So it sounds as if the TV ad business is on a slippery, data-driven, audience-based digital slope headed to an inglorious commoditized future like online display, with falling CPMs, where content-driven adjacency and brands have largely been kicked to the curb, doesn’t it?
No, TV’s not going to get kicked to the curb. But let’s face it: TV advertising is already commoditized. The majority of TV ads today are traded on the basis of sex/age demographics and gross rating points, allocated and rotated by network, day and daypart, and little else. They are not bought and sold according to the specific spot or show. The brand of the show only matters if it appears on an exclusion list.
It’s getting worse. As audience fragmentation worsens, fewer and fewer shows are able to break through and stand out with large audiences on their own. Already, more than two-thirds of all TV ad impressions are on episodes with national ratings under 0.5 -- a metric that’s quickly moving up to 75% of all ad impressions. So it's getting more and more difficult for most individual shows and episodes to stand out and be valued uniquely.
Data can only make it better. No one can argue that the best way to value a TV ad spot is solely on the basis of broad sex/age and day/day-part segmentation. More data about people viewing that episode, from previous purchase behaviors or actual sales caused by ads viewed from that episode, can only add value. More data means more markets for every spot. In a world where there is robust data on viewers and their behaviors, the notion of “non-endemic” advertising will go away. Every spot will have certain “endemic” advertisers who may find a special connection to a show and its audience, and many other advertisers who will find lots of counterintuitive spots to buy as they match a show’s deep viewer and performance data with their own data about target customers.
Finally, it wasn’t data that commoditized the online display market anyway. It was the way the online industry counted and sold advertising. Online ad impressions are everywhere, in numbers increasing faster than the amount of time people spend engaged with digital media and services. There may be 15 impressions on a single page. Many of them may not be viewable. Fundamentally, there are more online display impressions than there is demand for ads, and the ratio is getting worse.
TV, by contrast, sells on a time basis. TV’s ad impression load grows or shrinks with the overall amount of audience time spent on ad-supported TV shows, with a few exceptions. Today, largely because of careful management of ad loads, pricing, packaging, and commitments, demand for TV spots exceeds the supply.
Thus, data-driven, digital approaches to advertising won’t be TV’s demise. Instead, it has already met its nemesis by making sex/age GRPs its primary currency. It is already commoditized. Data could save it.
What do you think?
Great point Paula. I don't believe that audience data and computer-driven decision-making will be the only way TV advertising will be bought and sold in the future. Human judgment and content adjacencies will still be a critical part of it
Dave, I agree with many things you say here and in other posts, however I should point out that a large part of TV is already de-commoditized. I refer to most of the broadcast TV networks' primetime fare, some late night network and primetime cable programs, most sports, specials in general, most news shows, plus certain cable channels in their entirety. In these cases the CPMs and demographics are subordinated to the imagery, content compatibility and promotability of being seen in such contexts to the point where this far transcends any evaluation of cost efficient targeting. Indeed, were programmatic ever to come to national TV, it would almost certainly shun such fare and advertisers who prefer these types of buys----there are many-----would simply legislate them into their plans. So, de-commoditizing is already here and has been here for decades. Whether a greater reliance on data will break down these long established media preferences remains to be seen....but, somehow, I doubt it. The question then arises, what about other advertisers---mainly packaged goods types ---- who go more by the numbers and are CPM conscious. Here, I wonder whether using more sophisticated data to aid in targeting, evaluation of sales response, etc. will create the huge benefits that are envisioned as most of the programs that such advertisers buy are "commodity" shows---daytime talkers, game shows, off-network reruns, oft repeated/low budget documentaries, cable "reality shows", old movies, etc., which tend to be priced very cheaply----and competitively to eachother--- compared to the genres I mentioned at the outset. I'm not arguing, just making an observation based on how things actually are. There are, in fact, two kinds of advertisers and two kinds of program content----high CPM premium and low CPM eyeball tonnage collectors. Which of these is going to be most affected by the coming "data" revolution?
Ed, lots of good stuff there. Thanks. I think that the eyeball tonnage folks are likely to move faster than premium CPM folks, once they realize that they can drive more provable, profitable sales by adjusting their plan. However, the fastest movers will probably be folks who aren't in Tv heavily today - like e-commerce or mobile app developers - who can really benefit from TV's power, but need to see data and dashboards like they have in digital before putting money into a media channel. It will be interesting to see.
Dave, thanks for raising another important and timely point.
It might help us to widen the lense on what constitutes ROI across linear and IP-based video trading. In traditional (linear) TV planning, implicit to any ad buy was also brand awareness. We now have an impressive ability to use data for audience-based targeting, measurement and link to purchase. But we must also be cognizant of the limitations.
Tying ad exposure to sales provides great insight to marketers with short sales windows (e.g. CPG) and where purchases are made online. We now have datapoints where previously only direct response could quantify the impact of advertising. But where short-term or direct sales cannot be proven, attribution models discount the value of brand awareness – or where messaging didn’t work – as well as other touchpoints that may have influenced the consumer’s journey to purchase. Of course, this doesn’t even consider fraud and bot traffic that only perpetuates the 50% wasted ad dollar issue.
To be sure, data will benefit the overall media planning and measurement process, especially as media fragments. Together with a shift to audience-based buying, these move needle by informing WHO is viewing WHAT, (if and) WHY an ad should be placed WHERE – and WHEN. But as the media consumption options multiply, so do siloes of data and standards for metrics. We are closer than ever to having the tools and data to provide fair attribution of each touchpoint along the purchase funnel. Wouldn’t Wanamaker be proud if we worked cooperatively as an industry toward this aim?
I’d like to pose 2 further questions to the group as we continue the discussion:
1. Assuming data and automation will provide efficiencies for (human) media planners, how can we re-examine pricing of media and representative data to drive differentiation and bridge the gap between CPM models and traditional TV trading?
2. How long will marketers trust metrics provided by the same companies trying to sell them media*?
*In full transparency, I work for a WPP-owned company, but the Kantar division maintains best-in-class, market/media-neutral and auditable currency services.
Excellent points Susan. My take on your questions is that: 1) I expect hybrid pricing, probably based on a target CPM but guaranteed to readjust to a performance metric and, 2) some proprietary methods of ad measurement (like Google search clicks) maye survive as a currency if they are as self-evident to buyers, but the vast majority of ROI media will have third party validation.