When you first started, attributing an install to the last ad a user clicked was simple enough. But as you scale past 10,000 users, that model breaks down. Your users now see your app on TikTok, hear about it from an influencer, and then search for it on the App Store. The last click only tells the end of the story, not the whole plot.

To scale effectively, you need to understand the entire customer journey. This means moving from a simple, single-touch attribution model to a multi-channel one. The first step is to understand the two core methods for matching a user to an ad: deterministic and probabilistic.

Deterministic attribution is the gold standard. It uses unique device identifiers, like Apple's IDFA or Google's Advertising ID (GAID), to create a direct, one-to-one match between a specific user and an action. If a user with GAID 'xyz123' clicks an ad and then installs your app, the connection is certain.

Probabilistic attribution, on the other hand, is an educated guess. When a unique ID isn't available, this method uses non-unique data points—like IP address, device type, and operating system—to create a likely profile or 'fingerprint' of a user. It then matches the fingerprint of an ad click to the fingerprint of an app install that happens shortly after. It's not perfect, but it's a powerful tool when identifiers are absent.

The move away from device identifiers, led by Apple’s App Tracking Transparency (ATT) framework, has fundamentally changed attribution. In this new landscape, Apple's SKAdNetwork (SKAN) has become the primary tool for privacy-centric measurement on iOS.

SKAN works by letting the ad network know a conversion happened, but without revealing which specific user converted. It's like getting a report that says 'a person who saw your ad in London installed the app', but with no name or address. This preserves user privacy but introduces challenges for marketers, such as delays in reporting and very limited data on what users do after they install.

Because of the data gaps created by SKAN and other privacy measures, relying solely on device-level attribution is no longer viable. Marketers must now embrace a more statistical approach to understand campaign effectiveness. This is where incrementality testing comes in.

Incrementality testing answers a simple but critical question: did my ads actually cause new installs, or would those users have converted anyway? It works by dividing your target audience into two groups: a test group that sees your ads, and a control group that doesn't.

You then measure the conversion rate for both groups. The difference between them is the 'lift'—the incremental impact of your advertising. This method doesn't rely on user-level tracking. Instead, it measures the statistical impact of a campaign on a population, making it a privacy-safe way to prove the value of your marketing spend.

Another major challenge in scaling is cross-device tracking. A user might see an ad on their work laptop, research on their personal tablet, and finally install on their phone. Without a persistent identifier that links these devices, it's difficult to connect the dots. Probabilistic methods can help bridge this gap by matching device characteristics, but their accuracy can vary. The goal is to build a holistic view of the user's path to conversion, however fragmented it might be.

An attribution window is the period after a user interacts with an ad during which a conversion can be credited to that interaction. A typical window might be 7 days for a click and 1 day for a view. However, a one-size-fits-all approach is inefficient when scaling.

Consider the user journey for your specific app. A hyper-casual game might have a very short conversion cycle, meaning a 1-day attribution window is plenty. A finance or investment app, however, might involve more research and consideration. Users could take weeks to convert after their first click. In this case, a longer attribution window (e.g., 28 days) would be necessary to capture the full impact of your top-of-funnel campaigns. Customising your windows ensures you don't over-credit short-term channels or under-credit ones that build long-term value.

Time to see what you've learned about multi-channel attribution.

By moving beyond last-click attribution and embracing these more sophisticated models, you can allocate your budget with greater confidence. You'll know which channels work together to drive high-quality users, setting you up for sustainable growth as you scale.