How to run an ASO audit

The five passes

1. 1

   Baseline the metrics

   Pull 90 days of impressions, product page views, conversion rate, install volume and average rating from App Store Connect and Play Console. These are the numbers every finding will be argued against.

2. 2

   Audit the indexed fields

   Check the app name, subtitle, Apple keyword field, Play short and full description against their limits. Note unused characters, repeated words and terms nobody searches for.

3. 3

   Audit the creative

   View the listing at thumbnail size on a phone. Judge whether the icon is distinct, the first two screenshot captions are readable, and the video says what the app does within three seconds.

4. 4

   Compare with the top five

   For each priority term, read the listings of the five apps above you. Record the words they all use that you do not, and how their screenshot messaging differs.

5. 5

   Rank the findings

   Write each gap as a change, an expected metric it moves, and an owner. Sort by expected impact over effort, not by the order you found them.

What to score, and what good looks like

Reading competitor listings properly

Competitor auditing goes wrong when it turns into copying. The useful signal is not what one strong competitor does; it is what all five of them do that you do not. Terms that appear in every top listing are terms the category treats as load-bearing, and their absence from yours is a real gap.

The second signal is positioning divergence. If every competitor leads its screenshots on the same benefit, the opportunity is often to lead on a different one — the crowded message is already served, and differentiation converts.

Turning findings into shippable work

1. Group findings into metadata, creative, ratings and technical.
2. Assign each a single metric it should move — impressions for metadata, conversion rate for creative.
3. Ship metadata and creative in separate releases so attribution survives.
4. Re-read the baseline metrics two weeks after each release, not two days.
5. Keep the audit document and diff it next quarter; the delta is the report.

Worked example: two hours on a listing nobody had audited

An education app had never been audited. The team assumed the problems were creative, because that is what people notice. Two hours of structured passes found the actual distribution: one metadata issue worth more than everything else combined, two creative issues worth having, and a review-health problem nobody had connected to conversion.

The metadata pass found the app name spending eighteen of its thirty characters on the institution's full legal name. The searchable term arrived after the truncation point, which on a phone means it was effectively absent from the most heavily weighted field on the listing.

The review pass found the connected problem: a 3.6 average with strong recent sentiment, held down by a cluster of one-star reviews about a login bug fixed the previous year, all unanswered. Replying to them and prompting satisfied users at the right moment lifted the average over the following quarter, which raised conversion on every impression the app was already earning — the compounding effect that makes review health an ASO issue rather than a support one.

What makes an audit useless

An audit that produces no shipped changes was usually built wrong:

• Listing findings without ranking them by expected effect, so everything looks equally urgent.
• Auditing creative first because it is visible, and metadata last because it is not.
• Producing a document rather than a queue of specific field-level changes.
• Comparing against category leaders whose brand demand you cannot replicate, instead of the set you actually compete with.
• Failing to record the pre-audit baseline, which makes every later result unattributable.
• Auditing once rather than on a cadence — a listing decays because competitors move, not because anything broke.