ASO: ranking on the App Store and Google Play (find 10 differences in the algorithms)

I’ll say right away that no one will give you an exact answer to most of the questions discussed below (if your friend doesn’t work at Apple / Google and did not intend to violate the NDA), and the current current version of the answer can only be obtained by tests, preferably on a lot of traffic. And it’s never known exactly how long the relevance of the current correct version of these answers will last.

Therefore, the conclusions made below do not pretend to be true in the last resort, but are the product of analytics and numerous experiments on real traffic, plus the advice and opinions of members of the #asostack community .

In both cases, ranking (not to be confused with indexing ) for a search query is affected enough many factors (and many were brokenthe face of virtual dishes in disputes about their relative weight, so I will only describe the conclusions, the current relevance of which is not in doubt among most experts). So, the ranking is strongly influenced by:

1. weighing result for matching keywords,
2. total number of application downloads
3. the number and intensity of downloads from this particular search query,
4. Quality indicators, the list of which is different for Google Play (GP) and the App Store (upstore).

Today we will talk about the first point, slightly touching on the fourth.



Insight 1 : in both sides, the ranking (position in the search results) is strongly influenced by the weighted weight of keywords included in the request (on Google Play, the weight of the keyword in Short Description is greater than its weight in Description, and less than its weight in AppName, the position starting from the first the symbol inside each of these blocks affects the weight during the final weighing; in the App Store, the weight in the title is more than the weight in keywords and approximately equal to the weight in the sub-title - according to the current observations, although the results of the relative weight tests of the sub-title periodically diverge, due to the novelty of this block still not gain sufficient statistics; position inside the keywords in the App Store not yet taken into account)

Insight 2: in the App Store, unlike Google Play, it makes no sense to repeat keywords in the name and sub-title / keywords (when ranking the apple platform in case of duplicating keywords only takes into account the object with the maximum weight, the rest are ignored. Google acts completely differently, but do not forget about the probability of “cross the line", because no one has canceled the spam recognition algorithms. At the moment, according to our observations, the optimal key frequency for all indexed blocks of GPs is 4-5)

Insight 3: Both indicators relate to the Quality Indicators the number of application launches (frequency of use), the number of reviews, the average application rating score (in iOs 11 it will be much easier to manage this factor, some kind of cheating is directly compared to the GP), the number of application deletions (only GP ), the time since the last application update (only in the app), the number of external links and mentions / reviews on the web (only GP), the frequency of keys in reviews (only GP) and SBR (The total score of the developer is definitely in the GP, I can’t say about the App Store and b I'm glad if someone shares their research on this). About the last factor (RRF on Google Play) I will write a separate article when I find this time.

Insight 4: in the App Store, for “X Y” to appear in the search results for a query, it is necessary that “X” and “Y” are present in any of the blocks (different ones) involved in indexing (AppName / Subtitle / Keywords / Developer) in the main locale, either that “X Y” is present unchanged in (any) of the additional locales, or that “X” and “Y” are present in the names of the additional locales. In fact, it should have been written “necessary and sufficient,” only the position in the issue can be very different, and it depends on the ranking results. The weight of the keys in the primary and secondary locales does not change.



Bonus is not about ranking: The App Store understands plural numbers (and sometimes cases / tenses) in the English language, in all other languages ​​the same keyword in different numbers = different keywords (in iOs 11 so far everything has remained the same, despite the assurances of Apple from the screenshot above )

I would be glad if in the comments you share your observations on these and related issues.

All unlimited organics;)