Around 40 broadcast and cable network executives reportedly attended a meeting hosted by the Video Advertising Bureau where they heard from analytics company Data Plus Math about its system for linking multiscreen video ad exposure with actual product purchases (which has been dubbed “Project Thor”).
I haven’t seen any specifics on what they are doing, but I am wary whenever sellers get together to tell advertisers the “best” way to measure the effectiveness of their product. Advertisers and media agencies would obviously need to get on board for this to have any chance of seeing the light of day.
A few things need to be considered when using any type of new data for buying:
-- Most important is whether the data is projectible. The main reason networks will guarantee adults 18-49 — and not adults 21-30 — is because the larger age group is more stable and easier to project. Product purchase data, particularly when linked to network viewing (as reported by current syndicate research companies), tends to fluctuate widely from year to year.
-- Can you measure commercial effectiveness on an ongoing basis? I don’t remember what commercials I saw in a given program I watched yesterday. Unless it is a new commercial, appearing for the first time, even if I do recall a commercial and it affected my behavior, you have no way of knowing whether I saw it in a certain show.
-- How do you know which exposure spurred action? If I see a commercial and then go out and buy the product, while you see the same commercial and do nothing, it could simply be because it’s the eighth time I’ve seen it and only the first time you’ve seen it. Will the system be able to measure how many times I’ve seen a particular ad before making a purchase?
-- Ad creativity is a big factor in recall and action. Will one advertiser be charged more only because they produced a more effective ad?
The real issue goes beyond just whether or not standard age/sex data is anachronistic. The real issue should be whether samples can still measure the viewing habits of the overall population in the first place.
If I’m part of Nielsen’s national people meter sample, for example, I theoretically represent thousands of my cohorts (those who fit into a dozen or so categories that Nielsen has determined impact viewing behavior).
Twenty or thirty years ago, when people had fewer than 30 channels, there was no time-shifting via DVRs, and no streaming services such as Netflix, Hulu, Amazon Prime, CBS All Access, YouTube TV, etc. This worked quite well. Today, I have serious doubts. It’s interesting that Nielsen has announced it will begin measuring Netflix viewership, while Netflix has already said the initial viewing data released by Nielsen isn’t even close to the real numbers.
I think the problem is best illustrated by looking at common television references. Every generation has had iconic phrases from various TV shows that most people (across age groups) were familiar with. Say “There’s a gremlin on the wing,” “Meathead,” “What you talkin’ about Willis?,” or “I’m the master of my domain” to viewers who watched TV in the 1960s, ‘70s, ‘80s, and ‘90s, respectively, and they know what you mean (even if they didn’t actually watch “The Twilight Zone,” “All in the Family,” “Diff’rent Strokes,” or “Seinfeld”). Today, if you say “We’re all Negan” or “You know nothing, Jon Snow,” you’re more likely to get blank stares. The last two examples are well-known phrases from the two highest-rated series on television, “The Walking Dead” and “Game of Thrones.”
The TV landscape is so fragmented today that the highest-rated series on television are watched by a fraction of the people who watched the highest-rated shows 20 or 30 years ago. With hundreds of channels, DVRs, on-demand, and multiple streaming services, I and my cohorts no longer watch the same things (and don’t always even have access to the same programming). How can samples still determine who represents the total population of TV viewers?
The idea that the industry is trying to measure anything related to ad effectiveness when we are still unable (or unwilling) to measure actual commercial viewing is kind of ridiculous. Maybe the industry would be best served refocusing on that issue.