Playbook: ASO Optimization¶
A 4-week route to more installs from App Store and Google Play search, then a monthly loop to keep improving. It's for apps that are already live. Not published yet? Do listing-app-store / listing-google-play (Level 1) first, then come back.
The funnel you're fixing¶
ASO moves three numbers. Know which one is weakest before you change anything.
| Step | App Store Connect (App Analytics) | Google Play Console (Store listing acquisition) | Fixed by |
|---|---|---|---|
| People see you in search | Impressions | Store listing visitors (by traffic source: search) | Keywords, title, subtitle |
| People open your page | Product page views | Store listing visitors | Title, icon, first screenshots, rating |
| People install | Conversion rate | Store listing conversion rate | Screenshots, rating, reviews |
Metric names change between console versions. Check the current help pages: App Store Connect Help · Play Console Help.
The route¶
| Week | Goal | Skill | You end the week with |
|---|---|---|---|
| Week 0 (1 h) | Baseline, so you can tell if anything worked | none, see Measure | Last 28 days of the funnel numbers, saved |
| Week 0 (optional) | Know what users love and hate | research-review-mining → case-study-competitor-teardown | User phrases, top complaints, what top competitors show |
| Week 1 | Rank for words people actually search | aso-keyword-research | Scored keyword sheet, keyword field |
| Week 1 | Turn keywords into a name that ranks and sells | aso-title-subtitle-optimization | Title, subtitle / short description; update submitted |
| Week 2 | Make the page convert | aso-screenshot-strategy | New screenshot set + one A/B test running |
| Week 3 | Ratings that help, not hurt | aso-review-and-rating-strategy | Review prompt live, 1–3★ reviews answered |
| Week 4 | Read the results, pick the next change | the monthly loop | A one-page ASO log with before/after numbers |
Where to start if you only have one week: find your weakest step in the funnel table. Few impressions → keywords. Views but few installs → screenshots. Rating under 4★ → reviews first, because a low rating drags down every other fix.
Measure the baseline (week 0)¶
- Pick a 28-day window that ends before you change anything. Avoid launch spikes, sales and featuring.
- Write down, per store and for your top country: impressions (or search visitors), product page views, conversion rate, installs from search, average rating, number of new ratings.
- Write down your rank for your 5 most important keywords (search them on a phone, signed out if you can, or use an ASO tool you already have).
- Save it as
growth/aso-log.md. Every later change gets a line in the same file.
Store rules that shape the plan¶
- App Store: title, subtitle and keyword field change only with a new app version you submit for review. Promotional text can change any time without a new version. Bundle your keyword and title changes with a release, and plan them a week ahead. App Store Connect Help.
- Google Play: you can edit the store listing without a new app release. Changes still go through review. Play Console Help.
- Both: don't stuff competitor brand names or misleading claims into metadata. It's the most common rejection. App Review Guidelines · Developer Program Policy.
- Tests: Apple's Product Page Optimization and Google's store listing experiments test screenshots and graphics on real traffic. Use them for week 2 instead of guessing.
The monthly loop¶
After week 4, run this every month. It takes about an hour.
- Read (15 min). Compare the last 28 days with your baseline in
growth/aso-log.md. Same window length, same countries. - Find the weakest step (5 min). Use the funnel table. One step, not three.
- Change one thing (30 min). Re-run only the skill for that step. One change per cycle, or you won't know what worked.
- Log it (5 min). Date, what changed, why, and which number you expect to move.
- Wait. Give a metadata change at least 2–4 weeks before judging it; rankings settle slowly. Let A/B tests run until the console reports a result.
Illustrative example. A made-up log line, to show the format. These are not real numbers.
| Date | Change | Why | Before (28 d) | After (28 d) | Keep? |
|---|---|---|---|---|---|
| 2026-03-02 | Subtitle: "Water tracker" → "Drink reminder & water log" | "reminder" had more searches in keyword sheet | 12,400 impressions · 3.1% CVR | 15,900 impressions · 3.0% CVR | Yes |
Rules of the road¶
- One change at a time. Two changes in one release = no idea which one worked.
- Compare like with like. Same 28-day length, same countries, no launch spikes or holidays in only one window.
- Fix the product before the page. If the top review complaint is a bug, no screenshot will outsell it.
- Localize last, not never. Once your main language converts, repeat weeks 1–2 for your next biggest country.
With an AI agent¶
Installed the skills? Tell your agent: "Follow the aso-optimization playbook for my app. Start with the week 0
baseline and keep growth/aso-log.md up to date."
Using a chat AI? Open each skill in order and paste its prompt, then paste your funnel numbers when you reach the loop.
Keep ASO in your repo¶
Save each output next to your code: growth/keywords.md, growth/listing.md, growth/screenshots.md,
growth/aso-log.md. Next month your agent can read the log and tell you what to change.