A solo Android developer published thirteen years of Google Play Console exports — monthly installs, uninstalls and active devices for three apps — and laid them against one question. In October 2014, one of his games did 2,591,467 installs in a single month and peaked at 886,208 active devices. In June 2026 the same game had 499. The essay is about what changed in between, and the answer is not that the game got worse.

The 2014 numbers, in his own exports:

  • Shooting Expert 2: released 25 Aug 2014 with 22 installs. A million users four weeks later. 13,559,906 lifetime installs.
  • The distribution was not marketing. The Top New Free chart sat on the front page of the store, and once a game landed on it the store did the work. His job was “to have something people kept.”
  • He learned what an ad banner is worth on an earlier app — a phone remote for a TV whose own remote was exposed as a web page. 774,898 installs, and 105 devices still had it in June 2026.

How it ended — nowhere near as fast as it started:

  • Google moved ads to the top of Play search and redesigned the store front. The chart engine kept running, just underneath the paid slots.
  • He tried paid acquisition once, with his own money. An ads-only game means buying traffic in order to show ads to it, and without charts carrying the rest of the distance it lost more than it returned.
  • Serbia could not open a Google merchant account until October 2021, so for the entire run those 886,208 peak devices could only ever produce ad revenue. No paid apps, no IAP, no subscriptions.
  • He left for web development in 2018, when ad revenue had come down to what he calls “intern money.”

The 2026 side is the same measurement problem, reversed. New game, press kit, a launch post written for the target subreddit: 493 views in eighteen hours — the fourth most-viewed post in that community that day — single-digit store visits, and zero installs, confirmed by the ad SDK logging no ad requests at all. Both old games have since been rebuilt in Godot and moved to the new account, which sets up a result he promises to publish either way: if 13.5 million installs and tens of thousands of reviews are worth anything on Play in 2026, that should show up before it shows up on a game nobody has heard of.

The two variables moved in opposite directions. Discovery got worse, and the launch checklist cannot substitute for a distribution slot that no longer exists. Monetization got better, and the thing that made 2014 structurally hopeless is finally fixed. “I do not think 2014 was better,” he writes. “It was easier in one way and impossible in another.”

All the CSVs and raw report zips are linked in the post. He is also explicit about which claims he can prove and which ones he cannot.