The measurement problem nobody wanted to solve
When Matt Scharf joined Adobe in 2013 as a manager of display media operations and analytics, the industry's attribution problem was hiding in plain sight. Marketers were handing out conversion credit to ad impressions that consumers had never actually seen. The model counted the ad. It didn't ask whether anyone was looking.
Scharf asked. He assembled what his team called Project Iceberg - a reference to the idea that the most important stuff in digital advertising is beneath the surface. Working with MarketShare and DoubleVerify, his team built a viewability-adjusted attribution model that stripped credit from out-of-view impressions. The finding was blunt: some publishers were great at targeting but terrible at inventory quality. Others were the reverse. Without isolating the two, marketers had no real picture of what they were buying.
"Attribution models aren't ingesting viewability data so they cannot exclude the events that are not in view."Matt Scharf, 2014 - AdExchanger
In 2015, the Association of National Advertisers handed Adobe its Genius Award for Pioneering Analytics. Scharf wasn't a director yet. He was a senior manager. The recognition came not for a product launch or a campaign - it came for exposing a structural flaw in how the entire industry measured digital media performance. That kind of win tends to set a career trajectory.
He stayed at Adobe. A decade later, he's VP of Growth Marketing Performance and leads the company's Global Media Center of Excellence. The Iceberg work eventually grew into something much larger - the internal measurement framework that Adobe's marketing team built to run its own programs, which became the foundation for Adobe Mix Modeler, now a commercial product. The team used their own tools on their own budget, delivering results that made the case for selling the capability to the market.