Get the newsletter
Google · Software · Filed Jul 3, 2025 · USPTO published Nov 19, 2026 · This week's release

Google Files Patent for Spotting Trending Apps Before They Go Mainstream

Google wants its app store's Trending list to tell you something you don't already know. Its patent application describes a statistical test that separates real surges from random blips, plus a model that guesses which apps will take off.

A smartphone displaying a ranked list of top trending apps in an app store. Drawing from patent filing US 2026/0368778 A1.
A smartphone displaying a ranked list of top trending apps in an app store.All 5 drawings ↓
Publication number
US 2026/0368778 A1(US20260368778A1)
Applicant
Google LLC
Filed · Published
Jul 3, 2025 · Nov 19, 2026
Inventors
Zilu Zhou, Dinglong Li, Bruce Ban Herm Chia
Status when we published
Examiner raised objections; the company can respond
Document
20 claims
US classification
705/14.52
Examiner
VANDERHORST, MARIA VICTORIA (Art Unit 3621)
Parent application
National Stage Entry of PCT/US2024/011290 (filed Jan 11, 2024)(PCTUS2024011290)

What Google's trending-app patent actually does

You open the app store and tap Trending, and the list is full of apps you've already heard of, because by the time an app makes the list it's usually a household name. That's not much of a tip-off.

Google's filing describes a fix. The store tracks how much people use each app, and how many people use it, day after day. It then checks whether the latest numbers are higher than earlier ones by an amount too big to be luck. If the answer is yes, the app gets labeled trending and shown that way in the store.

There's a second layer too. A prediction model tries to guess which apps are about to qualify, so the store can feature them early, before most people have noticed.

From the filing · CLAIM 13
determining, by a computing system and for each time period from a plurality of time periods, a trending metric for an application available through an application marketplace …

Translation: The system calculates a popularity score for an app across multiple different time frames.

How Google tells a real surge from random noise

The core of the claims is a repeatable test for "is this app really on the rise?" The system measures a trending metric for an app across a series of time periods. It then compares the last period to the first one and asks whether the increase clears a significance level (a statistical bar that rules out random fluctuation). If it does, the app is classified as trending and displayed that way in the marketplace.

The description adds detail on how this could work:

  • Two metrics: an engagement metric (time spent using the app) and an active user metric (how many people who have it installed actually used their device in that period).
  • Two time scales: a long-term check against the initial period at a level of 0.05, and a short-term check against a middle period at 0.1. The long-term bar is stricter because there is more data behind it.
  • Assumed statistics: engagement is modeled with a Gamma distribution and user counts with a Poisson distribution (two standard ways of describing how such numbers normally scatter).

The example windows are 7, 14 or 21 days, and the claims name 21 periods of one day each, either fixed or sliding forward day by day. Data may be smoothed to reduce seasonal swings.

A machine learning model can also predict, before the final period begins, whether an app will be classified as trending. If so, the store can show it as trending right away. The model is trained on its own prediction accuracy.

Why a spot on Google's Trending list matters

If you've ever scrolled a Trending list and found nothing new, this is aimed at you. Showing apps just before they break out could make the store a better place to discover things, and it gives smaller developers a real shot at visibility. The filing itself says a Trending slot has a considerable effect on an app's adoption, which is why it stresses reliability.

For Google, the practical upside is a Trending section people actually trust and use. The examiner's first review rejected the claims as filed, which is routine at this stage and usually followed by amendments.

This is the 1100th Google filing in our Google coverage since May, adding to work like blurry face sharpening and fall detection without face ID.

Source: US 2026/0368778 A1 at the USPTO (PDF). Full patent text and figures are in the official publication.

Get our take in your Top Stories

Liked this breakdown? Add Patentlyze as a preferred source on Google, and our plain-English take shows up more often in your Top Stories the next time Google patent news breaks.

There are more where this came from

We read every patent application Big Tech publishes and send you the ones worth knowing. Plain English, free, every week.

The drawings · 5 sheets

5 drawing sheets from US 2026/0368778 A1 · click any drawing to enlarge

Phones sending app activity across a network to servers, drawing sheet 1 from US 2026/0368778 A1
Sheet 1 Phones sending app activity across a network to servers
Server components that analyze app data using machine learning, drawing sheet 2 from US 2026/0368778 A1
Sheet 2 Server components that analyze app data using machine learning
A phone showing a ranked list of trending apps, drawing sheet 3 from US 2026/0368778 A1
Sheet 3 A phone showing a ranked list of trending apps
Tracking sudden metric spikes to classify an app as trending, drawing sheet 4 from US 2026/0368778 A1
Sheet 4 Tracking sudden metric spikes to classify an app as trending
Using AI predictions to spotlight future hit apps in stores, drawing sheet 5 from US 2026/0368778 A1
Sheet 5 Using AI predictions to spotlight future hit apps in stores
Patent filing page

Reader comments

No comments yet

Start the discussion

Real name or a handle, either is fine. Comments are read by a person before they appear, so allow a little time. Keep it about the filing.