The Rise of Marketing Attribution

Marketing attribution did not start as a discipline. It started as a default setting nobody questioned.

When Google Analytics launched in 2005, built on the acquired Urchin platform, it gave the entire industry a single, simple habit: whichever channel got the last click before a sale got the credit. It was easy to explain and easy to build dashboards around, and for years almost nobody asked whether it was true. A customer who saw a display ad, read three blog posts, clicked an email and then typed the brand name into Google would have their entire journey credited to that final branded search. The ad, the content and the email got nothing.

By the early 2010s, bigger budgets made that gap too expensive to ignore. Vendors like Visual IQ, C3 Metrics and Convertro built multi-touch attribution models that tried to split credit across every touchpoint a customer interacted with, using linear, time-decay or algorithmic weighting instead of a single winner-take-all click. Convertro’s acquisition by AOL in 2014 was a signal that attribution had become a business worth owning, not just a reporting feature bolted onto analytics software.

Multi-touch attribution ran into a wall that no amount of clever modeling could fix: the platforms themselves would not cooperate. Facebook and Google each measured conversions inside their own walled gardens and each, unsurprisingly, tended to claim more credit than an independent view would give them. Marketers ended up staring at three dashboards that added up to well over 100% of their actual sales, a running joke inside performance marketing teams for most of the decade.

Then privacy regulation and platform policy did in a few years what modeling debates never resolved. GDPR took effect in May 2018 and forced consent into the center of European data collection. Apple’s Intelligent Tracking Prevention had already started shortening cookie lifespans in Safari the year before, quietly eroding the tracking window attribution tools depended on. The bigger shock came in April 2021, when Apple’s App Tracking Transparency prompt went live with iOS 14.5 and asked every iPhone user, directly, whether apps could track them across other apps and websites. Most said no. Meta later disclosed the change would cost it roughly ten billion dollars in 2022 revenue alone, a blunt measure of how much of the attribution industry had been quietly built on cross-app tracking that a single dialog box could switch off.

Google’s own roadmap added to the pressure. Its 2020 announcement that Chrome would phase out third-party cookies kept attribution vendors on notice even as the timeline slipped repeatedly. And in March 2022, Google confirmed that Universal Analytics would stop processing data the following year, pushing the entire industry toward GA4, a platform built around events and modeled conversions rather than the session-based, last-click reporting that had defined web analytics since 2005.

None of this killed attribution. It forced it to grow up. By 2022, the smarter teams had stopped hunting for one perfect model and started triangulating: keeping click-based attribution for tactical, day-to-day reads while leaning on marketing mix modeling and incrementality testing, both of which work on aggregate data and do not depend on tracking any individual person, for the bigger strategic calls. That combination, imperfect and constantly recalibrated, is what took the place of the tidy last-click number everyone had trusted for over a decade.

The lesson sitting underneath all of it is not about tracking technology. It is that attribution was always an estimate dressed up as a fact, and 2022 was the year enough of the industry finally admitted that out loud.