In this talk, Kumar will explore the gap between measurement quality and decision quality, and why closing it requires a different kind of discipline than most organisations practice today. He will walk through five problems he has seen repeatedly across large-scale digital investment portfolios: the precision trap, where teams chase model accuracy while the business question sits unanswered; measurement fragmentation, where MMM, attribution and experiments each produce precise but contradictory answers; the wrong question, where "which number is right?" stalls progress while the capital allocation decision waits; confidence intervals that never leave the analyst's working file; and the reframe from measurement as truth-seeking to measurement as risk management. Drawing on his experience governing more than GBP 70m in annual digital media investment at Sky, Kumar will show how the shift from "prove you are right" to "bound the downside of being wrong" changes what you build, what you report, and what you ask for in the boardroom.

In this talk Tom will talk about the key mistakes marketers make when attempting to measure performance in AI - whether it be AI-augmented search, ChatGPT, Gemini, or specialist tools.

Tom will compare key differences with traditional SEO measurement, and new challenges specific to AI. In both cases, he'll offer tactical and technical workarounds to overcome these issues, as well as a strategic mindset for AI as it sits alongside other marketing channels.