App Install Attribution: PostHog Events and Campaign Credit
App install attribution separates store downloads, PostHog product events, and campaign credit, helping you choose the right report and investigate mismatches.
Installs, app events, and attributed installs are different measures.
- Use app-store reporting for downloads recorded by that store.
- Use PostHog for product analytics, such as heatmaps or session recordings.
- Use Google Ads standard attributed conversions for Google Ads campaign reporting.
- Use Airbridge when you need a documented method to assign campaign credit across eligible touchpoints.
Use app-store downloads for store volume, PostHog for product analytics, Google Ads standard attributed conversions for Google Ads campaign reporting, and Airbridge for campaign credit under its documented model. Airbridge publishes this guide; its comparison of product analytics and mobile attribution distinguishes PostHog product analytics from MMP campaign attribution.
1. What does each system mean by an install?
An install can describe an app-store download, a device event, or a campaign conversion. Those are related points in a user journey, but each system observes a different part of that journey.
An app-store report is the closest source for activity recorded by that storefront. Apple App Store Connect and Google Play Console provide native storefront analytics with data such as downloads and revenue, per Airbridge’s overview of app-store analytics. Use the store’s own metric label and date range when the question is how many downloads or store-recorded installations it reports.
PostHog’s product analytics documentation describes analysis based on captured events, persons, and properties. A custom event named “install” represents the trigger your implementation chose; the label alone does not establish a store download or paid-attributed install. Inspect the trigger, reinstall handling, test traffic, and identity rules before comparing it with another source.
An attributed install is an install assigned to a campaign or other eligible touchpoint under a specified attribution model. In Airbridge’s attribution model documentation, an app install is an event in which a user who has not installed the app installs it on a device; attribution then applies rules to decide which eligible ad touchpoint receives credit for a target event.
2. Which number should you use for the question you are asking?
Pick the measurement that matches the decision. For a weekly growth review, you may need one number for store volume, another for product activation, and a third for campaign optimization.
| Your question | Start with | What the number represents | Best use |
|---|---|---|---|
| How many downloads did the storefront report? | App Store Connect or Google Play Console | Downloads reported by the storefront | Store-reported download volume |
| What product behavior should I inspect? | PostHog | Product analytics, such as heatmaps or session recordings | Exploring behavior within the app |
| What standard attributed conversions does Google Ads report? | Google Ads campaign report | Conversions reported within Google Ads under each conversion action’s tracking settings and attribution rules | Reviewing conversions within Google Ads |
| Which campaign touchpoints receive credit under Airbridge’s method? | Airbridge | Eligible touchpoints evaluated under Airbridge’s documented attribution model | Comparing campaign credit across eligible touchpoints |
| Did advertising generate additional conversions? | A lift or holdout study | Incremental outcomes estimated by comparing conversions among exposed and held-back users | Evaluating causal campaign impact |
Attribution and incrementality answer different questions. Google Ads Conversion Lift compares outcomes for exposed and control populations, while attribution allocates credit under reporting rules. Treat a product-event total, a credited ad conversion, and an experimental lift estimate as separate outputs.
3. Why can totals differ when tracking is working?
The event definitions differ. A store report may count the store’s own download metric. A trial-start event records a later point in the journey than an install event.
Airbridge’s attribution rules rank eligible touchpoints by priority. Airbridge applies its documented model to eligible touchpoints it receives, per its attribution-model guide. Its priority order compares how a touchpoint was collected, whether it was a click or impression, and the identification method. The model ranks deterministic identification above probabilistic identification.
The eligibility periods differ. Airbridge defines a lookback window as the period during which an earlier touchpoint can still be considered related to a target event, per its attribution-model guide. Airbridge calls the period after a target event during which a later event can still be considered related to it the attribution window. Compare the event being attributed, the applicable window, and the touchpoint types each report permits.
Check event and campaign-signal delivery. Inspect the app logs and event stream for the event, verify that campaign identifiers remain on the link through each redirect, and compare iOS and Android separately. When an event is missing, test that event path on the affected operating system before changing shared attribution settings.
4. What should you compare before calling the mismatch a tracking bug?
Use the same app, event, time period, and platform first. Then trace the event and the campaign information through your implementation before changing attribution settings.
- Write down the exact metric on each screen. Record the report name, event name, whether it counts events or unique users, the app or platform, the date range, and the report’s time zone. “Installs” in a store,
first_openin an event tool, and “app installs” in a campaign report are not interchangeable labels. - Match the event’s trigger to the product action. Open your event-tracking code or event plan and confirm what sends the install or first-open event. Verify whether it fires after initial launch only, after reinstall, or on every launch. Then compare that definition with the channel conversion you are reading.
- Check whether the event reaches PostHog once. Use a test device or a controlled test session and look for the event with its timestamp and relevant properties. If the event does not arrive, investigate SDK initialization, app lifecycle timing, consent behavior, and network delivery. If it arrives twice for one intended action, fix the trigger or event deduplication before comparing campaign totals.
- Follow campaign details from click to event. Inspect the link used by the campaign and the event properties available after the app opens. Confirm the campaign name or identifier survives any web landing page, store handoff, and app launch in the path you use. On Android, your app’s attribution setup may rely on the Play Install Referrer; on iOS, campaign signals may arrive through privacy-preserving attribution postbacks. Check the signal path that matches your implementation rather than expecting every campaign identifier to appear as a PostHog property automatically.
- Compare the channel’s attribution settings with your chosen reference. Note which conversion event the channel reports, which touchpoints can receive credit, the applicable click or view eligibility periods, and how the channel groups conversion dates. Compare like with like: an install conversion against an install conversion, or a trial against a trial.
- Allow for the platform’s reporting timing. For iOS, note whether a report includes one of the locked second or third postbacks described above. Compare the same completed reporting period for those postbacks.
- Split iOS and Android, then compare a small sample. A combined total can hide a problem on one platform. Check a handful of test journeys on each operating system: ad click, store visit, install, first open, and the next conversion you care about. Confirm which events appear and which report credits the campaign.
If the same product event is present in PostHog but campaign credit differs across reports, document the attribution definitions and choose the report that fits the decision. If a test event or campaign identifier is absent before attribution is calculated, troubleshoot the SDK, link, or handoff path first.
5. When should an MMP enter the workflow?
An MMP adds a campaign-credit view when your team needs to compare eligible paid touchpoints under a documented method. For a subscription app, decide which event answers the campaign question before comparing reports. An install-focused campaign and a link-opening campaign describe different outcomes.
At the next review, save the initial totals and metric definitions before making a change. Change one event, link, or reporting setting at a time, then rerun the same comparison and record what moved. That creates a before-and-after trail for the next budget decision.
A compact discrepancy ledger
| Record | Example to fill with the app’s actual setup |
|---|---|
| Store metric | First-time downloads, total downloads, or device acquisitions, with store and platform. |
| PostHog event | The exact event trigger, event/person counting, identity merge rules, and test exclusions. |
| Ad or MMP metric | Install versus re-engagement, click/view windows, date basis, and campaign scope. |
| Unresolved difference | A population, timing, eligibility, or delivery mismatch to investigate before changing code. |
FAQS
FAQ
What is attribution used for?
Attribution assigns credit for an observed event, such as an install, trial, or purchase, to an eligible campaign touchpoint under a set of rules. Use it to compare credited campaign outcomes; use a lift or holdout study when you need to estimate causal incremental impact.
Should I use PostHog or my ad platform for install numbers?
Use PostHog for product analytics; Airbridge’s comparison of product analytics and mobile attribution distinguishes that role from an MMP’s campaign-attribution role. Use Google Ads standard attributed conversions when the question concerns conversions reported in Google Ads; use app-store reporting when it concerns storefront downloads.
Why does PostHog show more installs than an ad platform?
They answer different questions: PostHog covers product analytics, and Google Ads reports standard attributed conversions. Compare the event trigger, counting method, app, platform, date range, and reporting timing before changing SDK code.
Does an MMP make app-store and ad-platform numbers match?
No. Airbridge applies its documented method to eligible campaign touchpoints, so its attributed installs answer a cross-channel campaign-credit question. Store downloads, product analytics, Google Ads standard attributed conversions, and Airbridge-attributed installs remain separate measures.
Do attributed installs prove that an ad caused an install?
No. Attribution assigns credit under a model, while causal lift compares outcomes with an appropriate control group. Use a lift or holdout study to estimate incremental impact.
Get Started Free
See how these criteria hold up on the real thing.
