Re‑test
Proof that a fix worked.
A re‑test runs the same prompts on the same surfaces under the same conditions, so the before and after of a fix can be compared one to one.
Same conditions, before and after
A re‑test repeats the original measurement exactly, so nothing changes between before and after except your fix. That is what makes the comparison trustworthy.
Same prompts
The exact prompts from the original measurement run again, unchanged, so the comparison is one to one.
Same surfaces
The re‑test uses the same AI surfaces in the same market and language, so nothing but your change is different.
Same conditions
Market, language and setup match the original run. When something cannot be matched, rankyn labels the comparison instead of presenting it as clean.
Original measurement
Your fixThe only change
Re‑test
Before and after, side by side
rankyn holds the before snapshot and the re‑test result next to each other. You see whether the recommendation rate moved, and by how much.
ChatGPTNot namedRecommended
AI OverviewRecommendedRecommended- AI ModeNot namedRecommended
PerplexityNot namedNot named
1 of 43 of 4surfaces recommend your product
Re‑tests in every plan
A re‑test repeats a measurement you already ran, after a fix, under the same conditions. Every plan includes a number of them each month.
Included each month
- Starter
- 2re‑tests a month
- GrowthRecommended
- 8re‑tests a month
- Scale
- 20re‑tests a month
Prove the change worked.
Make a fix and re‑test it. Start with a free audit to find your first change, then show it moved the answer.
Questions about re‑tests
What a re‑test is, what makes it valid, and how many your plan includes. Can’t find your answer? Write to hello@rankyn.ai.
What is a re‑test?
A re‑test is when rankyn runs the same measurement again under the same conditions. rankyn repeats the same prompts on the same surfaces, so the before and after of a fix can be compared one to one.
What makes a re‑test valid?
A rankyn re‑test is valid when it matches the original run. rankyn keeps the same prompts, surfaces, market and language, so only your change differs. When something cannot be matched, rankyn labels the comparison instead of hiding it.
How many re‑tests do I get?
rankyn includes a set number of re‑tests in every plan. The recommended plan includes 8 a month.
What does a re‑test prove?
A rankyn re‑test shows whether the answer changed after your fix. The before and after sit side by side, so you see whether the recommendation rate moved, and by how much, on each surface.
When should I run a re‑test?
Run a rankyn re‑test once your fix is live and AI surfaces have had a chance to see it. rankyn runs it under the same conditions as the original measurement and shows both results together.