FEATURE5 December 2011
FEATURE5 December 2011
Chief research officer Bruce Goerlich responds to reader questions about the company’s new measure of commercial viewing.
The company described the Exact Commercial Ratings Module as an “advance” on the existing currency measure, produced by Nielsen, of average viewer numbers for commercial minutes within a programme.
But readers queried how “exact” the ratings could be, given the company’s reliance on set-top box data rather than household meters. We put their questions to Rentrak’s chief research officer, Bruce Goerlich.
How does Rentrak calculate the viewership for an individual commercial?
Rentrak receives from Kantar Media Services – a leader in providing advertisers and agencies with information on commercial occurrences – a detailed occurrence list of national commercials by network, by programme. Those occurrences are merged with our television data set to produce exact commercial ratings.
Given the use of set-top box data, is it more accurate to say that the company can report the total number of tuned TV sets, rather than the total number of people viewing a specific commercial?
We report on TV set viewing, household viewing, and households “with” viewing – what we call “advanced targeting”, e.g. households with children, households with income above $125,000, households that just bought a Ford, households that are politically conservative, etc. We do not report an individual in front of the set.
What data sources do you rely on to achieve this?
We have different sources for our advanced targeting, ranging from Polk automotive purchasing files to Epsilon household demographic tags. Depending on the data source, we employ models at the zip code level, data mapping and direct household matching via a privacy-protected process through third parties.
Is it correct to use the term ‘exact’ to describe these ratings? Isn’t ‘estimated’ commercial ratings more accurate?
We feel ‘exact’ is appropriate. The ratings reflect the viewership during the exact time period the commercial was aired, as reported by Kantar.
A predictive network model has the ability to predict object interest based on video footage, writes… https://t.co/HFYrIBTjoN