Proof loop: one illustrative before-and-after
The tracked prompt is the report's #1 gap: "best customer analytics platform for PLG SaaS". After the fix shipped, the same prompt is re-tested daily on the same engines - the before/after lives in dated observations, not a slide.
Baseline citations0 / 12before the fix
First re-test1 / 12First example re-test
Day 148 / 12third illustrative re-test
Cadencedailypaid proof loop
Observation log
Baseline · day 0
0 / 12 observations 4 AI services x 3 samples
Prompt returned rivals only; fix brief generated the same day.
Fix shipped · day 2
-
Answer block + schema deployed to brightmetrics.example. Re-tests scheduled daily.
Re-test · day 5
1 / 12 observations Perplexity
First sampled citation appeared on one AI service.
Re-test · day 9
5 / 12 observations Perplexity, ChatGPT
Five of the twelve sampled service-and-prompt cells cited the brand; the repeat count remains small.
Re-test · day 14
8 / 12 observations Perplexity, ChatGPT, Google AI Overviews
Eight of the twelve sampled service-and-prompt cells cited the fictional brand in this deterministic example.
How to read this. Engine answers are stochastic, so every figure is a sampled observation count, never a certainty. These are deterministic example counts. Keep each date and denominator visible without inferring effect size, persistence, or causation. The public Proof Ledger documents its separate product reporting rules and limitations.
Take RankEcho with you: Full walkthrough · Intro to RankEcho · How it works · Written product guide