Most indie developers look at one number after an App Store update: total downloads.

That is the wrong number to lead with.

Downloads are a lagging indicator. By the time downloads move, three other things have already happened, or failed to happen. If you read only the final output, you cannot diagnose what’s working or decide what to do next.

After six weeks of metadata tests, screenshot updates, and ratings prompt work, here’s how to actually read the data.

The ASO signal stack

App Store performance reads in a specific order. Think of it as a funnel with four layers:

  • Impressions: how many times your app appeared in search results or browse surfaces.
  • Product page views: how many users tapped into your listing.
  • Keyword rank movement: whether your positions for target terms are holding, rising, or falling.
  • Ratings velocity: whether review volume and score trend is supporting or undermining conversion.

Each layer informs the next. If impressions are flat but downloads fall, the problem is on the product page. If impressions rise but page views do not follow, users are seeing your app in search and choosing not to tap in. If ranks improve but downloads do not, the product page or rating volume is the ceiling.

Reading them in order is how you diagnose rather than guess.

Layer 1: Impressions - what they mean and what they don’t

Impressions measure visibility: how often Apple showed your app on a search result page, a featured shelf, or a Today tab placement.

For Pi Digits over the last 90 days: 2,758 impressions. For a small indie app in a mid-difficulty niche, that is a real baseline. It is not zero, and it is not dominant.

The useful question is not “are my impressions high?” It is “are my impressions growing after my last metadata update?” If you shipped a metadata test targeting a new keyword cluster and impressions did not move in the following 3 weeks, the terms in that cluster are not generating search discovery. Either the new terms do not match common queries, or the app is not indexing for them yet.

Impression data lags by up to 3 to 5 days in App Store Connect. Do not read a same-week change as a signal.

Layer 2: Product page views and conversion

Product page views tell you how many users cared enough to tap in after seeing your app. The gap between impressions and page views is your tap-through rate.

For Pi Digits: 192 page views from 2,758 impressions, a 7.0% tap-through rate.

That number deserves attention. What causes a user to skip your listing after seeing it? Usually: the icon, the app name, or the subtitle does not immediately match what they searched for. If your tap-through is below 5% in a category where competitors average 15%, the issue is the listing itself, not the ranking.

There is a subtler data point here. Pi Digits recorded 224 downloads from 192 product page views. Downloads that exceed page views signal that a meaningful share of installs came directly from search: users tapped Install without visiting the full product page. That is a strong search-intent match signal. When a user converts without reading your description or scrolling through screenshots, the keyword, app name, and icon together were enough. That is the result of a tight, specific metadata bet.

Layer 3: Keyword rank movement

Rank movement is the leading indicator that tells you whether your metadata changes registered with Apple before downloads move.

For Pi Digits, the current Marteso pull shows 138 ranked keywords in the US. The top cluster is tightly themed: pi digit challenge game at #1 (popularity 58, difficulty 63), pi number challenge at #2 (popularity 55, difficulty 62), pi challenge at #3, and pi digits quiz at #3. That consistency means Apple has a clear, stable model of what this app is.

The same pull shows what Apple has not assigned rank to: brain workout (popularity 99, difficulty 93), memory games for adults (popularity 97, difficulty 95), memory trainer (popularity 96, difficulty 88), and math games (popularity 93, difficulty 100). These are the terms many apps in this category compete for. Pi Digits does not rank for any of them.

That is not a failure. It is a roadmap. The app owns a niche. The question is whether to defend and deepen that niche, or attempt a first step toward an adjacent high-volume term with a targeted test.

Marteso’s keyword tracking is what makes this visible. Without a persistent ranking layer, you are checking position manually across dozens of terms and missing movement patterns. The value is not any single snapshot. It is the trend line over 3 to 6 weeks that shows whether a test actually shifted Apple’s model.

Layer 4: Ratings velocity

Ratings affect product page conversion, and product page conversion affects whether your ranked keywords hold.

Pi Digits currently has a 5.0 average from 2 reviews over 90 days. That is a good score from a small sample. The practical question: is that sample large enough to convert skeptical users?

For most indie apps, the ratings threshold that visibly reduces conversion friction is somewhere between 20 and 50 reviews at a credible score. Below that, users browsing a competitive keyword see the rating count before they read the score, and small sample sizes trigger doubt.

Velocity matters more than the static count. Two new reviews in 90 days is a different story than 20 reviews in 30 days. Ratings velocity is Apple’s signal that an app has an active, engaged user base, and that signal contributes to whether rankings in a more competitive tier hold over time.

If your keyword ranks are stable but installs have plateaued, ratings volume is often the invisible ceiling.

Reading the full stack together

For Pi Digits after 90 days, the picture looks like this:

  • Impressions are real and growing, driven by a specific keyword cluster Apple has indexed.
  • Tap-through is solid at 7%, and direct-from-search installs suggest the metadata is doing its job on a narrow, intent-matched set of queries.
  • Keyword ranks are strong in niche territory: 138 ranked US terms, many at #1 to #3, but absent from high-volume adjacent categories.
  • Ratings are a near-term ceiling: 2 reviews will not hold a ranking in a more competitive tier, even with clean metadata.

That tells you three things clearly: what’s working, what the ceiling is, and what the next test should be. In this case, the answers are metadata specificity and niche dominance, ratings volume, and one step toward a higher-volume adjacent term with a ratings prompt strategy running in parallel.

This is what an ASO review at the end of a 60-day cycle should produce. Not just a download count, but a diagnostic that tells you what to do next.