Understanding the Search Algorithm of Apple and Google

Understanding the search algorithm of Apple and Google is a million-dollar topic. This topic puts the ASO field at fixed odds against Apple and Google trying to get to know, tricks, and tips to better understand the mechanisms behind rankings on store. Three things should consider when considering how the algorithms work for ranking on app store and whether your app is ranked for the search term (and if so, where)

Read More: Leverage Keywords To Increase Visibility Through Search

Is Your App Eligible to Rank for a Specific Keyword?

Usually, this needs to specify the keyword in metadata of your app.

How Specific is the Added Keyword to Your App?

It is a secret sauce to how relevant your keywords are and the placement of keywords in your app metadata.

What is the Ranking Strength of Your App for Keywords?

App strengths are factors outside of the app’s metadata, such as keyword retention and install conversion rate.

What Does it Mean When an App is indexing?

Show in search results app stores index apps on the basis of keywords include in metadata e.g, app title, short description, subtitle, short description, keyword field, etc. this is the primary main factor that impacts search algorithms. Like search engine optimization, Google Play Store and Apple App Store consider keywords added in metadata to match and try search queries of users. That’s why it is significant to make sure you have added relevant keywords in metadata according to the behavior of users.

It is also possible to rank for a keyword even though you have not included it in your app’s metadata. The algorithm of Google employs machine learning and added embedding neural network models like sip grams. To Identify relevant relationships between keywords that go beyond the metadata of the app e,g., similar context, synonyms, misspellings, etc. the search engine guesses the intention behind a search by algorithm matches similar words. Your app can be ranked for that search term if a keyword is algorithmically associated with another keyword found in the metadata of the app. 

Apple’s algorithm is much simpler than Google’s but also allows apps to rank for many keywords that do not need to be mentioned in the metadata of app. As well as choosing the right keywords, you should also be aware of some specific rules regarding indexing in the app stores. We will now list some important observations regarding indexing by search algorithms. 

Apple’s Shortcomings in Ranking Plural or Compound Keywords

Apple does not always rank apps for the compound and plural versions of singular keywords. While algorithm of Apple is improving at matching plurals with singulars, this process is still a long way from being fully baked in the App Store, especially in respect of non-English searches. To decide whether to include both compound, plural, or compound keywords in metadata of app, it is significant to test, rather than imagine. First, you must do the analysis of both singular and plural variations of the root word e.g., “ox”, ” oxen”, as well as compound words e.g., “audiobooks,” “audio books”), using any ASO tool.

if one type of variation “compound words” closely matches the rank of the root word i.e., the inconsistency is less than 15%, then it’s highly likely that Apple will match these variations as well and that you will rank for both types of keywords by just specifying one. You can see this being the case for “notebook” and “note book” (compound) and “book” and “books” (plural). The reason these variants never quite match is that there are likely to be other matching rules at play. 

On the other hand, we see a big difference between the singular and plural forms of “mouse” and “mice.” It is therefore highly unlikely that Apple is matching the plural to the singular for the keywords “tooth” and “teeth”; to rank well for both keywords, you should add both variants to your metadata of app. 

An exact keyword match has more value and weight than the matching rules Apple offers. If you have a primary search term and both variants bring in high volume (i.e., “book” = 5.2 and “books” = 6.2), we suggest that you should add both the plural and singular forms in your metadata for desired results. As time goes by you can conduct tests of metadata iterations by adding and removing singular and plural keyword variants.

How to Optimize Keywords in the App Store?

The first week of the app after launching is not only a time to check its ability and credibility to fly or flop, it is the time that Apple boosts your app visibility artificially. So, any marketing effort or plan during this time will be more beneficial. For Apple, understanding how good a new app is, or how relevant the keywords in its metadata are, is a guessing game. The algorithm has the power to gauge metrics like conversion rates and retention from different search terms, in seven days, Apple will artificially give a high place to your app in the search results for keywords added to its metadata.

In order to take full edge of this keyword raise, you might want to target relevant but high-volume keywords in your metadata to get more downloads. You could ensure that your app covers the most relevant keywords so that your performance history for these keywords starts high. This will give your app a better chance of maintaining its higher rank for these keywords beyond the week-long keyword boost. 

Keyword Ranking Strength for Apple and Google

Now, imagine if the app stores immediately started with a clean slate and lost all their historical data. What would be the order of results for a search query of, say, the keyword “video maker”? The first step Apple and Google take to decide on search queries is to recognize which apps/games are allowed to show for the keyword on the basis of metadata all apps have added (indexation). This only concerns whether the keywords “video” and “maker” can be found in the metadata of an app.

The next step will be to decide how related your app is to a particular keyword. To unzip this second step, the algorithms initially look at the positioning of a specific keyword in the metadata of app to determine relevance. As a rule of thumb, the algorithms consider the following:

• How visible the metadata is, with the title weighing more heavily than the iOS keyword field, for example. 
• How many characters are allowed in a metadata element? 
• How many times do the keywords appear in your metadata (Google only)? 

On Google Play, keywords that are mentioned more often and earlier in the long description are 

more relevant, and thus get a better ranking than those found in the later describing lines. Let’s say, If you mention the word “restaurant” five times in the first few sentences of your long description, the Google Play algorithm will give your app a better relevance weight for “restaurant” than if you mention the keyword the same number of times, but spread out over the entire long description.

As far as the relevance of your metadata for a keyword such as“video maker”—and This is especially true for Google Play metadata and App Store titles/subtitles.—the algorithm also allocates more relevance weight if you have an exact match in your metadata.

Another Possible Ranking Factor

Some years ago, it was considered that short description have a greater impact on rankings than long description; however, evolving Google Play algorithms have led ASO experts to believe that the latter holds more weight when keywords are repeated multiple times.

While we have no evidence (yet) that reviews are indexed in the App Store, there have been cases where “keyword stuffed” black hat reviews have appeared in the App Store and have been preceded by an increase in keyword rankings.

The idea that Google might be able to accept text on the assets attached to store listings and then index apps for this text has been proven false in the past. Nevertheless, this concept remains one of the hypotheses of how screenshots can be used to rank keywords.

A test was run to decide if the file name of screenshot could also influence keyword rankings. However, changing the file name of screenshot from “IMG0001.png” to “new-keywords-to-rank-for.png” was not noticed by Google and the test failed to cause an increasing in keyword rank. 

Conclusion

Earlier, we discussed a theoretical situation in which the app stores would start out with a clean slate and no history of how they ranked apps. There is no historical data to increase search results. The app stores would have to rank apps based only on their metadata. For both Apple and Google, however, giving a good user experience is a first priority. Therefore, the app stores use historical performance data to help organize searching in order to give users high-quality search results.