Android vitals and ranking: what happens past 0.47%.

Crash and ANR rates aren't just engineering metrics. Above Google's thresholds, your app gets harder to find. Here is how to read them and how to get back under.

By Arsh Goyal, Growth Engineer 3 min read
Android vitals and ranking: what happens past 0.47%.

I came to app growth from engineering, and the first thing I noticed is how often ASO conversations skip the build. Teams rewrite the title and redesign screenshots while the app quietly crashes on a fifth of their users' phones. Google Play notices that before the marketing team does.

On a healthcare product I worked on, slow data fetches were causing drop-offs at the booking step. We stopped treating latency as an engineering detail and put it on the product scorecard next to conversion. Fetch time fell 40%. Android vitals deserve the same treatment: they are a product metric with a direct line to your store ranking.

The two lines that matter

Play Console's Android vitals documentation sets the bad-behaviour thresholds:

MetricOverall thresholdPer device model
User-perceived crash rate1.09% of daily users8% of daily users
User-perceived ANR rate0.47% of daily users8% of daily users

Google states that exceeding these thresholds reduces your app's discoverability on Google Play. "User-perceived" is the important phrase: these count crashes and freezes a person actually experiences while using the app, not every background failure.

Why the per-device line catches people

The overall rate can look healthy while one popular budget phone is a disaster. In India and other emerging markets, a single low-RAM model can carry a meaningful share of your installs. Fix that model, and the overall number follows.

Gyan Studios learned this at launch in 2020. Its first kids trivia game shipped at about 45MB, which sounds light. But roughly 38% of its target users ran phones with under 1GB of RAM and Android versions below 7.0, and the app crashed on them consistently. First-week reviews were one-star complaints about performance, which suppressed search ranking exactly when it mattered most.

The fix was a full rebuild in 11 days: compressed assets, code-driven animations instead of frame-based ones, an APK under 22MB, and Android 6.0 as the minimum version so phones three to five generations old were covered. After the relaunch, crash reports fell 91%, the ratings curve reversed within ten days, and the new build crossed 10,000 installs in its first week. (case study)

How to triage

  1. Open Android vitals and sort by device. Find any model above 8%. That is usually the fastest win.
  2. Group crashes by stack trace, not by report count. A handful of clusters normally explain most of the rate.
  3. For ANRs, look at the main thread. Android raises an ANR when an app doesn't respond to input for about five seconds. Disk access, network calls and heavy parsing on the main thread are the usual causes.
  4. Test on the low end. Use the oldest, cheapest device your audience actually owns, not the one on your desk.
  5. Watch the 28-day trend after each release. Vitals are measured over a rolling window, so recovery takes a few weeks to show.

What healthy looks like

On a 440K MAU client app, the crash rate sat at 0.23% and the ANR rate at 0.13%, comfortably under both lines. That headroom is what let listing work compound: a 31.7% custom listing conversion and steady rating gains were never undone by a stability penalty. (case study)

The free ASO Score checker includes vitals as one of its twelve Google Play checks, because a listing can't outrank the build behind it.

Questions people ask

What is a good ANR rate on Google Play?

Below 0.47% of daily users overall, and below 8% on every device model. Aim well under the line, because one bad release can double the rate.

Do Android vitals affect Play Store ranking?

Yes. Google says apps that exceed the bad-behaviour thresholds have reduced discoverability on Google Play.

How long does it take to recover after fixing crashes?

Vitals use a rolling 28-day window, so expect a few weeks before the rate falls fully below the line. Ratings can turn faster once new reviews stop reporting the crash.

Want this mapped onto your product?

See how our ASO work runs, or send your store link and a founder will reply personally.