What attention data can tell advertisers

An editorial interpretation of the conversation, distinguishing the guest’s arguments from verified business outcomes.

The question that led to Lumen

Mike Follett traces Lumen Research’s origins to a recurring agency argument: how large a brand’s logo should be. Across London, New York and Mumbai, he found confident opinions without a reliable basis for choosing between them. A Volkswagen presentation in India sharpened his doubts about strategy that could be rationalised around whichever creative work a client bought. Returning to university helped him combine experimental methods with advertising experience. His central research objection is that asking people to study an ad can remove the very problem advertisers need to understand: whether anyone notices it in ordinary life.

How exposure differs from a human glance

Follett describes attention as scarce, brief and frequently distracted. An opportunity to see an ad therefore differs from actually looking at it, with consequences for both media selection and creative development. He argues that established formats deserve comparison on observed behaviour rather than assumptions about their age. He cites Lumen’s work with Dentsu as evidence of strong attention and recall for linear television, while discussing cinema’s viewing conditions and live sport’s shared, appointment viewing. Cricket, in his view, is especially suited to advertising. These are his interpretations of the research and media environments, rather than universal guarantees for every placement.

Creative quality inside the media average

Creative quality remains central to that comparison. Follett warns that media averages conceal the difference between work that attracts attention and work that loses it. His illustration of stronger creative increasing the effective value of a media budget is an economic analogy, not a promised cash saving. He challenges the idea that platform optimisation alone can replace creative ambition or build lasting value. He also distinguishes attention as useful buying information from attention as the sole currency on which media should be traded. According to Follett, agency planning tools and verification partners already offer ways to apply Lumen’s data.

Attention models and what they cannot prove

The discussion separates observed panel behaviour from predictive models applied to campaign data. Follett reports consistency in platform use across several markets while stressing variation in creative performance. On generative AI, he questions systems that produce and assess ads without meaningful human response. For advertisers exploring attention measurement, he suggests requesting planning data, investigating partner measurement products and testing attention-informed buying against business as usual. He considers human attention harder to manipulate than technical proxies, and describes reducing a publisher’s ad load to improve attention and support higher prices for remaining inventory.