching, your report flags it, so you know before you pay for more.
Designed and trained in-house for this one job, before they ever meet your data. So your first report starts from experience, not a blank page. How they work is ours. What they find is yours.
What you get
Every report ends with what to do next: who to rebook, who to test again and who to pause, judged against what a paying customer is worth to you.
Rebook when even the cautious end of the range beats your cost limit. Pause when even the hopeful end misses it. Otherwise, test again.
Cost per paying user with its likely range. Dashed line: your limit, US$49,99.Simulated sample
A read-only key and the post links you already have. Nothing changes in your product, your checkout, your captions or your bios.
Views are captured every hour, so the timing of each post can do the work a link would.
The biggest posts are measured on their own, with ranges, and the report marks which ones were large enough to tell.
With Monitoring, the models rerun on fresh data every day, and you hear about it when a post starts converting.
Every result comes with its likely range. When the data can't support a call, the report says so instead of guessing.
Code redemptions are exact but incomplete: most viewers never type a code. Every report shows them next to our models' results so you can compare.
@cramclub
Privacy
WhichPost measures campaigns, not people.
It never identifies a device or a viewer. Everything runs on hourly counts, so there is nothing to ask permission for.
Keys can read, never change. Revoke them in Stripe or RevenueCat at any time and syncing stops.
WhichPost is run from the European Union and built for GDPR from day one. Where your data is processed is listed in the privacy policy.
Pricing
Every audit starts with a free data check. If your data can't support a clear answer, you aren't charged.
One campaign, analyzed and explained.
US$199
per campaign
For apps working with 10 or more creators a month.
US$79
per month, one product
For studios and agencies with several products.
US$199
per month, up to 5 products
The first 20 customers pay half price for as long as they stay.
Prices are in US dollars.
Questions
Straight answers about your Stripe key, setup and accuracy.
WhichPost accepts only restricted keys with read access and turns away full secret keys. A read-only key can't charge anyone, refund a payment or change your account. Revoke it in Stripe at any time and syncing stops.
Checkouts, customers, subscriptions, invoices, refunds and promo codes: when someone started a checkout or a trial, when they paid and how much. It keeps times, amounts, countries and Stripe's customer IDs. It doesn't keep names, emails or card details, and it never contacts your customers.
No. Paste a read-only Stripe key, or connect RevenueCat for a mobile app, then paste links to your campaign's posts. WhichPost fetches the rest and keeps it current. If you prefer files, the upload tool still works.
Each post's views arrive on their own schedule. When sign-ups or installs rise above your normal level in the hours after one creator's views grow, and not after others', that creator gets the credit. Across many posts it learns each creator's typical paying users per 1.000 views, which is the number you need for rebooking.
Every result comes with its likely range, and when the data can't support a call, the report says so instead of guessing. Single posts are measurable only when they're large, and the report marks which ones are. Where you use promo codes, the report shows them next to our models' results so you can compare. We'll publish benchmark results here once they're complete.
Stripe for a web product or RevenueCat for a mobile app, about 20 new sign-ups or installs a day or more, and at least a week of history before the campaign started. Campaigns with staggered posting times give sharper answers.
Keep them; they're exact. But creators get one bio link for every brand, captions can't hold links, and most viewers never type a code. They search for you on Google or the App Store instead, which no link sees. WhichPost shows code-confirmed customers next to its models' results so you can compare.
Before any paid audit, a free check confirms your data can support a clear answer. If it can't, you aren't charged, and the check tells you what's missing. If a paid audit finds a post made no measurable difference, that's an answer too: it tells you not to rebook.
Connect Stripe in about two minutes with a read-only key. A free data check before you pay anything.