Experiment debrief

A test result is forgotten when the person who ran it moves on. The debrief closes that loop. When you end a test, Pertento offers a short prompt that shows the result and the hypothesis, then asks what you expected, why you think it happened and what you are doing next. Every question is optional, and ending a test never waits for it.

Pertento drafts what it can see, and you confirm or correct it. It compares a full week of website traffic against the first week, for the devices the test targeted, and flags a shift of 15% or more on tests that ran 14 days or longer. It also notes tracking gaps: days with no visitors recorded, or a visitor split that drifted from the planned weights. Your answer decides which outside factors the debrief records.

Afterwards the test has a Debrief tab, a single card led by the lesson. The facts from your statistics sit on one side (uplift, confidence, visitors and the days the test ran) and your team’s answers on the other. The headline is written from the statistics, never typed, so it cannot overstate a result. A variant that looks good alone but misses the stricter bar needed when several variants are compared is described as not by enough to call it, rather than as a win or as flat.

The tab is there on any test that has ended or had a winner deployed. Anyone with access to the test can fill it in or change it at any time, and the card shows who wrote it last and when, so a new teammate finds the lesson where the test lives.