Getting Found in the Store, and Installed Once You Are
ASO is two jobs wearing one name: being findable for what people search, and being convincing in the few seconds after they find you. Neither is a one-off task, and neither can save an app people do not want.
Store Search Is Not a Small Version of Google
People arrive with SEO instincts and most of them are wrong here. The mechanics are genuinely different.
You have one page, not a site
There is no content strategy to fall back on, no blog, no internal linking, no long tail of articles pulling in traffic. There is a single listing per store and a small number of fields. Every word has to earn its place, because there is nowhere else to put the rest.
Queries are short and blunt
People type two or three words on a phone keyboard, often a category or a brand. There is very little of the conversational query that web search has trained everyone to write for, and almost nothing that a long explanatory paragraph would match.
Behaviour is a ranking input
How many people install after seeing you, what they rate you, and whether they keep the app all feed back into where you rank. That is the loop nobody can shortcut: a listing that converts badly gets shown less, which is the store agreeing with your users.
Browsing matters as much as searching
Category charts, editorial placement, similar-app recommendations and the apps your users already have all deliver installs that never involved a query. Some of that responds to the same work, and some of it responds to nothing but being good.
The result is a picture
Search results show an icon, a name and, on iOS, the first screenshots or a video. The decision to tap happens in about a second and it is mostly visual. Text optimisation gets you shown; the imagery gets you opened.
Two stores, two rulebooks
Different indexed fields, different character limits, different testing tools and different review cultures. Copying a listing from one to the other is the most common wasted opportunity we see.
Which Fields Actually Affect Ranking
Effort spent on a field the store does not index is effort wasted. The split between the two stores is the single most useful thing to understand.
Apple App Store
- The app name carries the most weight of any field, and it is short
- The subtitle is indexed as well, and it is the one line under your name
- A separate keyword field that users never see, with a tight character budget
- Your developer name and, in some cases, in-app purchase names are also read
- The description is not indexed for search, so it is pure persuasion rather than keyword real estate
- Category matters for browse placement and for what you are compared against
Google Play
- The title is the strongest field here too, and just as tightly limited
- The short description is both indexed and shown, so it does double duty
- The full description is indexed, which is the biggest difference between the stores
- Repetition is read but it is not free, because stuffed copy costs you conversion
- Install and engagement signals feed back visibly into ranking
- Category and tags shape which browse surfaces you appear on
Store mechanics change, sometimes without announcement. We check behaviour against current documentation before rewriting a listing rather than working from folklore.
Finding the Words Without a Web Keyword Tool
Web keyword volumes are the wrong data. Someone searching for advice reads an article; someone searching in a store wants an app, and they say it differently.
Start with the job, not the product
List what people would say they are trying to do, in their words rather than your feature names. Support tickets, sales calls and the phrases in your own reviews are a better source than any brainstorm.
Read the store autocomplete
Type the first few letters of each candidate into both stores and record what is suggested. Those suggestions are drawn from real searching, and they are free, current and specific to the store you are optimising.
Study who already ranks
For each term, look at what comes back. If the top results are enormous apps with years of installs, that term is a long-term ambition. If the results are weak or barely relevant, that is a gap worth taking.
Use paid search as a demand signal
Store advertising tools expose how popular a term is. Even a small test campaign tells you which words people actually use and which convert, which is far better evidence than an estimate.
Be realistic about difficulty
Ranking is relative to your install volume and your conversion. A new app should aim at specific phrases it can genuinely win, gather signal there, and move up to broader terms later rather than fighting for the category word on day one.
Do not repeat yourself
On iOS a word only needs to appear once across the indexed fields, so repeating it wastes a scarce character budget. Combinations form across fields, which is why the keyword set should be planned as one thing rather than three.
The Most Valuable Text You Will Ever Write
A handful of characters, shown everywhere, indexed heavily, and awkward to change later.
Brand plus what it is
A bare brand name tells a stranger nothing, and a listing stuffed with descriptors looks like spam. The pattern that works is your name, a separator, then the clearest short statement of what the app does, in the words people search for.
Write for the truncation
Names are cut off in search results, on the home screen and in charts. Check how yours appears at the shortest width before falling in love with it, because the first few characters are what most people will ever read.
The subtitle is not a slogan
The line under the name is indexed on iOS and read by humans on both stores. Spending it on something abstract about empowerment wastes the second-most valuable field on the page. Say what the app does for the person reading.
Changing the name has a cost
It resets recognition, breaks word of mouth and confuses existing users who search for the old name on a new phone. Worth doing when the current name is genuinely holding you back; not worth doing as an experiment.
The First Two Screenshots Are the Whole Fight
Search results on iOS show the first few screenshots in place, and most people never open the full product page at all. Everything after the second image is for the minority who scrolled.
- One idea per screenshot, with a short caption above the device frame
- Text large enough to read at thumbnail size, not at full size on a monitor
- The strongest benefit first, the proof second, the detail afterwards
- Real screens from the real app, because inventing features gets you rejected
- Regenerated whenever the interface changes, so the listing is never a museum
- Captions translated for every locale you publish in, not just the app itself
The preview video
Worth making when the value is hard to show in a still — a game, a drawing tool, anything with motion. Each store has its own rules on length and on what may appear in the frame, and a video that opens with a logo animation wastes the seconds that count. If the app is a form and a list, a good screenshot set beats a mediocre video.
The icon
Judged at a size smaller than a fingernail, next to eleven competitors. One clear shape, strong contrast, no words, no screenshot of the interface. Test it as a thumbnail on a busy home screen rather than on a white background, and remember that changing it resets the recognition your existing users have built.
How to Ask Without Becoming the App People Complain About
Rating is both a ranking input and the first thing a stranger looks at. It is also the easiest thing to damage by asking badly.
Ask after a win
Prompt when something has just gone right: an order delivered, a workout finished, a document saved. Never after an error, a failed payment, a long spinner, and never on first launch before anyone has any opinion to give.
Use the native prompt
Both platforms provide an in-app rating control that does not throw the user out of your app, and both limit how often it can appear. Those limits are a budget: spend a prompt on the wrong moment and you do not get it back.
Never buy or gate it
Paid reviews, incentives conditional on a good rating, and funnels that route unhappy users to a private form while sending happy ones to the store all breach store policy. The risk runs to the account, not just the listing.
Reply to the bad ones
Developer replies are public and they are read by the next person deciding whether to install. A calm, specific reply that names the fix does more for conversion than the original complaint did damage.
Reviews are a defect queue
The same complaint from six people is a bug report with a sample size. We feed review themes into the backlog, and note the release that fixed them, which also gives you something true to reply with.
Recent weighs heavier
A rating that reflects the current version matters more than a historic average, which is why a fix plus a considered ask can move a score that looked permanent. Both stores treat rating history differently, so the tactic differs by platform.
From Seen, to Opened, to Installed
Ranking is only half of it. The other half is the share of people who install once they have seen you, and that number is usually easier to move.
Impressions
You appeared in a search result, a chart or a recommendation. Driven by keywords, category and the store deciding you are worth showing.
Product page views
Someone tapped through. Almost entirely the icon, the name and the first screenshots doing their job in about one second.
Installs
The page convinced them. Screenshots, rating, size, and whether the description answers the one question they arrived with.
First open and retention
Where ASO hands over to the product. Installs that never open, or open once, quietly undo the ranking the listing earned.
Both stores report these stages in their own console. Watching the step-to-step rate tells you which asset to work on next far better than watching installs alone — more on measurement in analytics and monetisation.
Experiments, and Then Other Languages
Two ways to grow a listing once the basics are right: prove which version converts better, and then do it again in another language.
Listing experiments on Play
Real store traffic split between variants of your icon, screenshots, video or text, with the result reported as a difference in install rate. The closest thing to proper evidence available in this discipline, and the reason we usually start testing on Android.
Product page testing on iOS
Apple provides its own way to trial alternative icons, screenshots and videos against the default page, plus custom product pages that can be linked from a specific campaign so an advert leads to matching imagery.
One change at a time
Change the icon and the first two screenshots together and you will never know which one moved the number. Where traffic is too thin for a clean test we run changes sequentially and accept a rougher answer rather than pretending a tiny sample proved something.
Localise the listing first
Translating the store listing is cheap compared with translating the app, and in some markets it lifts conversion on its own. It also opens keyword sets you cannot reach in English.
But do not lie by omission
A translated listing leading to an English-only app produces installs, one-star reviews and refund requests. Either localise the app too, or say plainly in the listing which languages it supports.
Keywords are not translations
The literal translation of your best English term is often not what people in that market type. Each locale needs its own research pass, and screenshot captions need redoing, not just the text fields.
ASO Cannot Rescue an App People Do Not Want
This is the part most agencies leave out of the pitch, and it is the most useful thing on this page.
When we look at an app with weak numbers, we look at retention and reviews before we touch a single field. Quite often the listing is fine and the first two minutes of the app are not, and we would rather tell you that than sell you a keyword report. That fix usually starts with design or a rebuild.
What Sits Either Side of the Listing
Launch & Submission
Getting the listing published in the first place, on both stores.
Learn moreAnalytics & Monetisation
Retention and revenue, which is what decides whether more installs are worth having.
Learn moreSupport Plans
Where ongoing listing work, review responses and release notes actually live.
Learn moreASO Questions
Is ASO a one-off job or something ongoing?
Ongoing, because everything it depends on moves. Competitors rewrite their listings, the words people search for shift with the season and with whatever is in the news, your own ratings drift up or down, and the stores keep changing what a search result looks like. A first pass sets up the name, the keyword set, the screenshots and the icon, and that pass does most of the heavy lifting. After that the useful rhythm is a monthly look at what changed and a deliberate test of one element at a time, tied to your release cycle rather than done at random.
How soon does ASO work show up in the numbers?
Conversion changes show up fast and ranking changes do not. Replace your first two screenshots and you can often see the effect on install rate within days, because you are changing what happens after someone already found you. Keyword and metadata changes need the store to re-index, then need enough traffic to judge, so a few weeks is a fair expectation before anyone should draw a conclusion. The mistake we see most often is changing five things at once on both stores, which guarantees you learn nothing from any of them.
Do keywords belong in the app description?
It depends on the store, and the two work differently enough that the same text should not be pasted into both. Apple takes keywords in a separate field that users never see and does not index the description, so your description is pure persuasion. Google reads the title, the short description and the full description, so the words people actually search for need to appear in that copy, written as sentences a human would want to read. Keyword stuffing hurts on both: it reads as spam to a person deciding whether to install, and that decision is itself a ranking input.
Should the app ask users for ratings?
Yes, but rarely and at the right moment. Use the rating control the platform provides rather than sending people out to the store, ask after something has gone well such as a completed order or a finished session, and never after an error, a payment failure or a long wait. Both platforms limit how often the native prompt can appear, so spending one on a user who is mid-task wastes it. What nobody should do is condition a reward on a good review or route unhappy users to a private form while sending happy ones to the store, which breaks both store policies.
Can we A/B test the store listing?
On Google Play you can run listing experiments that split real store traffic between variants of your icon, screenshots, video or text and report which converted better. Apple provides its own testing for product page assets, along with alternative product pages that can be linked from different campaigns, so a listing arriving from one advert can show different screenshots to the default one. Both need enough traffic to reach a conclusion, which is the real constraint on a small app. Where traffic is thin we test sequentially and accept a slower, rougher answer rather than pretending a tiny sample is significant.
Our downloads are low. Will ASO fix that?
It will fix a findability problem and a listing problem. It will not fix a product problem. If people search, find you, install and then leave within a day, better keywords simply deliver more people to the same disappointment, and the ratings that follow will pull your ranking down rather than up. So we look at retention and reviews before touching metadata. Sometimes the honest answer is that the listing is fine and the first-run experience is not, and that is a build problem, not a marketing one.
Have Your Listing Looked At Properly
Send us your store links and we will come back with what we would change first, in order, and what we would not bother with — whether or not you hire us for it.