Updated Nov 2, 2026
Not every day. AppCurves has read the exact lifetime download counter behind Google Play listings every night since Aug 4, 2026, for between 114,000 and 192,000 apps a night. On a median night, 38% of those counters changed from the night before. On six nights in September, not one did, including three nights in a row, Sep 21 to 23.
Every Google Play listing carries an exact lifetime download counter in its page data; the store shows only a rounded band such as 1B+ (where to find it). Each night we read that counter for the apps on Google Play's top charts across the 203 markets we scan. Chart membership changes, so the number of apps read varies from night to night.
For each app read on two consecutive nights, we compare the two readings. A counter changed if the two numbers differ. We count the same way for the rating count, which tells us whether the listing itself was fresh that night. Nothing here is modelled or estimated: every figure is a count of readings we kept.
Across 88 nightly comparisons from Aug 6 to Nov 1, 2026, the share of counters that changed ranged from 0% to 81% (Oct 15). The median was 38%. In other words, a typical app's counter changes every two or three nights, not every night, and the updates arrive in batches rather than spread evenly.
Low nights happen in every month: 3.8% on Aug 29, 3.0% on Oct 19, 0.6% on Oct 30. Completely flat nights were rarer. We recorded six, all in September: Sep 7, Sep 8, Sep 13 and Sep 21 to 23.
From Sep 19 to Sep 28, the counters barely moved. Three nights in that stretch had no changes at all, and no other night in it reached 10%. Normal batches resumed on Sep 29, and from Oct 3 to Oct 9 every night was above 45%.
| Reading | Apps compared | Counters changed | Share | Rating counts changed |
|---|---|---|---|---|
| Sep 17 | 156,688 | 7,125 | 4.5% | 41,274 |
| Sep 18 | 114,947 | 46,625 | 40.6% | 39,770 |
| Sep 19 | 116,046 | 8,767 | 7.6% | 42,356 |
| Sep 20 | 113,535 | 11,199 | 9.9% | 41,726 |
| Sep 21 | 117,083 | 0 | 0.0% | 39,262 |
| Sep 22 | 112,887 | 0 | 0.0% | 35,935 |
| Sep 23 | 112,757 | 0 | 0.0% | 35,267 |
| Sep 24 | 112,810 | 4,383 | 3.9% | 35,523 |
| Sep 25 | 116,834 | 8,423 | 7.2% | 36,709 |
| Sep 26 | 115,950 | 835 | 0.7% | 37,569 |
| Sep 27 | 120,066 | 2,098 | 1.7% | 37,566 |
| Sep 28 | 107,346 | 5,472 | 5.1% | 34,791 |
| Sep 29 | 106,392 | 43,516 | 40.9% | 34,950 |
| Sep 30 | 108,363 | 4,650 | 4.3% | 38,192 |
| Oct 1 | 115,148 | 58,651 | 50.9% | 40,777 |
The listings themselves were not stale. On each of the three flat nights, the rating count changed on 35,000 to 39,000 apps, about a third of those we compared. When we checked live listings on Sep 24 from two networks and four storefronts (US, Germany, Japan, Brazil), the counters matched our readings to the digit. Google simply had not updated them.
ChatGPT's counter read 1,567,613,533 on Sep 16 and the same number on each of the next ten nights. Over those eleven readings, its rating count rose from 60,312,341 to 61,839,284, an extra 1.5 million ratings. On Sep 27 the counter jumped to 1,589,755,232: 22,141,699 downloads appeared in a single update.
TikTok followed the same pattern: 4,676,450,126 from Sep 16 to Sep 26, then 4,688,806,830 on Sep 27, a jump of 12,356,704. The jumps show the downloads kept coming during the pause; the counter just did not show them until the next update.
A flat day is not a day without downloads. If an app's counter did not move overnight, the likeliest reason is that Google did not update it, not that nobody installed the app.
A big jump is usually several days at once. Daily changes in the counter are lumpy. To compare apps or measure growth, use a week or more, and compare apps over the same dates so they share the same update batches.
The band moves even less often. The Play Store's rounded band changes only when the counter crosses the next threshold, so it can sit unchanged for months.
Play Console is a different measure. Developers see install metrics in Play Console that are defined differently from the public counter; the two will not match (why).
Every night in this study is in one file: google-play-download-counter.csv. Columns: the reading date, apps read that night, apps also read the night before, how many of those counters changed, the share as a percentage, and how many rating counts changed.
You are welcome to use the data and the charts you make from it. Please cite AppCurves and link to this page. For any single app, its full record is on its page: ChatGPT, TikTok, or search any Android app.