Your 2026 App Store Optimization Checklist

Keyword research is the foundation of every ASO project, and it is also where most teams cut corners. Start by listing 30 to 50 candidate keywords that describe what your app does, the problems it solves, and the categories it competes in. Pull these from your own product copy, customer support tickets, App Store autocomplete, and competitor listings. Tools like AppTweak, Sensor Tower, and AppFollow give you search volume estimates and difficulty scores for each one.

For 2026 it matters that you separate informational queries from commercial ones. A user typing "habit tracker" is browsing, while "best habit tracker no subscription" is closer to download intent. Map your keywords to user intent and prioritize the ones with high commercial value, even if their volume is lower. Cal AI famously ranks for "calorie counter ai" rather than the generic "calorie counter" because the AI variant converts at almost twice the rate.

Build a working spreadsheet with four columns, keyword, monthly volume, difficulty, and intent. Sort by volume divided by difficulty to surface your best opportunities. Aim for a final shortlist of around 15 keywords that you can realistically rank for in the next 90 days. Save this list, you will reuse it across the next 7 steps.

Your App Store title and subtitle carry the most ranking weight of any metadata field. Apple gives you 30 characters for each, and every character should be working hard. The pattern that consistently wins in 2026 is brand plus value plus a top keyword, like "Notion, notes, docs, AI" or "Headspace, sleep and meditate".

Lead with your brand if it has any recognition, otherwise lead with the strongest keyword. The subtitle should be a benefit-focused phrase that includes a secondary keyword, not a repeat of the title. "Track habits and build streaks" works better than "The best habit tracker app" because it tells the user what the app does and includes two ranking targets.

Avoid generic adjectives like "best", "ultimate" or "amazing", they take space without adding ranking signal. Run every iteration through a character counter so you don't lose 4 characters to a forgotten emoji or non-breaking space. Refresh your title and subtitle every release cycle, frequent updates are correlated with stronger ranking gains.

Your icon shows up at 60 by 60 pixels in search results, smaller than a fingernail. Test every icon design at that size before approving it. Open your shortlist on a phone, set the screen brightness to 50%, and hold it at arms length. If you can identify the icon in under one second, it passes. If not, simplify.

The icons that win in 2026 follow three rules, single focal element, two or three colors, no thin strokes. Compare Linear's clean L glyph against any cluttered indie productivity icon and you will see the gap immediately. Avoid text on the icon unless your brand is the text, like Cash App. Apple consistently rejects icons with literal screenshots of the UI on them.

Generate every required size from a single 1024 by 1024 master file. Never upscale a smaller icon, the App Store will display it pixelated on Retina screens. Export at full quality, the App Store re-compresses but starts from your source. Run the icon past 3 to 5 people who don't know your app and ask what they think it does. If their answers are all over the place, the icon is too abstract.

Screenshots drive 60 to 70% of the install decision according to repeated A/B tests on the App Store. Your first three screenshots are non-negotiable, they appear in search results and they get roughly 7 times more views than the last screenshot in your gallery. Lead with the single benefit a new user cares about most.

In 2026 the patterns that perform best are large headline text in the top third, a single phone mockup centered, and a saturated background color that contrasts with surrounding listings. Headspace, Notion and Duolingo all follow variations of this layout. Six words maximum on the headline. If you cannot say it in six words, the message is not sharp enough yet.

Generate screenshots at the required portrait sizes, 6.7 inch (1320 by 2868) and 6.5 inch (1242 by 2688) for iPhone, plus 12.9 inch (2048 by 2732) for iPad if you support it. Export PNG, never JPEG. Use ScreenMagic to produce a full localized set in minutes rather than spending a day in Figma per language.

Apple autoplays your preview video on the App Store, and the first frame of that video replaces your first screenshot in some search contexts. That makes the poster frame just as important as the video itself. Design the poster to work as a standalone screenshot, with the same headline treatment as your other creatives.

Keep the video under 20 seconds. The first 3 seconds need to communicate the core value, because most users swipe away before second 5. Show the actual product, not stock footage or animated brand logos. Apple rejects videos that don't show real app interactions. Cal AI's preview video shows three calorie scans in 15 seconds, no narration, no music swell, just the product working.

Export at the right specs, 1080 by 1920 portrait for iPhone, H.264 codec, max 500 MB. Add captions for accessibility, since 80% of users browse with sound off by default. If you cannot produce a video that beats your screenshots-only conversion rate, remove it. A weak video hurts more than no video at all.

Apple uses your average rating and review velocity as a ranking signal, especially in the 4.0 to 4.7 range where most apps sit. Going from 4.2 to 4.6 stars can lift category ranking by 5 to 15 positions for competitive keywords. The fastest way to get there is to ask happy users at the right moment, not after a crash or a paywall.

Use SKStoreReviewController on iOS to trigger review prompts after a clear success event, like completing a goal, sharing a result, or finishing onboarding. Apple allows up to 3 prompts per user per year, use them sparingly. Never offer rewards or incentives for reviews, Apple removes apps caught doing this and competitors will report you.

Reply to negative reviews within 48 hours. Public replies don't change the original star rating in real time, but they often prompt users to update their review once their issue is resolved. Track your weekly review volume in App Store Connect, a sudden drop usually means a recent update introduced a bug your prompt logic is now suppressing.

Localized listings convert 25 to 40% better than English-only listings in non-English markets, according to Apple's own data and independent studies from Apptentive. The mistake most teams make is localizing into 15 languages at once with poor translations. Pick 5 markets, go deep, then expand.

Prioritize Japan, Germany, France, South Korea and Brazilian Portuguese for most app categories. These five cover roughly 40% of non-English App Store revenue and they reward localization more than markets like the UK or Australia where English-only listings already work. Use professional marketing translators rather than generic translators, marketing copy needs cultural adaptation, not just word substitution.

Localize four assets, keywords (the 100 character field is per-locale), title and subtitle, screenshots, and the description. Skipping screenshots is the most common mistake, translated metadata with English screenshots leaves most of the conversion lift on the table. ScreenMagic and similar tools can regenerate localized screenshot sets in minutes once your base templates are ready.

Apple's Product Page Optimization (PPO) lets you run native A/B tests on your icon, screenshots and preview video against the live version. Google Play offers Store Listing Experiments with similar functionality. Most apps run zero tests per year, the top 1% run a new test every 30 days. The compounding effect is enormous.

Test one variable at a time. Start with the first screenshot, since it has the largest measurable impact. Run each test for at least 7 days and until Apple shows a 90% confidence level, this usually means at least 1000 page views per variant. If your daily traffic is too low for that, focus on lower-traffic markets first or test bigger creative differences to detect a signal faster.

Document every test in a single spreadsheet, hypothesis, variants, winner, lift, sample size. Over a year you accumulate a private playbook of what works for your audience that no agency can match. Notion, Linear and Cal AI publish about their PPO learnings on their team blogs, study their patterns then test similar hypotheses on your own listing.