Commentary
Programmatic TV: Doesn't Work Perfectly, But Close Enough
- by Wayne Friedman , Staff Writer, June 10, 2016
One-to-one marketing is still too hard for TV -- at the moment, anyway. But programmatic TV deals can work -- generally well.
Speaking at the MediaPost TV Summit, Oscar Garza, global director of programmatic of Essence, says, with regard to true one-to-one TV marketing: “It’s easier on digital, but not on linear TV. Addressable will improve on that.”
In response to drilling down to specific consumers, Manny Hernandez, global director of programmatic of Ogilvy, says: “I struggle with how granular we get with data. You can cripple your buy.”
That said, Hernandez says, with regard to programmatic TV: “it does work, but doesn’t work perfectly. [For example] maybe you can’t get a frequency cap across an entire campaign. But you can get close enough.”
Wayne Friedman is West Coast Editor of MediaPost. You can reach Wayne at wayne@mediapost.com.
It doesn't have to be "perfect." There is no such state in the real world we mortals inhabit. But in general, if you get to 80 percent or so accuracy, that is enough to show results worth the investment of money, time and people. At 80 percent accuracy, you are hitting the mark four times out of five. And when doing any analysis, planning or buying in this industry, wouldn't we accept that as great performance? Getting it right 80 percent of the time is much more efficient than trying to get to ninety percent or so. I have personally seen over the years that while achieving 80 - 85 percent of the goal out of 100 consumes x time and effort, diminishing returns set in beyond that, so getting from 80 to 90 costs much more than x that got you to 80. Often not worth the extra effort, till something happens from outside that alters the curve by dramatically improving productivity. And that is infrequent, unpredictable. So don't wait for perfect. Very good is close enough.
Henry, I doubt that "programmatic TV", which has yet to be applied in any buys involving many TV options as opposed to those where only a single seller and a few low rated shows are involved, can be evaluated in terms of its degree of perfection. As for the targeting aspect, so far, neither the proposed programmatic or the so-called "addressable TV" approach can deal with viewing as opposed to set usage. Which raises the question, are these "new" targeting schemes really better than using single source data from MRI, which defines both viewing and product usage/buying from the same respondents and applying the resulting indices to Nielsen viewer ratings?