Free audit
Which of your A/B winners were real?
Every test's winner in your export is re-checked from its own counts, the way a statistician would. Below is a worked example on real public tests. Drop in your own export to see yours.
Check your own export
A CSV with one row per variant: the test or campaign, the variant, how many it was sent to and how many clicked. Message text is optional.
Runs in your browser. Your file is never uploaded.
| Variant A | Variant B |
|---|
Could a back-test on these tests give a clear answer?
| If AheadWise's skill on your copy is | Hit rate, all judged | Hit rate, clear winners | Chance it shows skill |
|---|
What do your tests say together?
| Versions with | Tests | Rate with it vs without | Pointing that way | Verdict |
|---|
Plan your next 90 days of tests
| Test this change | Tests | Why |
|---|
Uses your file as history: how much your versions really differ, and the changes your tests already made. Change one countable thing per test and keep the rest the same, so the results pool into lessons. Planning estimates, not promises.
Were your real winners predictable?
Send the tests file for a free blind back-test. AheadWise seals a prediction for every test before reading any result, then shows how often it was right and how often "confident" meant right.