ASOpatents.com compiles a list of patents that are likely used to shape the algorithms of the Apple App Store, Google Play Store, and other major platforms. While it's uncertain whether these patents are actually implemented in the algorithms, the site provides insights into potential clues about search results, recommended apps, and other data points.

US11853306B2: What Apple Actually Got for Personalized App Store Recommendations

Patent numberUS11853306B2
Official titleTechniques for personalizing app store recommendations
AssigneeApple Inc.
InventorsJayasimha R. Katukuri, Peter Leong, Chandrasekar Venkataraman, Rabi S. Chakraborty, Hardik Vala
Priority dateJune 3, 2018, via provisional application 62/679,949
FiledSeptember 28, 2018
GrantedDecember 26, 2023
StatusActive. Anticipated expiry February 25, 2040
FamilyUS20190370345A1 (published application), US20240086412A1 (continuation, pending), related to US11422797, Chinese application 201910447549.8
Sourcepatents.google.com/patent/US11853306B2

What This Patent Covers and Why It Matters for ASO

Feature it describes: personalized app recommendations, built by matching a user profile against application profiles.

This site already covers US20240086412A1, which is the pending continuation from this family and which carries the full disclosure: user profiles, cluster labels, the installation retention factor, the trending and stability factors, accolades, and the query modifier that rewrites a user’s search before ranking.

This article exists because that continuation is an application, not a patent. It confers nothing. US11853306B2 is the granted patent from the same family, it survived examination, it is in force until 2040, and it is what Apple can actually enforce.

The instructive part is the gap between the two documents.

The specification describes an elaborate apparatus: semantic information extracted from app text, engagement metrics, derived quality scores, machine learned user clustering with readable labels such as “sparse gamer” and “reader”, demographic attributes, and a query modifier that adds, changes or removes words from a search query based on user history.

Almost none of that is in the granted claim. Claim 1 covers the skeleton: a server receives a recommendation request, identifies the user profile, accesses application profiles, analyzes the user profile against a subset of those application profiles to pick apps to recommend, associates the recommendations, and causes them to be displayed.

For anyone reading patents as evidence, that gap is the lesson, and it applies to every article on this site:

  • What is claimed is what was novel enough to defend. The narrow claim here suggests the examiner found the specific profiling machinery either obvious or already disclosed elsewhere.
  • What is described is what the team was building. The installation retention factor and the query modifier are real design decisions even though they are unprotected, and they are the parts with the most ASO consequence.
  • Citing “Apple’s patent” for the query modifier is imprecise. The honest citation is the specification, disclosed in this family’s published applications, not the granted claim.

One more signal worth recording. The family has a Chinese counterpart and a live continuation filed in late 2023, five years after the priority date. Apple is still spending money on this line, which is a stronger indication of ongoing relevance than any single document’s status.

Patent Summary

A server computing device receives a request for at least one software application recommendation from a user’s device. It identifies, among a plurality of user profiles, the profile associated with that user. It accesses a plurality of software application profiles, each associated with an application the server manages.

It then analyzes the user profile against a subset of those application profiles to identify at least one application to recommend, associates the recommendation with that application, and causes the user’s device to display it.

Note the word subset. Candidate selection happens before matching, and neither the granted claim nor the available record specifies how that subset is chosen. In systems of this kind, candidate selection frequently determines the outcome more than the ranking does, so this is a substantive gap rather than a formality.

A detailed breakdown of the claim set, the prosecution history and the differences from the continuation will follow in a later update to this article.

Related reading: US20240086412A1 carries the full disclosure, including the app profile signals and the query modifier.

One response to “US11853306B2: What Apple Actually Got for Personalized App Store Recommendations”

  1. App Store Ranking Factors: What Apple’s Patents Actually Say – ASO Patents

    […] US9405832B2, US9280595B2, US20120191694A1, US10394838B2, US8364694B2, US10678832B2, US9104670B2, US11853306B2, plus US11360958B2 and […]

Bir yanıt yazın

E-posta adresiniz yayınlanmayacak. Gerekli alanlar * ile işaretlenmişlerdir