Commentary
The Gross Rating Point Begot The Audience Rating Point -- And The Father Is Digital
- by John Piccone , August 1, 2014
The GRP has ruled the TV measurement roost for quite some time. And there is a good chance it will crusade to be the ruler of the digital landscape, too. Its focus on measuring how many people watch a TV show a certain number of times allows planners, buyers and sellers to make decisions quickly and efficiently. The simple and now-classic formula – Reach (R) x Frequency (F) = GRP – has kept TV big and round.
The digital landscape grew up in the shadow of the GRP and saw the opportunity to separate itself by looking at and controlling for frequency against the people reached. That magic little cookie allowed the machinery to put a stop to drowning people in repetitious ads, and more importantly, allowed media sellers to move past the generic impression to focus on the person. This allowed planners to control for frequency and buyers to guarantee how many people they could reach.
Thanks to the increase in the number of channels and growing diversity of programming in the past two decades, audience fragmentation has come to TV in a big way. Although more people are watching TV today, when and how much they watch is all over the map. As a result, more ads are reaching fewer people. In formulaic terms, the R in R x F = GRP has become very small, and the F has become bloated in an attempt to maintain GRPs in the face of declining ratings per show.
Unlike digital, where the cookie allows for frequency caps to be implemented, viewers are subjected to multiple views of the same ad, frustrating consumers and advertisers alike. Until the cookie comes to TV and frequency can be capped, it is up to the business community to address the problem of too much frequency to too few people.
Enter the Audience Rating Point (ARP). The ARP takes the best of TV and digital measurement concepts and smushes the two together to create a new formula: R=ARP. In case you missed it, the F is gone. Now it would be folly to think that Frequency does not have an important role in brand building. But at the same time we know for sure that R does equal results. And since the TV technology can’t control for frequency yet, let’s at least have the business discussion about eliminating it from the pricing formula.
Firstly, it would maximize the true value of the only media platform that is available in virtually every household by reaching audiences in the outer edges of programming.
Secondly, it would allow planners and buyers to plan and buy on the same currency: audiences and audiences only.
Thirdly, sellers would have the option to sell against impressions (F) or audiences (R) and thereby focus the conversation on the marketing tactic of reach, resonance or reaction.
Lastly, and most importantly, we the consumers will not be bombarded by the same ad over and over again, thus improving the overall watching experience.
At the end of the day, the market doesn’t need advanced technology to transact against metrics that are already available in today’s toolsets.
This seems pretty simplistic and a little naive about the ability to manage to a certain extent, reach and frequency on television. Your equation reminds me of King Arthur's logic in Monty Python and the Holy Grail. "If she weighs the same as a duck, she's made of wood, and therefore a witch."
Frequency is alive and well. It will continue to remain as vital in marketing theory as reach. And your view that there is little we can to to control frequency in television is flat out wrong.
Television reach and frequency can be controlled through daypart mix and program ratings parameters. For example 200 TRPs of 100% prime on programs with ratings over a 5, will deliver a lot of reach, but very little frequency. Conversely, 200 TRPs only during morning, daytime, and early fringe with ratings below a 1 are going to give you a lot of frequency, but much lower reach.
Granted, you cannot control TV ratings as much as you can digital, but frequency in digital is still extremely important. If it wasn't, retargeting would not be near as popular...or effective.
I am more of a recency advocate than a reach/frequency theorist. Recentness suggests that the marketing message you hear right before you make a purchase decision is the most important. However even recency requires frequency to be effective.
Direct response TV relies heavily on frequency, and very little if any on reach or total rating points. It only takes frequency to generate conversions. That's why you see Anthony Sullivan hawking Oxyclean 10-15 times between 2AM and 5AM. The DRTV formula hasn't changed since we learned that Slim Whitman has outsold Elvis or the Beatles.
So via condios to your absurd notion.
TV is measured on households with PPM in only 26 markets. Therefore, all "audience" measures are fuzzy to begin with. Moreover, since we cannot buy TV on a program by program basis but must make deals with networks and dayparts our ability to implement to reach among audiences is compromised in execution. Finally, our ability to implement dynamically is non-existent. Other than, it is a great idea.
Agreed with Evan Brown. I'm not sure how much TV buying you've done, but with a properly optimized daypart mix, you can effectively control projected frequency.
The "bombardment" (frequency) problem you mention is either due to advertisers who are dumping money into TV because of highly competitive environments (insurance, telcom) or because they have failed to balance their budget and network dispersion. You comment about the inability to manage frequency in TV, but how many times do you see the same ad repeatedly in FEPs or VOD? Same problem, different format.
Second, "R" does not *necessarily* mean results. If a consumer sees an ad once, it's not guaranteed to elicit a marketing driven response. That's why Effective Frequency and Recency Period are critical to establishing communication levels.
This is true for any format: TV, Digital, Print, whatever. I highly encourage people to study up on Erwin Ephron's work to understand such basics.
I have to agree with Evan, the logic seems tortured here and somehow we arrived at "she's a witch" in two easy steps. I could go on about the math but replacing a flawed, very flawed, measurement system called a GRP with something that is more measurable and completely eschews frequency (show me any digital campaign that moved any needle at an average frequency of 1 or less) seems very retrograde, however, the real problem we need to change our thinking about measure to take into account the full impact of the channel and how people connect to them, not just if they have an opportunity to see it and how many opportunities they might see it. The impact of an :30 TV spot ≠ a banner ad ≠ a radio spot ≠ magazine ad ≠ social media placement ≠ newspaper ad ≠ pre roll video et ceterta et cetera et cetera.
Our striving for a new measurement system needs to embrace that fact.
Where does one begin? First, regardless of what media scheduling theory an advertiser advocates, it has never been possible to make TV audiences watch your commercials at whatever frequency rate is desired. People watch what they want to watch and only when they feel so inclined. Moreover,the audience surveys are unable to tell you whether anyone is watching a commercial, let alone being influenced by it. So you may think that you are "reaching" 70% of your target audience an average of 4 times per month ( 280 GRPs ) but, in reality, only 60% of the target group is probably watching----in the sense that they are paying any attention--- about 2-3 times. Second,, it is not true that TV ads are reaching fewer and fewer people due to rating fragmentation. If you are an advertiser who is willing to pay the CPM increases required to attain the same number of GRPs as before, you can split up your buy among many channels and attain almost as much reach as in the good old days---only it may take you five weeks, not four to get there. Finally, all of this talk about "audience buying" being so different than the GRP method is rather misleading. When an advertiser buys a given number of GRPs he is merely stating an audience tonnage goal in the form of percentage points and delivery of same is usually guaranteed by the seller, give or take a point or two. When an online buy goes down, the buyer can specify a certain amount of audience tonnage---like 2.5 million impressions---- and this is what is delivered. What's the difference? What's more, when a TV network or cable channel guarantees a certain number of GRPs, the commercials actually appear on the viewers' screens---in their entirety. The same can hardly be said of many online media buys where 30-50% of the time the ad isn't even "delivered".
One last observation. TV GRPs are not measurements. Rather they are the sum of actual ratings ( audiences ) attained per minute by various TV telecasts that carried an advertiser's commercial. If you want to challenge the validity of the rating surveys, perhaps on the ground that they produce vastly inflated results, that's one thing. So far, there is no reason to believe that the ratings are grossly inaccurate, though there are issues that should be dealt with in an ideal world. However, once you accept the ratings as more or less accurate, GRP compilations or goals are nothing more than a way to express audience tonnage in a more readily understandable manner. That's all.
I agree with Ed to a certain extent. However the inability to measure actual viewing (reading or listening included) and engagement is consistently unmeasurable across all media, digital or otherwise. How manny banners go unnoticed (even though we pay for "impressions")? How many clicks are "accidental?" Magazine readership estimates from MRI or Simmons are considered the currency for that medium, but they only measure readership of the publication...not of the ad within the publication. The TAB is trying to more accuately measure OOH ratings with "eyes-on" numbers, but so far there are still many advertisers who use DEC as the measurement; and that only measures how many cars drive by a billboard. I also agree with his position that TRPs/GRPs are not measurements. The real measure in the success of a marketing effort is the individual set of KPIs set for that effort. We project the quantity of impressions, GRPs, pageviews, DEC, or whatever needed to deliver our KPIs. Finally, probably the most honest measurement is that of newspaper, whose currency is "circulation." This is probably more accurately measured than any other medium, and the definition, "circulation" makes no promise whatsoever of readership...merely that "your ad is printed in a publication that will go out to x number of households....nothing more...nothing less." Maybe all these newer media should take a lesson from one of the oldest media.
Evan, I agree that frequency can be managed on TV, but disagree that it is a simple as buying large rated shows. First, because there aren't many 5+ rated shows to buy anymore. Today, according to Nielsen data, 65% of all TV viewing occurs on shows with ratings under 0.5. And second, as Erwin Ephron told us years ago, buying lots of smaller rated shows actually delivers a better ration of reach to frequency, since higher rated shows in prime have much higher percentages of heavy TV viewers and thus deliver lots of frequency against those same viewers.
You're right, David, but this forum (and most other forums) doesn't allow for that granular of a discussion as to how to manage R/F. There are duplication issues, unique audience issues, etc. I was using those two approaches solely as 30,000-ft. examples. It is a very complex level of planning that those siloed into the digital world do not even want to understand. Even as we talk about convergent and integrated media planning impacting, as Charles Pinkerton suggested, the value and purpose of the impressions in different media, there are still those who refuse to step out of the digital vacuum into the real world. We need more discussion on the appropriate unit of measure that takes into consideration issues like active vs. passive viewing, message complexity, impact of pictures and sound vs. just pictures or just sound, etc. Can we get there? I don't know. It may be as fruitless as trying to come up with the Grand Unification Theory in physics.
Agreed Evan. I'm hopeful that we'll get there before the Grand Unification Theory :-)
Regarding the ability of controlling one's frequency by buying time in top rated primetime TV shows, it should be remembered that an average advertiser garners only 10-15% of its total GRPs---or ad "impressions" ----via the broadcast networks' primetime fare. The rest is attained with primetime cable buys as well as dollars spent in other dayparts----all offering much lower ratings. Moreover, with average minute primetime ratings now hovering in the 2-3% range for the broadcast networks for typical target group definitions, one can question whether it is really possible to control viewer frequencies at such low coverage levels. Even if it's deemed worthwhile to give it a try and use primetime broadcst exclusively in an effort to manipulate frequency levels---say by "roadblocking"----running spots on three or more networks at the same time----you must consider the consequences.To do this, you will have to give up all of your cost efficient "other daypart" buys and even primetime cable may have to go. But at what price? Your cost per viewer will probably double, compared to last year and you will get many, many fewer ad exposures. Which raises the question: Is ordering your frequency so some people see your message only once in a week really worth it when you are sacrificing so much audience tonnage over the course of your campaign?
And just to clarify, the formula is AVERAGE Frequency = GRPs / Nett Reach (i.e. 1+ reach). This is subtly but importantly different for two reasons. First, it is the Frequency that is being derived (i.e. the 'output' variable). Second, it is an estimate of the AVERAGE frequency - or the weight - of the campaign. It has no cognisance of the frequency distribution - merely the average. Most communications plans are framed in terms of how many people (the reach) how often (the frequency) given the marketing goals and the creative execution that is to be used. GRPs is simply a shorthand way (an 'input') into achieving the campaign goals (the 'outputs' of net reach and average frequency) as cohorts for sales and financial success.
This is part of my last post on my blog. I'm writing about online GRPs for more then a year. So please tell me am I right.
First, behind GRP there are also measures like net reach, OTC, frequency. Media-planers use all of those measures to optimize TV campaigns to reach specific target groups. GRP is just an end product, one number which will tell CMO did they reach their advertising goal in any country because 250 GRPs in one country is the same (or close enough) to 250 GRPs in any other country (percentage of population).
Second, as 30” is “standard”/by default for TV ads and GRP was invented by TV guys, time is not in E.Q. GRP formula. If it would have been in formula than GRP would have been measurement of how much time population or person have spent on average with an ad/brand on TV.
Third, that time spent with an ad/brand on TV generate branding results for brands, like brand awareness, purchase intent… So GRP in the end is promise of some branding results. Results are not always the same – better (more creative) ads have better results, but everybody gets something, on average there are always results for brands.
Fourth, because of those results and experience (number of GRPs/branding results/sale of product) which CMOs have, they can establish (and they have established) number of GRPs needed to get to branding or sales results in some country. So talk between CMO and his agency account is something like this: CMO asks his account “how many GRP did we get in that country?”, account answers “around 200″, and CMO says “that is fine, we have maintained our brand awareness on a same level, we can expect same results in sales like last month”.
So if we, online advertising professionals, want to get to the point where TV is with branding advertisers, we have to learn from them. We have to add time to a viewable impression, we have to research what time and which formats drive and for how much branding results on average and then we can start development of online GRP and present our findings to the CMOs.
There is many more things we have to do beside this to form online GRP but this should be a good start.
Jan, I think that I understand what you are saying. If not, please correct me.The basic point is that over many years and many ad campaigns, branding advertisers have learned how to manage their audience delivery with TV, mainly, and, to a lesser extent with radio and magazines. Of course, reach times frequency, which equals GRPs , does not automatically tell you what the result will be in terms of overall ad message awareness and certainly not whether you will improve your share of market. The deciding factor is what you have to say in your ads and how effectively you grab the audiences' attention and motivate them. Still, there are certain rules of thumb----let's call them guidelines--- about what levels of exposure and frequency are required to make the sale as well as other issues, such as wearout----the point where you no longer gain added buyers no matter how many GRPs you employ. Online sellers could learn a lot from the experience of branding advertisers if they bothered to explore how "old media" like "linear TV" is planned, bought and evaluated. This obviously applies to the question of ad viewability. It's all well and good for some industry committee to set up a standard that defines an online ad as "viewable" if at least half of it can be seen for one or two seconds. That's a nice compromise that keeps the ad sellers happy but don't think for a moment that such a definition is going to impress branding advertisers who know that you must engage the viewer with a commercial before you can sell him. A few seconds of partial visibility isn't going to get the job done.
Exactly! Engagement is part where creative people have to do their part of a job. On a media side, we have to provide them with opportunity to do so. And I'm not talking about 2'', more like 10'' minimum. Also format is important, as I know EQ GRP is 70% of screen, there is also lean back VS lean forward "mode", targeting, frequency. But we have to start somewhere, and I think that time spent with an ad is a good place to for that.
Loving the discussion. I'd like to add that "not all GRPs are created equal". Put another way, one minute with one medium is not necessarily worth one minute on another. For example, the average person may read something like one weekly magazine for 45 minutes and 3 monthly magazines for 90 minutes each. That is, 7.5 hours a month or around 15 minutes a day compared to (say) three hours a day of TV. That doesn't mean that 10+ times "better" than magazines. So while time is the one common metric not all time is equal. I also echo Ed's comments around GRPs - they are rules of thumbs for the CMO, CEO, CFO etc. For example, I had a client here (years ago) who 'budgetted' for 3,000 TV GRPs per annum (around 60 per week). It was up to us to work out how to 'flight' those GRPs most efficiently and effectively. We'd "burst" with new creative or 'news' or trade support, and then have weeks off. The CMO 'knew' from previous experience that 3,000 was his "magic number" (and it turned out he was pretty much on the money as our econometric modelling later showed). But always remember that a GRP for one medium is not necessarily equal to a GRP for another medium. Further, while I agree that an ad "can't work" in 2 seconds, that is conditional agreement. Read the work of Robert Heath on low-involvement processing. Using the old Ehrenberg principles of a frequency of three for 'effective' advertising, that is necessary for new products or new ads. Roughly put, it takes three exposures to 'imprint' on the brain. But once the consumer has taken their message from the ad (and all consumer's out-takes are different) it doesn't take long for the human brain to recall all the brand associations (which are often sounds, colours, logos, emotions associated with the brand rather than the brand itself). When you consider that pundits say that the average person receives 3,000-4,000 ad messages a day, if you relied on 30-seconds for each message ... then that is more than the number of minutes in a day. Even if they are just 1-second we're still talking around 1 hour a day cognitively processing advertising ... which we clearly are NOT doing. Most of it is sub-conscious - apart from those few precious first exposures. The out-take is that once you have your ad hooks imprinted with consumers you only need regular (not necessarily frequent) low amounts of reminder exposures to keep your campaign bubbling along (until those creative hooks eventually wear out).
I agree, John, this is turning into an interesting discussion. One of the distinctions that should be pointed out is between an individual "exposure" and the campaign in its totality. When we talk about GRPs, we are referring to the aggregate number of exposures garnered against a target group over a particular time frame. This may be for a week, a month or a much longer interval. If you have, say, 100 GRPs per week, this probably involves upwards of 50 TV 'hits", each reaching anywhere from .5% to 4.0% of the target group. The same thing applies to online campaigns, only with smaller numbers per exposure. So, when we are discussing metrics like time spent for online ads, as a surrogate for real world exposure, we are talking about an average exposure, not how the campaign builds up awareness and gets its point across over time. John is correct that it takes about three thorough exposures---some place the figure a bit higher----for a TV commercial to fully register its message, but how one applies this fact in media planning depends on the campaign. Is it the beginning of a totally new campaign or is it the second year of an ongoing campaign? Also, is the sales pitch a competitive one ( "Brand is better than Brand B " ) or is it primarily focused on the advertised brand and its supposed attributes. In the former case---a competitive pitch -----even if the brand gets its message across after the initial three exposures, every time "Brand B" runs its ads, they function as a form of rebuttal, which means that Brand A, may need to do much more reinforcing----more GRPs, more weeks--- to sustain the impact of its effort. This may not be the case for a non-competitive pitch or for one featuring a totally new product where much of the message explains the way it functions and the benefits without any competitor to worry about.
@John, yes and no. I heard a lot of times that adex should follow media consumption and I don't agree, better to say that I only partially agree. Why? Because I might spend 2 hours with magazine and during that time I might be in contact with an ad for 10’’ on the other hand I might spend 1 hour watching TV and I will be in contact with and ad for maybe 60’’. I wrote a post about that “Time should be common denominator for all types of advertisement” on my blog Jilek-on.com. So the problem for establishing time through all media formats is that you can’t measure time spent with an ad in all formats e.g. in print. Another thing is that every media format has different ad format, only sound, only picture, sound and picture, different sizes. Third thing is that every media format has different “exclusivity” of an ad, when you are watching TV than ad is only thing there – 100% audio and video in print you can have article on one page and ad on the other. There is also differences in way how why people use different media formats – relaxing, searching for information, communication and that also have influence on how they will perceive a message… So no because time spent with media is not the same as time spent with an ad and yes, because every media format have different abilities and different magnitude for transferring an ad message even if time spent with an ad is the same. Based on that I don’t believe in GRP for print (better to say EQ GRP for print). So there are two GRPs, one for measuring reach of media vehicle and one for measuring ad reach in that media vehicle. You can’t add one to another because we know every media format have different branding results even if time spent with an ad is the same e.g. branding results (on average) for radio ad 30’’ are not the same as TV ad 30’’. But if you would know that (on average) you have to get 10 times more on radio to be on a same level as one TV ad that might be a good information. Especially for internet where we can measure every exposure and find how much time spent and by which frequency we need for which (an average) result. And yes time is limited, and that is beauty and challenge of our business, and that is something which is pushing us to be more efficient and more creative, and because of that ads (with time incorporated) will have price. And also it will get more interesting when you will ask for a campaign with reach of e.g. 10 mil people and with 30’’ ad frequency capping 1 and after that retargeting of those people with 5’’ version of a same ad. But to get to that point first we have to build infrastructure for that on internet and that means that we have to ditch page views as “container” for ads and reload ads without new page view – you will have one page view and only content will be changed (and ads every xy‘’).
@Ed yes, for every industry/product “TV strategy” – advertising strategy can be different, but they know something and they have idea what could bring results, from one of my posts: “In addition, TV measurement systems measure reach on higher frequencies and a rch 2+, 3+ or exact 2 or exact 5 can be calculated but cannot be ordered and paid for – this is delivered via effective media planning (agency). GRP has been in use for decades and cases as well as scientific research have established some known models for TV planning. For example, that a level of 200 GRP’s weekly over reach 1+ on 70% is sufficient for retaining brand awareness. Telco industry experience, on the other hand knows that they need a frequency of 10+ over 50% of the target population to introduce a new service/product – this implies a certain level of GRP’s needed, depending on the country and TV viewing patterns. The premises of these and other similar theories have their base in the fact that TV is picture, sound and motion vehicle and if something interesting for a viewer is presented the viewer will shift his/hers full attention to the TV screen.“ So we need to make a lot of research on how different format exposures and different time spent on internet influences users. And we can easily compare exposed group to a control group and see the differences.