Use the same app, keyword, country and data source. Check the observation dates, separate missing data from measured positions, then investigate what else changed before editing your listing.
1. Compare the same app, term and country
A US result and a Spanish result answer different questions. Keep the storefront fixed when comparing positions, even when the search term is identical.
Record the app ID, exact keyword, country, source and observation date. Storewake tracks keywords by app and storefront; its collected public search rankings do not provide separate iPhone and iPad positions. A manual device search may not be directly comparable.
2. Check what was actually observed
Rank 12 becoming rank 28 is a measured decline of 16 positions. A missing rank is not rank zero and should not be converted into an arbitrary last place. Check collection status and the returned result set before interpreting it.
- Are both snapshots present, successful and dated?
- Did the keyword start tracking only recently?
- Was monitoring paused, or did a provider or usage limit prevent collection?
- Are you comparing consecutive observations or dates with a large gap?
Storewake schedules snapshots weekly on Free and daily on Grow. Manual refreshes remain subject to usage limits. Compare the dates actually returned rather than assuming every calendar day has data.
3. Put listing edits beside the ranking history
Record when an edit became visible, which localization it affected and what changed. Include releases and any relevant campaign changes in your own timeline. If you did not record the earlier metadata, say so instead of reconstructing it from memory.
Ask whether the decline began before or after the edit. A change that follows an edit may justify a hypothesis, but timing alone does not prove that the edit caused it. Avoid reversing several changes in response to a single observation.
4. Look for evidence beyond one keyword
Compare related tracked terms in the same country. Is the movement isolated or shared? Review relevant competitors, while distinguishing today’s search results from any historical evidence you actually have.
Use available acquisition, conversion and revenue data to test whether a ranking movement coincides with a business change. Rankings alone do not measure downloads. Connected RevenueCat snapshots can add revenue context, but do not attribute a subscription to a particular keyword.
Apple popularity may be unavailable for a term. An unknown score is not proof of low demand. Keep measured ranks, missing values and estimated difficulty separate in your analysis.
5. Write one testable hypothesis
Choose a change small enough to evaluate and keep a record before you make it. For example:
Observation: [keyword] declined from [rank] to [rank] in [country] between [dates]. Hypothesis: [possible explanation, not a conclusion]. Evidence for it: [dated observations]. Missing evidence: [what still needs checking]. Planned change: [one specific listing change]. Evaluation: [keywords, country, business metric and review date].
Choose the review window based on your snapshot cadence and available traffic. Preserve the earlier observations, note other changes during the test, and avoid claiming an uplift that the data cannot isolate.
Use this prompt with your connected agent
Investigate ranking drops for my app in [country]. Start with saved keyword history and report exact dates. Separate measured declines, missing ranks and stale snapshots. Compare related terms and available competitor evidence. Ask me for the timing of listing edits if it is missing. Return observations, hypotheses and one proposed experiment. Do not claim causality or change tracking without my approval.
The goal is a decision you can explain: what moved, how reliable the comparison is and what you will test next.