Meta, Google Ads and AppsFlyer (AF) can report different install and purchase totals without a broken event integration. First align the action, date and attribution definition. This guide adapts Xiaowei's 2023 article; its historical default lookback periods are not universal 2026 settings.
1. One user can receive different channel credit
A user may click a Meta ad, click a Google ad and then first open the app. Each platform can report a conversion it can observe, while an MMP assigns a source under its cross-channel rules. Do not add platform install totals to estimate unique users. See AppsFlyer's Meta discrepancy guide.
2. Lookback windows differ
Record the current click-through and view-through windows in each platform and the MMP, separately for installs and in-app events. A conversion can appear under a seven-day window but fall outside a one-day window. The old article's 28-day example is not a current default to copy.
3. Interaction, download and first-open dates differ
Some Google Ads conversion views group by ad interaction date, while AppsFlyer records an install at first open. A click on day one and first open on day three can therefore appear on different dates. Compare a full period and inspect both date definitions; see AppsFlyer's Google Ads guide.
4. Align time zones
Write down the ad-account, MMP-app and warehouse time zones. Convert each export to one time zone and use the same day boundary before reconciling.
5. Reinstalls and retargeting may appear separately
A return can be a reinstall, re-attribution or new acquisition, depending on user history, the re-attribution window and retargeting interaction. Inspect the relevant report columns and AppsFlyer's reinstall definitions.
6. Then inspect event delivery and filters
If a material gap remains after alignment, inspect duplicate or missing SDK/S2S events, names and values, modeled data, plus country, OS, campaign and organic filters. Compare raw event counts before channel-attributed counts.
A repeatable reconciliation order
- Define the metric and deduplication rule.
- Use one period and time zone, allowing for the full attribution window.
- Align lookback, interaction versus conversion date and acquisition versus retargeting views.
- Break down the remaining gap and inspect raw events and settings.
For event setup, see our Meta app ads checklist and Google App Campaigns guide. Adapted from Xiaowei's original article.
FAQ
Why can Meta plus Google installs exceed AppsFlyer installs?
Each platform can report observed ad interactions while the MMP assigns one source under cross-channel rules. Align install definitions and windows before calling it duplication.
Why is the daily gap large but the weekly gap smaller?
Interaction versus first-open dates, time zones and delayed reporting may move events across days. Normalize dates and extend the observation period.
Does a reinstall count as a new install?
It depends on the MMP re-attribution window, user history and retargeting engagement. Check the current report classification.