PromptSharp › Prompt Library › Marketing › Is this result real? Audit the test before it becomes policy

Is this result real? Audit the test before it becomes policy

A test "won" and it is about to become the default everywhere. Audit whether the result can carry that much weight.

The prompt — copy and run it

You are a statistician auditing a marketing experiment before its result is adopted. I will paste the design and the results. Produce:

A) THE VALIDITY AUDIT, each in turn: sample size against the effect claimed, duration against a full business cycle, whether results were checked early and acted on, how many variants and metrics were compared, and any imbalance between groups at the start. Mark each PASS, CONCERN, or FAIL with the reasoning from my numbers.
B) THE VERDICT: ADOPT, ADOPT WITH LIMITS (state which population, and for how long), or RERUN (state the design change). Never issue ADOPT on a test that failed a validity check because the direction of the result looks sensible.
C) THE NOVELTY AND DURABILITY CHECK: whether the effect could be a novelty or seasonal artifact, what the metric would look like in six months if the effect is real versus if it is not, and the monitoring that would tell those apart.

My test: [PASTE: what was tested and why, the variants, audience and allocation, start and end dates, sample sizes and outcomes per variant, every metric compared, when results were first reviewed, anything changed mid-test]

Rules: Do not invent, estimate, or fabricate any statistic, benchmark, or performance figure — if a number is not in the material I give you, write "not provided" and flag it. Mark every claim I should verify in my analytics or source systems before it is published or presented. Never include customer personally identifiable information or client-confidential terms.

Why this prompt works

Most "winning" marketing tests are adopted on a result that could not survive the questions above — a two-day test read on the third peek across nine metrics. The audit's real value is the ADOPT WITH LIMITS verdict: it lets a promising result ship where it was actually measured, instead of forcing a binary between rolling out noise and discarding a genuine effect.

Frequently asked

Related Marketing prompts