Review miningThe complete guide to optimising your app title
What is in this resource:
How to build your ASO strategy from end to end
Learn to pick target keywords, write metadata, follow rank and base every decision on data.
From data to action
Simplify the keyword data and stay in line with App Store rules.
Tools that lift rank
Carry the keywords drawn from competitor reviews into your metadata and take your app out of the cusp band into the top 10.
The three indexed fields
The indexable space Apple gives you is limited, and half of it usually goes to waste. The causes are dully mechanical - which is exactly why all of it is recoverable.
The fields that make an app findable in App Store search are limited and they do not weigh the same. Any edit made without knowing what counts where is a shot in the dark.
The app name is the strongest. Visible to the user, indexed, and the place Apple gives most weight. The character limit is tight and it carries brand value, which makes it the most expensive field too: changing it often confuses users and disturbs your non-search traffic.
The subtitle is second. Visible, indexed, and far easier to change than the name. It is the field most often used to carry cusp-band phrases to the first page. It has to stay readable - a subtitle that turns into a word list gets indexed and loses the user.
The keyword field is third. Invisible, 100 characters, its contents open only to the developer. The words in it combine freely with one another to form phrases - the field's most important and least known property.
There are unindexed fields too, and knowing them saves time. The long description is not indexed for App Store search; stuffing keywords there tires the reader and gains nothing. Text in screenshots is not indexed either. Those matter for conversion, not for being found.
The five shapes of waste
The reasons half the keyword field goes to waste are specific and countable. Here are all five, because each one recovers a different amount.
- Repeating a word from the app nameWriting a word that already appears in your app name into the keyword field again gains nothing; Apple has already indexed it. This is the most common waste and usually the largest: with a long app name you may have spent a real share of the field on repeats.
- A space after a commaWords have to be separated by commas, but no space is required. Every space is one character doing nothing. Across a twenty-word field that is nineteen characters - room for another phrase.
- Both singular and pluralApple does not keep the singular and plural of the same stem apart. Writing both spends the longer one's characters. The same holds for close inflections.
- Writing a phrase wholeThe words in the keyword field combine freely. Writing "learn" and "english" as separate words recovers the space between them and makes other combinations possible as well.
- Empty country fieldsEach of the 175 countries and regions Apple is open in has its own keyword field. Leaving one empty spends all 100 characters there. This is the largest waste and the most invisible.
Planning the three fields together
Filled separately, the three fields repeat one another. Planned together, the same total characters produce far more phrases.
Planning starts with the app name, because it is the tightest and most expensive field. Which words are in it? Those count as indexed and should not be repeated in the other two.
Then the subtitle. It should be a sentence, not a word list - the user reads it. But watch which words the sentence contains: it is possible to build a readable sentence that carries the phrases you are targeting and are not in your name. Words in the subtitle also count as indexed.
The keyword field is filled last, with only the words that appear in neither of the first two. Follow that order and the whole field is spent on new words.
One warning: this planning has to be repeated per country. App names can be written differently in some storefronts; if the name differs, which words are indexed differs, and the keyword field has to change with it.
- App name first: the strongest field, and the most expensive.
- Then the subtitle: a readable sentence that carries your target phrases.
- Keyword field last: only words absent from the first two.
- Repeat the plan per country; if the name differs the field must too.
What a metadata audit counts
Finding the waste needs no guessing; counting is enough. That is exactly what an audit does, and every number it produces can be verified.
The first count is repeats. The words in the app name and subtitle are extracted, compared with the words in the keyword field, and the intersection found. Each shared word is waste equal to its own length.
The second is spaces. The spaces after commas in the keyword field are counted. That number is directly recoverable characters.
The third is stem repeats. Different forms of the same stem are found and everything but the shortest is flagged. This count depends on the language: with rich suffixes it finds more waste than it does in English metadata.
The fourth is character usage. How much of the field is filled and how much of the limit is used. Not filling the field is a waste too; every character left empty is an opportunity not taken.
The fifth is country coverage. How many countries have the field filled and how many are empty. This can be the part of the audit that finds the most characters.
The audit's output is not an estimate, it is a list: which word repeats where, how many characters come back. That list can be produced by hand as well; it is only repeating it across 175 storefronts that is impractical.
Six mistakes made again and again
The same mistakes recur while metadata is being edited. Here are six; every one of them happens to everybody at least once.
One: stuffing keywords into the description. It is not indexed for App Store search. Putting words there gains nothing in search and loses the reader.
Two: writing a rival's brand into the keyword field. Apple does not accept it and it leads to rejection at review. Even if accepted it brings irrelevant traffic and poor ratings.
Three: targeting phrases describing something the app does not do. You can rank, but the arriving user does not find what they came for. Low-rated reviews affect your strength across every phrase; you win one and lose all of them.
Four: changing every field in the same release. When the rank rises you will not know which one carried, and so you will not know what to repeat next time.
Five: copying one country's metadata into another. The same language being searched differently in different countries is the rule, not the exception. Copying spends one of the two.
Six: not saving the current state before a change. When a move drops your rank you need the old metadata to go back, and remembering is not enough.
| Mistake | What it costs |
|---|---|
| Keywords in the description | No gain in search; readability lost. |
| A rival's brand name | Review rejection, or irrelevant traffic. |
| Targeting work you do not do | Low ratings; strength lost across every phrase. |
| Every field in one release | Which field carried cannot be measured. |
| Copying between countries | One of the two is wasted. |
| Not saving the old state | Undoing becomes impossible. |
Measuring the effect of a change
The metadata is edited and the release has shipped. One question remains: did it work? That question can only be answered with a disciplined measurement.
First, a starting point. You need to have recorded the ranks of every phrase on your tracking list before the change. If you did not, this release's effect cannot be measured; record it for the next one.
Then waiting. Apple's index does not update immediately; a metadata change usually takes a few days to reach search results. A second change inside that window mixes two moves together.
Then the control group. Watch how the phrases you did not touch moved. If they all fell two places, your phrase falling two places is not the move failing, it is the market.
Then the comparison. Subtract the control group's movement and write what is left down to the move. If you won, apply the same kind of move to the next phrase on the list. If you lost, undo it and check a week later whether the old rank returns.
The real output of that loop is not the ranks won, it is the record accumulating. Three months in you have real knowledge of which field works for your app - and that knowledge is in no general guide.
Release rhythm and metadata
A metadata change needs a release, and release rhythm directly sets the pace of ASO. Not planning around it wastes most of the work.
A team shipping weekly can change one field a week and take one measurement a week. That is the ideal rhythm and it is not realistic for most teams.
A team shipping monthly takes one measurement a month. Choosing the move well matters far more then: a month is too long to spend on a wrong move. Go to the phrase at the top of the band, through the field with the highest odds.
A team shipping quarterly should think about ASO separately from releases. At that rhythm metadata moves happen four times a year; the real gain will come from the product side and from gathering ratings.
There is also the possibility of changing without a release: some metadata fields can be updated without shipping a new build. Which fields those are has to be verified against Apple's own documentation, because the scope changes from time to time. If a field can be updated without a release, the ASO loop can run independently of release rhythm - and that is a large advantage.
- Weekly releases: one field a week, one measurement a week. Ideal.
- Monthly: move selection is critical; top of the band, most likely field.
- Quarterly: the gain will come from the product; metadata is secondary.
- If any field updates without a release, build the loop on that one.
Frequently asked
- Should I separate keywords with commas or spaces?
- Commas, with no space. A space after a comma spends one character per comma and gains nothing.
- Should I repeat the words from my app name in the keyword field?
- No. Apple has already indexed them; repeating only spends characters. This is the item that recovers the most in most audits.
- Is there any point putting keywords in the long description?
- It is not indexed for App Store search, so there is no gain for being found. Write the description for the user; the three indexed fields do the search work.
- How often should I change my metadata?
- It depends on release rhythm, but one rule is fixed: one field at a time, and at least a week between changes. Changing more often makes measurement impossible.
- Do I really need to fill the keyword field for every country?
- For every country you receive downloads from or intend to, yes. An empty field spends all 100 characters there - and that waste has no visible sign at all.
Explore our resources
The turning point in ASO: a guide to raising visibility and staying in the race
Ready to rethink ASO from the start? Our new guide explains step by step the road from rank to download. We show how to find the keywords in the cusp band, which phrases to draw out of competitor reviews and how to build the 100-character field. Download the ASO guide now and pass the app ahead of you in search!
See resource
Metadata optimisation - Part II
When an app falls outside the top 10 on the App Store, it loses most of the visibility that comes from search. The real gain is hidden in the 11-30 cusp band: the words only one step from the first page. This guide explains step by step how to find that band and carry it up.
What is inside?
- A clear frame for how the cusp band is measured.
- The method for drawing keywords out of the user reviews of competitor apps.
- An analysis of the mistakes most often made when building the 100-character keyword field.
See resource
A guide to review mining
User reviews are the App Store's least used keyword source; where the tools everyone looks at give the same list, the reviews of competitor apps show which words users describe that app with. A large part of those phrases appears in no keyword tool. So how does one draw sound keywords out of reviews? How are the resulting phrases prioritised as a whole, and in which country does which word have a real counterpart?
See resource