On November 16, 2022, the Advertising Research Foundation (ARF) gave a webinar in conjunction with the Market Research Council in which they revealed that very large percentages of “advanced audiences” are invalid. Here are two slides as examples. A score of “0.33” means that only a third of the targets were valid. If basic demographics are that distorted, what hope is there for modeled targets built on them?
Unique among providers of advanced audiences, RMT/Semasio makes no use of such data: we are taking full text grabs of the content that ID uses. We are semantically analyzing their motivations based on the content. This is not lookalike modeling. And none of it is based on demos.
Other than RMT/Semasio, the others offering “advanced audiences” are using clues to infer demographics, then they create all kinds of audiences they know people want, by modeling lookalikes from demographics that are themselves impaired.
But the ARF has just shown the industry how much people are willing to fool themselves. Low validity for the base demos which support the lookalike modeling. Hundreds of millions of dollars going for those lookalike targets.
RMT/Semasio is the better alternative.
In 2018 Simmons did a study validating RMT methodology and showing that RMT DriverTags increase the predictivity of demo-based modeling by +83%. Happy to share that study as well as the other third-party studies proving that DriverTags are real science (Nielsen NCS, ARF Cognition Council, 605, Neustar, in addition to the Simmons study).
Demographics themselves, if validity were 100%, only account for 6%-16% of variance in sales/usage data. (Sources: Simmons, Henry Assael NYU, studies upon request) The ARF Cognition Council study found that RMT Motivational Types account for 48% of IRI sales.
The Neustar study was a random control trial and showed RMT/Semasio was +95% higher than popular lookalike targets in Return on Ad Spend.