Why API impressions read lower than App Store Connect
The gap is real, it is Apple's, and it is not a bug you can fix at your end. Apple's analytics reports drop rows that fall below its privacy thresholds, so impressions read low — and because the threshold removes a roughly fixed amount of data rather than a percentage, the smaller your app, the larger the gap looks. On a steady iOS app we measured the API within 1.4% of the dashboard. On a low-volume macOS app, the same method came out 39% low.
What we measured
Over a week in November 2026, we compared the App Store Connect API against App Analytics for two apps on the same account:
| App | Impressions vs App Analytics | Product page views | First-time downloads |
|---|---|---|---|
| A steady iOS app | within 1.4% | matched exactly | matched exactly |
| A low-volume macOS app | 39% low | matched exactly | matched exactly |
Only impressions diverged. Everything else reconciled exactly, which is what makes this diagnosable rather than mysterious — if the API were simply behind, or on a different day boundary, downloads and page views would drift too.
Why it happens
The API's rows for the low-volume app were all small search and browse counts: exactly the shape of data Apple's privacy rules suppress. Apple does not publish a threshold, but the behaviour is consistent with rows below some small count being dropped rather than rounded.
That produces the pattern worth remembering: the gap is roughly a fixed quantity, not a fixed percentage. A handful of suppressed rows a day barely registers against a busy app's impressions and swamps a quiet one's.
We also ruled out the obvious alternative. Apple's Web Preview report carries no impressions at all, so impressions arriving through a different report is not the explanation.
The other thing that is not the explanation
Time zones. Analytics days are in UTC and Sales and Trends days are in Pacific Time, which genuinely shifts figures between dates — but it shifts them between adjacent days rather than removing them, and it does not touch the analytics-to-analytics comparison here. If your impressions are low and your downloads match, you are looking at suppression, not a day boundary. More on that in why the latest days keep changing.
One way to make it worse by accident
Apple's Discovery and Engagement report has both a Counts column and a Unique Counts column. Impressions are the sum of Counts. Summing Unique Counts instead gives a smaller, different number that will never match anything, and the mistake is easy to make because both columns look plausible.
What to do about it
- Treat impressions as indicative, not exact, especially below a few hundred a day.
- Compare like with like over time. The suppression is consistent, so week-on-week movement in the API's impressions is still meaningful even when the absolute figure is low.
- Don't build a conversion rate on it without saying which source the impressions came from. An impressions-to-downloads rate mixing an API numerator with a dashboard denominator is wrong in a way nobody will spot.
AppStato shows impressions from Apple's analytics reports and says so in the metric's own definition, rather than quietly presenting a figure that will not match the dashboard. See AppStato.