Thinking the Unthinkable about Video Ads

supposedly addressed Wanamaker’s question for television ads. It’s the foundation supporting a $70 billion annual business. It’s supposed to tell advertisers which programs viewers are watching and for how long. Given present computer technology it’s downright stunning when one learns Nielsen’s techniques for collecting its data.

With so much money on the table there’s a lot at stake. Some industry constituents want to update the measurement technology while others want no changes at all. Any changes are likely to upset somebody’s apple cart. That’s probably why progress in measurement methodology has moved at the pace of continental drift. Nonetheless, sponsors pay the bills and ultimately they’re going to demand more for their money.

Nielsen’s flagship service tries to estimate the viewing habits of our entire nation from a polling process based upon imperfect diaries of 18,500 homes. Given the sample size, three-fourths of the 400 cable and broadcast networks are simply not watched enough to be accurately measured. Except for the top 20 markets, Nielsen mostly relies upon paper-logs that panel members (i.e. participating consumers) maintain on the honor system. Even in larger markets where Nielsen provides electronic log-entry, the company acknowledges that users sometimes press the wrong key resulting in a measurement error of up to eight-percentage-points.

Owing to the shortcomings of Nielsen data, a number of vendors sell supplemental information. TiVo is one example. With several million DVRs deployed across the nation they provide more granular and real-time data. For example, TiVo can tell how long a viewer watches a commercial.

Similarly, the CATV industry hopes to provide far more detailed and accurate data via its Project Canoe initiative. Unfortunately, Project Canoe faces a significant technical challenge. Data formats are inconsistent among different CATV systems. Some set-top boxes are incapable of capturing data and passing it back to the head-end. Each Multiple System Operator (e.g. Comcast, Time-Warner, etc) has proprietary elements in its networks. Proprietary inconsistencies are amplified to the nth degree considering that each MSO is by definition an amalgamation of independent systems typically acquired over decades.

In sum, the problem of getting universally accurate measurements out of the legacy CATV and Satellite platforms is going to be as convoluted as the Gordian Knot. Ultimately, the solution could well be to simply cut the knot by putting all video on the Internet. Standards on the Net are decades old, uniform, and well understood by an abundance of developers all over the planet. Measurements can be in real-time and sliced & diced nearly infinitely.

Most significantly, Internet sponsors are increasingly demanding Cost-Per-Action (CPA) advertising. Google AdWords conditioned them to expect that it is only necessary to pay when a visitor clicks on ads. Once video resides on the Net, sponsors may insist that they only pay for video ads that get watched.  The new paradigm will nearly eliminate the utility of audience measurement statistics as we have known them. We’ll want different data, but it will be more readily attained on the Net where the granularity and accuracy of measurement is infinitely better.

To advertising executives who don’t want to be accountable for the performance of their ads, such a paradigm shift is so horrible as to be unthinkable. So they may choose to simply avoid thinking about it. To them, it’s a creation of Satan anyway, more destructive to their business as the ability of the iPhone to display its Internet Videos on a television screen. Yet, when sponsors pay for video ads only when they actually get watched, the ghost of John Wannamaker will break into a happy grin.

show less...