What a Paid Install Is
A paid install is an app install credited to marketing you paid for. A user taps a Meta ad, a Google App campaign, a TikTok video, an affiliate offer or a creator's tracked link, and the attribution platform matches their install to that source. Everything with no tracked source is an organic install.
Paid installs are the unit most growth economics are built on. Cost per install is spend divided by paid installs. ROAS compares revenue from paid installs with the spend that bought them. CPI deals with networks are billed on them.
How Paid Installs Are Attributed
Attribution follows a set order:
- Deterministic signals. A click ID in the store referrer, a device ID where the user allowed tracking, or a claim from a self-attributing network such as Google, Meta or Apple Ads.
- Eligible clicks. Clicks from tracked links inside the channel's lookback window. When several channels have clicks, the last eligible click wins.
- Impressions. Where view-through attribution is on, an impression inside the view-through window can win, but only if no click can be credited.
- Estimated matching. When no deterministic signal exists, IP and user-agent matching can link a click to an install within a short window.
If nothing matches, the install is organic. If something matches, it is paid and credited to exactly one source.
Paid Installs on iOS
On iOS, many paid installs never get user-level attribution, because users who decline ATT cannot be matched by device ID. Those installs are measured in aggregate through SKAN and AdAttributionKit postbacks, which report installs by campaign with a delay and no user detail. Apple Ads installs are the exception, because Apple's AdServices framework attributes them without ATT.
A complete view of paid iOS installs therefore combines device-level attribution for consented users with SKAN data for the rest.
Reconciling with Ad Networks
The paid install count in your MMP will rarely match what each network reports. Networks count installs inside their own windows, often include view-through and modeled conversions, and do not know when another network's click came later. The MMP looks across every network and credits one source per install. Some gap is normal and mostly expected. See attribution discrepancy.
Paid Install Quality
Not every paid install is worth what it cost. Measure quality after the install, not just the count:
- retention at D1, D7 and D30 against organic users,
- revenue and return on ad spend by campaign and creative,
- click-to-install time distribution by source,
- share of installs flagged by fraud rules.
Fraud is concentrated in paid installs, because that is where payouts are. Click injection and click spamming steal credit for organic users, and SDK spoofing fakes installs that never happened.
Paid Installs in Linkrunner
Linkrunner attributes installs deterministically first, using the Google Play Install Referrer, ad click IDs such as gclid and fbclid, device IDs where the user allows it and Apple AdServices for Apple Ads, and falls back to IP and user-agent matching only when no deterministic signal exists. Spend syncs automatically for Meta, Google and Apple Ads, so CPI and ROAS are calculated down to ad set, creative and keyword.
Fraud Control, on every plan, flags suspicious installs with rules and templates and can withhold postbacks so networks do not optimize toward bad traffic.
On pricing, attributed installs are billed, including re-attributed reinstalls. Re-engagements are free, and the first 25,000 attributed installs are free with every feature included. See mobile attribution and pricing.