Commentary
Programmatic Measurement: Confirming The Outsize Value Of TV
- by Walt Horstman , December 4, 2014
The term “measurement” used here is the digital definition, meaning measuring the outcomes of the TV campaign, such as driving sales lift, search or social media activity. Historically, TV has not been held to precise outcomes measurement, only needing to demonstrate that the correct number of GRPs was delivered, because GRPs equaled business results.
I remember when I was a product manager quite a while ago, we developed a highly accurate linear regression model to correlate TV GRPs by week in market with the number of new customer acquisitions. If we were not tracking to reach our target number of new customers by the end of the quarter, we knew how much to add to the TV budget to hit our number. It was a brute-force approach, but it worked for that era.
advertisement
advertisement
Today agencies and advertisers want to precisely measure outcomes for all types of TV campaigns, from major brands across all categories to traditional response campaigns. They want to know not only if the campaign is successful, but why is it successful, and which parts are more successful than others. Which networks drive better outcomes, which dayparts, which target audience? Demand for this type of precision measurement is growing because it has been table-stakes for so long in digital. Marketers now have the same expectations from their programmatic TV campaigns.
I’ve spoken with a number of data company execs who have measured various outcomes of TV campaigns. They say it’s quite common for sales lift or search traffic lift to be well into the double digits. Measurement is simply confirming what many have known from years of experience: TV advertising is extraordinarily effective at driving intended behavior. Measurement is an opportunity for TV to reinforce and re-prove its value.
The fourth quarter brought a great deal of consternation in the TV advertising ecosystem, with a lower-than-expected upfront followed by a sluggish scatter market. Meanwhile digital, social and mobile seem to be charging ahead and stealing greater share. Unilever has committed 20% of its budget to digital, and Pizza Hut has allocated 40%. Further, TV network executives have remarked that advertisers are deciding on their TV spending closer and closer to airtime, sometimes within a few days. This lack of visibility on demand has taken a toll on media stocks. All of these changes speak to a shift in the advertising marketplace that demands more flexibility, agility and accountability, especially for TV.
Programmatic TV, the union of automation and audience data, including measurement, delivers on this need for flexibility, agility and accountability. The outsize value exists inherently in TV advertising, and programmatic will help unlock it further.
Major broadcasters have exclusively relied on closed distribution environments to generate revenue, i.e. upfronts, retran fees based on bundles and Nielsen polling, hence how scale audience aggregation is achieved. An open environment where digital and programmatic lives fostering audience fragmentation, is therefore highly probable to decrease the value in traditional TV advertising. Adapting programmatic TV appears to be counter-intuitive.
The ability to make media spend decisions closer to broadcast and adjust campaigns very quickly based on data driven insights will be the key driver in adoption of programmatic techniques in the broadcast marketplace. This will accelerate as more granular data become available due to the growing prevalence of digital broadcasting and the data collection it will enable.
All of these are interesting points, but I still wonder where the data is coming from that will motivate advertisers to walk away from the upfront, or long term program sponsorships, and invest their TV dollars on a purely opportunistic, day to day or week to week basis. Moreover, a major part of TV ad spending is not driven by precise sales expectations tied in to precise variations in media weight. I'm talking about advertisers who are using various TV content platforms because of their "environment" and other intangible "benefits". Even for the P&Gs, however, are we envisioning a world where the brand manager tracks the sales results of each set of TV audience "impressions" and raises or lowers the required ad tonnage, or adjusts the targeting, accordingly, day after day and week after week. What happens if the brand determines that it can up its sales by boosting its ad weight next week, only to discover that the networks can't accommodate this request as they are mostly sold out? And how are such determinations made in the first place? Is Nielsen going to be replaced with a new "big data" measurement service that tells each advertiser the exact response for every ad exposure more or less as it happens----assuming that such correlations are really valid----or possible? Where is that service? How will it get its data? Who pays for it? How many advertising categories can it cover? I'd really like to see some answers, even if they are theoretical, that tell me how this is expected to work. It seems to me that even if a miracle occurs and such a infinitely granular, real time audience/sales result measurement were somehow created and TV time sellers could accept to the very risky idea of never knowing what their sell out rate is from one week to another, that a totally new breed of automated brand manager and media planner would be needed to use such a system. Leave it to normal human beings and you'll never come close to seeing its full potential attained.
The narrative and questions becomes quite complicated when one applies linear thinking from a linear broadcast model.....