YouTube partners with Nielsen Catalina Solutions (NCS) to measure in-store sales lift
A new machine-learning model points to the future of measurement.
Nielsen Catalina Solutions (NCS) has partnered with Google’s YouTube to help consumer packaged goods (CPG) advertisers measure the offline sales impact of their campaigns on the streaming video service. This is part of Google’s larger effort, involving other Measurement Partners, to show offline sales lift from exposure to YouTube ads.
NCS is a joint venture of Catalina and Nielsen. Catalina recently introduced its own offline attribution solution, which shares some common elements with NCS’s methodologies.
Two approaches to measurement. Essentially, NCS has two approaches to offline sales measurement, neither of which rely on mobile location data. The more “traditional” approach uses retail purchase and loyalty card data that is matched to U.S. households and uses Experian as a third party agent to do matching and anonymizing on the back end. It utilizes a test and control approach to see whether and how much the campaign being measured affected the exposed group’s in-store purchase behavior.
The methodology developed for use with YouTube, by comparison, still uses retail purchase data but is considerably more advanced, according to NCS EVP of Strategy, Carl Spaulding. The company calls this approach its “next generation Sales Effect solution.” It seeks to provide an apples-to-apples comparison of ad performance across campaigns running on other networks measured by NCS, but relies on a very different methodology.
Closed environment required new approach. Spaulding told me that NCS needed to develop a new approach because Google won’t release data to third parties. So NCS had to run measurement inside the YouTube/Google environment. The process also had to be automated and dispense with human data analysts.
Accordingly, it uses a complex machine-learning approach that seeks to answer the question, “What would sales have been without exposure to the campaign?” Using lots of modeling, it provides accuracy “comparable to historical NCS lift results,” says Spaulding.
Spaulding also told me that NCS and Google validated the accuracy of the model before deploying it. He added that NCS will release a report later this year that will offer more explanation and detail. Spaulding believes that NCS’s next-gen solution is unique in the market and can be used in any publisher environment without significant modification.
Why should you care. Privacy rules are getting tighter and that inevitably means the exchange of data in the ecosystem will become more constrained. Marketers and brands will need to adapt. Cookies and other traditional tracking and attribution tactics are waning in their ubiquity. Device IDs are gradually replacing them but there are also some privacy concerns there.
NCS’s next-generation approach offers a new algorithmic model for sales lift measurement within first-party publisher data environments. It doesn’t require any transfer of data “outside.” So in that way, it’s inherently privacy compliant.
