Foxes, Henhouses, and Homework The State of Measurement Part 2
Anytime a heated debate over media measurement occurs (oooh, aren’t those fun) you are almost guaranteed to hear at least one of the favored idioms. “We can’t let the fox in the henhouse” or, “we can’t let them grade their own homework”. These statements are applied in reference to the measurement of a media properties’ audience, and by extension the audience exposed to various advertisements. The sentiment behind this is in order for there to be a fair counting and representation of what a buyer (marketer) receives, it is not right to allow the seller (media owner) to do the counting and representation, that’s grading their own homework.
Folks in the media measurement space have used this logic for as long as I can remember, but perhaps it is time to reconsider the notion. If we were to go back 25 or 50 years, the idea that it makes more sense for one independent, neutral party to count and define audiences than media owners is understandable. There was one dominant form of media at the time, television. It was not so difficult to apply a single methodology and process to the collection and interpretation of data across all major media owners. Of course a big media buyer like Kellogg’s would not want to rely on CBS, NBC, and ABC to independently and accurately tell them how many women 25-54 saw their ads. Thus, the third party standard (in this case the Nielsen GRP) was established and deemed not a fox that would threaten the hens (buyer’s investments).
But my, how things have changed in the past 25 years. Media choices seem endless - television, digital display, digital video, CTV, mobile browser, mobile app, and on and on. All of a sudden, the idea that a single methodology and process could accurately collect all the data necessary to produce reliable audience measurement seems impossible. As I noted in my last post - The State of Measurement Today - audience measurement fundamentally requires two things - accurate counting of people, and accurate assignment of descriptors about those people (demographics, behaviors, etc). Let’s examine each of those and really critique the “no grading your own homework” mindset.
First and foremost we have to consider the accurate counting of people. If we just look at three major formats of media - television, mobile video, and streaming/CTV - in order for a single company to actually count the number of people exposed to content or ads they need to:
Collect linear TV viewing data either through an opt-in panel or aggregation of set-top box data
Deploy tags to all media properties’ web pages as well as advertisers’ ads themselves
Deploy an SDK to any and all major mobile apps where video will be played
Deploy an SDK to all major streaming apps, and/or passively collect viewing data through things like audio content recognition (ACR) and smart TV device manufacturer data
In order for one entity to successfully do the above is hard, very hard. Just getting your SDK included in a significant number of apps alone is a major challenge and fraught with issues. These challenges are particularly hard when you consider that one might watch NBCU content via a ROKU TV or an Amazon Fire Stick, two very interested parties in the value of the viewership data.
On the other hand, each major media owner (think NBCU) or distribution platform (think Roku) is in a fine position to count the number of people exposed to their content. After all, they are the one distributing and/or serving the content. Of course I still want an independent evaluation of their ability to count, and this is exactly the role of the MRC. It’s kind of like, I am ok with you grading your own homework so long as a teacher looks over your shoulder on a regular basis to ensure you do so correctly.
But if you are still not comfortable with the idea of media owners providing their own count of people, consider this - pretty much every major advertiser has already demonstrated they are ok with it! Who do you think provides the Reach/Frequency metrics for your Facebook/Instagram media buys? Whether it is reported through a third party or not, those audience figures are coming from the platform themselves, and we seem to have just come to accept that, so maybe it’s not so bad?
The other aspect to audience measurement is the accurate assignment of descriptors about the people. Advertisers need to have confidence these descriptors are accurately applied, and increasingly rely on more than just age and gender when buying an audience. If you are going to achieve this, you cannot just know a little bit about a few people, that’s the legacy model of small panels producing age/gender descriptors. You need to know a lot about a whole lot of people. Well here is where our focus rests at Truthset. By partnering with the world’s leading consumer data companies, collecting massive amounts of consumer data, and evaluating it all for accuracy, we have produced an asset which delivers panel level accuracy at census scale. Today, we are matching the people, counted by media owners, to the accurate descriptions of who they are - age, gender, ethnicity, household income, children in home, pet ownership, etc. Our descriptor data provides a standard which allows buyers and sellers to transition from guessing about to knowing the audience they have reached.
I am willing to say, maybe it’s finally ok to let the fox into the henhouse, he just needs a set of checks and balances. Fortunately those checks and balances are more in place today than ever before. The MRC is there to validate correct counting ability, verification services like DoubleVerify and IAS verify it is actually people (with an opportunity to view) that are being counted, and Truthset is verifying the description of those people are indeed accurate. Now, get back to your homework and let me know how you scored.