What is the difference between App Units and Installations in App Store Connect?

App Units (Sales and Trends) count first-time acquisitions of your app — one per Apple Account, excluding redownloads and restores. Installations (App Analytics) count the app actually landing on devices that opted in to share analytics — including redownloads and additional devices. Different questions: App Units ≈ how many new people got it; Installations ≈ how often it hit a device.

Why don't my App Store Connect downloads match my own analytics?

Different denominators: App Analytics engagement metrics cover only opted-in devices, your in-app SDK counts app launches (missing people who never opened it and double-counting reinstalls depending on setup), and Sales and Trends counts Apple Accounts. Add Apple's 2–3 day reporting lag and time-zone differences, and disagreement is the expected state. Pick one source per question and compare it only with itself over time.

Which download metric should I use to judge ASO?

Total Downloads (or First-Time Downloads) in App Analytics, because it sits in the same reporting family as impressions and product page views — the funnel stays internally consistent: impressions → page views → downloads from one system. Use App Units when you care about unique new customers, and proceeds when you care about money.

FunnelHound / Guides / App Units vs Installations

App Units vs Installations vs Downloads: why your numbers never match

By Emin — indie iOS developer building FunnelHound · Published 13 Sep 2026

Apple gives you three different numbers for "how many people got my app" — App Units, Installations, Total Downloads — and they never agree. None of them is wrong. They're answering different questions, and the panic starts when you compare answers across questions. Here's the decoder ring.

The three metrics, decoded

App Units (Sales and Trends): first-time acquisitions, one per Apple Account. Redownloads and restores don't count. Closest to "new customers". This is the report family that also carries your money.
Installations (App Analytics): the app landing on a device — from opted-in devices only. Includes redownloads and second devices. Closest to "how often the app arrived somewhere".
Total Downloads (App Analytics): first-time downloads plus redownloads, in the same reporting family as impressions and page views. Closest to "store funnel outcomes", and splittable into first-time vs redownloads.

Why they legitimately disagree

Four honest reasons, no bugs required. Opt-in coverage: chunks of App Analytics engagement data come only from users who agreed to share — a structural undercount versus Sales and Trends. Redownloads: a user returning after deleting counts in Installations and Total Downloads, not in App Units — apps with comeback traffic see wide gaps. Family Sharing and multi-device: one purchase, several installations. Timing: reports close on different schedules with Apple's usual 2–3 day lag, so "yesterday" is a different day in different reports (the same trap covered in why did my downloads drop). Your RevenueCat or Firebase numbers add a fifth question — "who launched the app with our SDK alive" — which matches none of Apple's three. Five instruments, five songs, zero scandals.

Which number for which job

Judging ASO and the store funnel: Total Downloads / First-Time Downloads — it shares a reporting family with impressions and page views, so the funnel divides cleanly. Never mix families in one ratio: analytics page views over Sales-and-Trends units produces a very confident nonsense percentage.
Counting new customers: App Units. Investors and spreadsheets mean this one.
Counting money: neither — proceeds. Downloads of any flavor pay no rent.
Trend-watching: any of them, compared only with itself. Every metric is a fine thermometer as long as you don't swap thermometers mid-fever.

The one-consistent-funnel shortcut

The practical fix for metric soup: pick one internally consistent chain and live in it. FunnelHound builds its funnel entirely from matching report families — impressions → page views → downloads → proceeds, with deletions alongside — so every stage divides by the stage above it without a definitions asterisk. The three-way disagreement doesn't disappear (it's structural), but it stops being your morning problem.

One funnel, consistent definitions, no soup

FunnelHound assembles your funnel from matching App Store Connect report families — the ratios actually mean something.

Get FunnelHound

Share this on X

Data notes: metric definitions per Apple's App Store Connect documentation as of 2026; Apple occasionally renames and re-bases analytics metrics, so re-check definitions after dashboard redesigns. Opt-in analytics coverage varies by audience and iOS version.