Commentary
Ch-Ch-Ch Changes: Tomorrow's TV
- by Charlene Weisler , Op-Ed Contributor, December 16, 2014
See a short video of some of the highlights of
TVOT here.
For those of us in the traditional TV sector, changes can be divided into “positive” and unsettlingly “disruptive.” Here is how they parse out for me:
Positive Changes
Measurement Getting More Focused and Transparent
With all of the big and small data sets available and the advanced technology to drive analytics, we are now able to get to the core of consumer tracking behavior instead of relying on age
and gender proxies. CIMM’s Jane Clarke believes that there are two big positive trends in measurement: profiling and access. "One big trend is to profile customers based on the data and
linking datasets across platforms such as linking purchase data with media use data. The media data revolution started in digital and is now moving into TV,” she said. “The second trend is
the proliferation of ways to access data. Planning and buying are evaluating effectiveness across platforms. This is driven by the desire to move away from linear rating points, away from surrogates
of age and gender, to target consumers."
Now We Can Measure Sales Funnel Instead of Demographics
Along those lines, we can now attribute actual sales results to specific
media campaigns. This frees us to valuate audiences based on actual purchases. Dunnhumby’s Lung Huang explained, “It is all about big data. We are ultimately doing matching exposure to
purchase based on a verified household or person. We are taking the guesswork out of it. When Ted Turner first started CNN and it wasn’t rated, he sold ginsu knives. He didn't care about
ratings. He wanted to see how many knives he sold. It is freeing today not to be tied to legacy. There are many different audiences and we can show you that they saw the ad and bought [the
product.]”
Disruptive Changes
advertisement
advertisement
The Long View Is Much More Important
We tend to have a short attention span in the TV business. Facebook’s
Patrick Harris noted, “We tend to overestimate in short term and underestimate in the long term. Where is the content where you are living three years from now? The creative bar has changed. It
is now adding value when it used to be getting attention.”
Is Programmatic Destined to Include TV?
Can we expect programmatic to expand to television any time soon? Visible World’s Seth Haberman says this has already happened. But Beth Rockwood of Discovery believes that “the television business is different in that the inventory is generally tight and therefore enjoys very high CPMs, especially in broadcast.” Meanwhile Dave Morgan of Simulmedia sees both sides, and noted, “If programmatic TV means that a large portion of TV advertising will soon be bought and sold on a data-driven, audience-denomiated basis on metrics other than sex/age demographics, the answer is yes. That day is coming fast. However, if programmatic TV means that TV ads are about to be bought and sold machine-to-machine on a dynamic, real-time auction basis like online display: no. That kind of programmatic TV is years and years away."
Technology Brings Its Own Timing Challenges
Mike Willner of Penthera noted the technological challenge of streaming ad-supported video content: “When you add advertising to downloaded content, it becomes more complicated because of the ad flights. Christmas ads can be streamed on Dec. 24, but viewed on Dec. 26. So we would need to download the next flight of ads. We have technology that helps to do this, we can ID the user and what content they are viewing, and we can coordinate with the advertiser. But we can only upload new ads in a streaming environment, not when [video] is downloaded for future use.”
Despite the dizzying spiral of change, it is possible to carve out a path to assured future television success in the evolving media ecosystem. The secret is to stay on top of the trends and take calculated long-term risks.
Thanks for your optimism. A gloomier future would be what happened to Kodak, a company that truly got way out in front of a disruptive change, did all the right things, stayed on top of the trends, and still went out of business. Mobile devices are likely to create the same havoc for broadcasters as for real estate agents (Zillow), travel agents (Priceline), cab drivers (Uber), and hotels (Airbnb).
You could be right. I am hoping we are talking all of these test cases into consideration in making our business decisions. If we can evolve into content first and platform second, we might be able to tilt the business model in a way that still provides for a viable future for television companies. Of course the potential revenue might shrink and with it some jobs. But there will also be many more media outlets, albeit smaller.
The basic idea behind the application of "big data" to TV time buying is the assumption that all you need to know is who is buying you now and what are they watching. This means that current product and/or brand users are the ones you should target---which is not always the case. Take a typical car advertiser. The last time I ran through such data, current brand owners were, by far, the most likely to return to the same brand for their next car purchase. Doesn't that suggest that with such high loyalty factors, it might behoove a car brand to seek out other mindset segments---like people entering the market for the first time or potential brand switchers or upgraders who now have another make? Also, there are a lot of advertising categories where "big data" isn't available or where the buy is based on factors other than matching viewing to product usage---merchandising, corporate image, program environment, etc. As for the idea of matching independent "big data" information with audience surveys, this concept needs to be validated. There are many "technical" issues to be faced and resolved. So far, I haven't seen any movement along these lines. Instead, we are getting lots of speculation and theories----all of which may prove out, but these assumptions need to be tested....don't they? As far as the notion that computers will replace humans in TV buying and selling, I agree with Dave. That's not likely. A far more probable outcome will be the more sophisticated application of automation and computerized communication, to support the buying and selling process and give both sides greater flexibility, choice of options, etc.
With target segments, I think advertisers often do take into consideration potential switchers and those open to messages. And I am not sure how many categories are totally without some form of data even if it is custom study or qualitative. All of that can go into the mix for some form of targeting, I also agree with Dave and expect TV to go more programmatic in that area of buying and selling much sooner than later.
Charlene, the agency creatives and client marketing people certainly take into account many factors in fashioning their brand positioning strategies and their commercials, however, almost 90% of TV time is bought using absurdly broad 18-49 or 25-54 "demographics", and this takes nothing into account. So, once again, we have to distinguish between what agency functions we are talking about. If it's TV time buying, then one questions whether superimposing "big data" findings of detergent usage or cake mix purchases upon Nielsen ratings to produce detergent user or cake mix buyer ratings is a major innovation. The problem is that the actual targeting improvement across TV show genres and individual programs using such metrics are not nearly as large as is supposed. This is easily demonstrated by looking at MRI's findings. Sure, you will find some surprises and a few cases where demos, alone, don't go far enough, but most of the time, merely marrying product user---or even heavy user-----data with TV viewing data, as opposed to using finer demo breaks---like age within income----does not yield the great leap forward that is anticipated.
Hi Ed, I think we will slowly get away from age gender proxies and as there is more big data, especially shopping data, added to the mix for certain product categories we will be able to more easily map the consumer path from exposure to purchase or some other call to action. The challenge is to scale this. But that is being worked on now as the technology improves. MRI is valuable but maybe not so much in this instance since it is recall. But maybe there will be a way to incorporate some of the MRI data into the process - maybe attitudinal etc?