Lift-Test Reader

Interpreted

Finished a test? Drop in what your platform gave you. The tool reads the result, recomputes iROAS and the power the test actually had, and flags anything that can't be trusted.

Most platforms (Meta/Google) hand back a lift % and a confidence interval — use this.

What your platform reported

Reported liftRelative, in % (e.g. 12)

My result includes a confidence interval

CI low %

CI high %

Baseline conversionsIn the test window

AOVAvg order value

SpendOn this test

Baseline rate %For power

Audience per armFor power

The read

Significant

Reported lift

12.0%

CI 4.8% – 19.2%

iROAS (recomputed)

0.34x

$9,600 ÷ $28,000 spend

Incremental conversions

120

$9,600 incremental revenue

Retrospective power

90.1%

Power the test actually had

Honesty flags

Interpreted

The confidence interval (4.8% to 19.2%) excludes zero, so the lift is statistically significant.
Retrospective power was 90.1% — the test was well-powered to detect a lift of this size.

Report & save

Generate a plain-language write-up to share, and save the result to your history.