Commentary
TubeMogul Research Tracks Evolution Of The GRP
- by Tobi Elkin , Staff Writer @tobielkin, November 22, 2016
TubeMogul has conducted some interesting research on the changing nature of GRPs (gross rating points).
The findings are rather surprising:
--Advertisers need to buy 37% more prime-time ads to get the same reach on TV as they did four years ago.
--Frequency has increased. The average prime-time viewer saw a given ad 13.9 times in 2016, compared to 12.9 times in 2013, the three-year period the research took into account.
--Prime-time TV CPMs are also on the rise, with broadcast up to $47 in May of this year, compared to $42 in May 2015.
TubeMogul’s analysis finds a loss of 7.3 million households across prime-time buys in four years. This might help explain why many marketers are turning to software to reduce oversaturation and reach viewers that TV buys missed. How are they doing this? By taking a data-driven approach to TV buying (programmatic TV) and building cross-screen plans that include digital, social, and other media in an effort to have more consistent frequency across audiences quintiles.
TubeMogul’s analysis was conducted in the last three months. The data from comScore TV Essentials Exact Commercial Ratings analyzes the period from 2013-2016.
“The data tells a clear story,” Taylor Schreiner, TubeMogul’s VP of research, told RTBlog via email. Schreiner said “TV buyers face a daunting challenge” in the years ahead. “Marketers have to run [many] more prime-time ad spots just to get the same audience reach on TV as four years ago. Simultaneously, as frequency rises and people see the same ads over and over, marketers risk oversaturation.”
This dynamic may motivate many advertisers to look for software-based solutions to build audience reach and manage frequency across screens. Schreiner said that by using software, marketers can “flatten” the reach curve without demanding more from agencies or buying more ads manually. The reach and frequency issue continues to be just as relevant as it ever has been--and with increasingly granular targeting, it's only going to become more complex.
Tobi Elkin is the Editor of MediaPost's Real-Time Daily and the RTBlog. You can reach Tobi at telkin@mediapost.com.
Tobi, I assume that you and TubeMogul are referring to the broadcast TV networks, not to primetime as a whole. Also, the CPM cited is not for the broadcast networks' total average minute audience in prime, including all viewers, but for some demographic like adults 18-49, let's say. If one counts cable, which has majority of the primetime audience, not the broadcast networks, the 18-49 CPM would drop to something like $36-37. As for the reach factor, it is true that TV reach curves need revising and that it takes more GRPs, now to attain previous reach levels---we include a report on this in our upcoming annual, "TV Dimensions 2017"--- but the "finding" that it takes 37% more GRPs to attain the same reach as four years ago, needs a much better definition. Are we talking about the broadcast networks alone, or does this still apply if an advertiser uses a mix of broadcast networks and cable---if so I doubt that the finding is correct. Also, are we referrencing daily, weekly or monthly reach? It makes a real difference.
Ed, I believe TubeMogul is referring to the bcast networks but that's a good question. I think it requires clarification.
Ed:
First, re: the CPMs, you are absolutely correct. The numbers I was mentioning were actually yours from your press release here (http://www.mediadynamicsinc.com/media-matters/july-15-2016/) and refer to your figures for A18-49 broadcast. That should have been better communicated.
As for the 37% figure, I think it is clearer in the actual output soon to be up on our site. We assumed a 1000 GRP buy across each year, assuming an advertiser bought primetime broadcast and top 25 cable, and used whatever mix was optimal. That mix has obviously shifted to cable and to a wider diversity of shows which is part of the reason for the increase in spots required.
The biggest takeaway for me in this is that TV is getting harder to efficiently "by hand." It remains the premier medium, and if anything, the number of ways for a marketer to use TV has increased. But the need for thoughtful, transparent, analytical and (dare I say) programmatic use of the medium is clearer every day.
Thanks, Taylor, for weighing in on this issue. The clarification is very helpful to our readers and me.
Thanks, Taylor, for that clarification. As you probably know, I'm not exactly a big fan of programmatic time buying for national TV---as currently constituted--- though I see it as a distinct possibility for national spot buys which are, essentially, GRP-driven add-ons to national TV cmapaigns in various markets. I think that we should remember that most of today's national TV buys---broadcast networks, cable and syndication, whether in the upfront---about 70% of the dollars---or the quarterly scatter market---about 30% of the dollars ---are made on a "corporate" or a consolidated multiple brand basis---often involving 10, 15, 20, 25, or more brands. In such cases the agency-of record ( AOR ) that handles the placements--whether they be for primetime, late fringe, daytime or some other categories, is dealing with a large number of GRPs and the buys are spread out among many networks, cable channels and syndicators. So mass reach---for the corporation, as a whole---is a given. the trick, later, is how the GRP tonnage that has been purchased is allocated to the various brands by the corporate media director and his/her minions. Some of the large agencies have developed so-called "allocation" models which take into account various factors---reach, demography, CPS, etc and try to determine the best fir for each brand within the confines of the corporate buy. I have always felt that here was an area where a reasonably priced independent computer model---perhaps a "programmatic" one---could be of value to advertisers but this requires the use of formulas, not tallies of actual data as we are talking about the future and there is no data to process. I wonder if your company is exploring this aspect with any advertisers.