January 25, 2023
Commentary
The Reports Of TV's Death Have Been Greatly Exaggerated
- by George Simpson , Featured Contributor, April 13, 2017
In the column, he basically says that FANG has data that marketers want and TV doesn't, and so TV will lose. Says he: "Brands have never wanted to buy CPMs (cost per thousand impressions) or GRPs (gross rating points); they want to sell stuff. The data-rich FANG and other tech giants are offering data that can be turned directly into sales."
I think even the TV folks will confess that buying just demo-based GRPs is fast becoming old-school thinking. But where has Shelly been the past five years, as each and every network has begun to append their minute-by-minute viewing data with third-party data that runs the gamut from Oracle Data Cloud (which tracks $3 trillion worth of purchase a year, including Visa) to Acxiom's 5,000 data elements from thousands of omnichannel sources to Nielsen Catalina Solutions' purchase data from more than 90 million households.
advertisement
advertisement
Then there is Nielsen Buyer Insights and Rentrak-TV Solutions Outcome data sources including credit cards, app downloads, first-party data, point-of-sales data, loyalty cards, CRM systems and digital conversions -- not to mention prebuilt syndicated audience segments from IRI, Mastercard, and Datalogix.
And what are the networks and other performance TV companies doing with all this data? They are showing brands that their commercials result in actual in-store and online sales -- or better "business outcomes," if you will.
Mr. P is right in that when all TV was measured in bulk, blunt metrics like age/gender demographics, it was very difficult for marketers to truly understand or optimize media performance across networks. But data, data science, and software have changed marketers’ ability to address these challenges.
Even the big agencies are reformulating how they target TV. Look at GroupM's Modi Media, IPG Mediabrands’ Cadreon, Dentsu Aegis’ Amplifi. Both Omnicom and Publicis Groupe have added TV data expertise. Even independent agencies like Horizon Media are building advanced TV centers (which, one could argue, is aimed more at OTT TV than linear TV, but I think illustrates the momentum toward data-driven targeting.)
Because not all data is equal and often one data set doesn't sync with the next one (or, frankly, some privacy protections), it is true that the TV biz is still a ways away from being able to report: "Ok, as a result of this flight of commercials, the brand achieved this volume of sales." But it is getting closer every day.
A better argument might be what happens if ratings continue to fall, and viewers stay off the big screen in favor of other devices and/or OTT programming. Will TV then have a scale problem that only FANG (or smarter configurations of data-driven targeting) can resolve?
I'd ask Shelly, but I don't think he has a clue.
5 comments about "The Reports Of TV's Death Have Been Greatly Exaggerated".
Read more comments >
Next story loading 
Good one, George. As I've been pointing out, the likely outcome if the new skinny bundles proliferate and gain critical mass will be a dramatic reduction in the number of channels available per TV home, with the broadcast TV networks, their stations and owned cable channels being the primary beneficiaries of greatly reduced competition from other cable channels. As a result of fewer channels being available, overall TV usage will decline, somewhat---say 10%---, but the way the remaining time is divided up will yield much higher ratings and more GRP's to sell, for the incumbent TV programming establishment .
Also, all of this talk about how silly broad age/sex targets are misses the point. When an advertiser's 25 brand national TV buy goes down, the buyer and seller must agree on a single metric to account for guaranteed audience tommange delivery. So they use adults aged 18-49 or women aged 25-54, but these are not the true targets of the individual brands and everyone---but the trade press----realizes that. Later, when the corporate buy is divvied up among the brands, an attempt to satisfy their individual targeting needs is attempted---but not as part of the buying process. I think that this is an area where improvements in the ways buying is evaluated can be made and have described a process that deals with this in "TV Dimensions 2017" for those who are really interested.
It's a semantic argument, death versus healthy. Yesterday an analyst at Sanford Bernstein said bleak future, but not death. Is that what you want to hear your family doctor say?
I have great respect for Shelly. But the data advertisers want is data they can trust. Those who trust data from those who grade their own homework, and that resides in silos, are doing what may look good, but it's not as good as it looks or they deserve.
Absolutes. Absolutes. Pthu Pthu. Changes by the second. It's complicated and the more variables added, the more complicated it becomes, not more targeted. Only when the closer it becomes to control what people are allowed to see, do and buy will targeting have the preciseness predicted.
Totally agree Jack.
Age/gender is a convenient brand cohort that is used for trading currency. But then so are many, if not most of these 'new generation' targets using first-party or third-party data. Some of these systems are fused or hybridised to the trading currency, while many remain 'stand alone'.
To the best of my knowledge (based on my Aussie contacts in the US and similar systems in Australia) none of them are independently audited, and few if any have MRC accreditation.
Further, when we were playing around with this sort of targeting back in the early 2000s using third-party single source data, we found that while some programmes did show improved indexed efficiencies, across a campaign the improvement was generally within the sampling standard error - that is, it may have helped some of the time but no definitive assurances could be given that it would help all the time.
What we found definitely WAS more efficient was the client giving us sufficient lead-time to execute a buy (8-12 weeks out) so that we could cherry-pick from the inventory as soon as it was released. We were better able to achieve the campaign weight, the campaign reach and at better CPMs. While this typically delivered double-digit gains, precision targeting was in the low to middle single digit gains.