TV attribution results can be all over the place due to inconsistent data inputs, says a study issued by the Coalition for Innovative Media Measurement with the 4A’s Media
Measurement Task Force.
“Methodology, rather than technology, is [the] root cause of TV attribution outcome differences,” according
to the study, which was conducted by Sequent Partners and Janus Strategy & Insights.
Eleven leading TV attribution providers — which were not
named in the study — were found to have highly inconsistent occurrences and exposure data. Research says the accuracy of spot detection and all exposure data elements — gross rating
points (GRPs), reach, frequency — differ from provider to provider.
In addition, the report says, brand-lift outcomes are not consistent among providers
because of the differences in both occurrence and exposure data. Also, methodology of converting data into final ad occurrence and exposure data — including weighting, editing
and other data-processing rules — is believed to be the cause of the differences between providers.
The report calls for standardization, such as commercial IDs
similar to Ad-ID, for identifying occurrences and in defining exposure and reach.
"While the findings of our study do not necessarily tell us how to solve for the
attribution inconsistencies our industry currently faces, they do clearly indicate the need for the standardization," states Alice K. Sylvester, partner, Sequent Partners.
Sylvester says this pertains to naming, definitions, and categorization and quality assurance procedures.
“This report provides a roadmap for
providers to improve their offerings and at the same time reduce the dramatic differences between providers,” says Howard Shimmel, president of Janus Strategy & Insights.
The paper — titled “Getting Attribution Right: An Exploration and Best Practices for Television Data Inputs in Attribution Modeling” — was presented as part of
The ARF’s AudiencexScience virtual event.