January 25, 2023
It's About Time, Literally: Ad Industry Eyes Temporal Planning
- by Joe Mandese @mp_joemandese, December 1, 2015
In a move that signals a potential
shift in the way the ad industry thinks about its most fundamental principles of media planning -- reach and frequency -- a leading media planning software provider has created a way to calculate it
based on the amount of time consumers actually spend with media. The new tool, which will be introduced by Telmar early next year, isn’t intended to replace conventional reach and frequency
models that are gospel for many advertisers and agencies, but is designed to give them another way of looking at it: a temporal way.
The approach, which Telmar calls RFT (for reach-frequency-time), comes at a time when a number of industry stakeholders are beginning to rethink the value of the time consumers actually spend with media, and how it can be leveraged to assign more value to the media that deliver it.
advertisement
advertisement
Earlier this year, a number of premium publishers began adopting advertising models that charge advertisers based not just on “impressions” -- who and how many people are exposed to their ads -- but for how long they are. The Financial Times introduced its “cost per hour” (CPH) model to advertisers in May, offering it as a replacement for the cost-per-thousand (CPM) impressions models historically used by publishers.
In July, mobile video and rich media format developer Sled Mobile introduced a “cost-per-second” model enabling advertisers and agencies to “de-risk” their media budgets by paying only for the time consumers were exposed to their ads on mobile screens.
In November, The Economist introduced an “attention buy” format that charges advertisers based on how much attention readers pay to their digital ads.
While the ad industry has been organizing around the concept of a minimal viable exposure of time -- the Media Rating Council recommends one second for a banner ad and two seconds for a video ad to even be counted as an impression -- the shift to more explicit measures of time-based exposure has been slow to gather steam, in part, because there is little foundation in media planning theory for thinking about it that way. The original precepts of reach & frequency planning were based on static assumptions of media exposure and did not account for the amount of time individual people actually spend with them.
“This is a giant step forward for brands in the age of global planning,” asserts Paul Gold, vice president-strategic data operations at Telmar, who helped develop the new RFT analyzer. To illustrate the point, Gold says, “imagine if Coke found that its drinkers used their smartphones 20% less than Pepsi drinkers in a part of the world where no media planning data existed. That would have a huge impact on their ability to evaluate alternative media investments in the market.”
That’s exactly what Gold says Telmar found when it conducted a pilot study to serve as a “proof of concept” for the RFT analyzer. Not surprising, it picked a highly topical advertising category where there are clear lines of division among the time that consumer segments spend with various media: politics.
The study surveyed 3,299 Americans on the amount of time they spend with various media, as well as their party affiliations, and divided them into Democratic-leaning, Republican-leaning and Undecided voters based on the amount of time they spent with media.
The study looked at all the major media used by the respondents -- digital, print, radio, and TV -- but some of the most interesting analysis come from looking at RFT planning within a medium, utilizing one of the most fundamental forms of temporal media planning: the daypart.
On that basis, Telmar analyzed three different TV schedules of 200 gross rating points each applied to three daypart plans: daytime, prime-time and one dispersed throughout the day. The test showed markedly different reach and frequency curves for each voter segment. Daytime-only scheduled performed the worst. Prime-time schedules delivered the greatest reach among Democrats and undecideds. Dispersed schedules delivered the greatest reach among Republicans.
One of the more interesting -- and potentially topical -- parts of the Telmar study is that it also analyzed the consumer segments based on whether they installed and used digital ad-blocking software. On that basis, Telmar found that Hillary Clinton and Ben Carson would fare better than Bernie Sanders and Trump in terms of getting their online ads seen on user’s browsers, because those candidates had lower percentages of supporters who installed ad blockers.
Telmar President Corey Panno says the new RFT analyzer isn’t intended to be used as a form of media-buying “currency,” although it will be replicating the studies and offering the tool in markets where there is little or no audience data available, including Third World markets in Africa, Asia and South America.
But the real purpose of the analyzer is simply to give planners and ad executives another way to think about the role of consumer time spent with media plays in planning reach and frequency.
“We should always be challenging those original principles,” says Barry Lowenthal, president of The Media Kitchen, who hasn’t been briefed on the new Telmar product yet, but said he agreed with the idea of thinking beyond static reach and frequency. “They were the best that we had at those times, but what happens is that they sort of shift from principles and become laws. We shouldn’t forget that they’re not laws, they’re just the best way we had to do something at that time.”
Lowenthal said measuring time spent as a “first filter” is a “good start,” but he said the ultimate goal is to “figure out how we reach uniques.” By that, he means, unique media consumers within a medium or across media channels. One of the problems doing that, he said, is the data isn’t available for some big analogue media to do that.
“We spend a lot of money on TV and out-of-home and there’s not a way to do that,” he said.
“I think everyone is struggling with what have been relatively blunt evaluations tools that we been using to develop and analyze various media and the plans we put together,” adds Audrey Siegel, managing partner of sister MDC Partners agency Assembly.
Siegel, who also had not been briefed on the new Telmar product was a little skeptical about its potential value, adding: “What would make an actionable tool is if we could find something beyond a rating point and ge to a more meaningful understanding about what those impressions are, and how they play out with the user/listener/viewer, in terms of their actions. What we’re really trying to do is get to some kind of comprehensive measurement of all screens.”
8 comments about "It's About Time, Literally: Ad Industry Eyes Temporal Planning".
Read more comments >
Next story loading 
This reminds me of the days when we weighted primetime GRPs more heavily than daytime GRPs based on presumed impact/attention levels!
Lots of different ideas, Joe, but some running in different directions. In digital, it makes some sense to value more time spent on a page containing an ad as this suggests greater ad exposure. For decades, magazine publishers whose editorial output fell into the "reading book" category tried to sell buyers on the added value of their total reading time, relative to "picture books" or fast read weeklies---but this did not always translate into better ad awareness findings. In the case of TV many advertisers have tried to downplay the values of daytime TV and low rated cable channels, relative to primetime TV on the grounds of ad clutter differences which, they assumed, diminished ad exposure----yet ad recall studies often failed to support this assumption and certainly not in the face of the huge CPM differentials favoring the less desirable forms of TV. In fact, research shows that your average daytime TV viewer is considerably more ad receptive than your typical primetime viewer.
So I'm not sure exactly where this is heading. It may be a very good way to look at things----I'm not against that----but we need a little clarification. For example, doesn't time spent with the ad message---if we even have such measurements---trump time spent with the vehicle---be it a TV telecast, a website visit or a magazine reading experience? Or are we saying that a one hour TV show is automatically better than a half hour show for the advertiser ---because the one hour show's audience spends more time watching it?
While I agree with the concept that 'time' is the one common link or metric across and between media, such a model seems to be predicated on the assumption that the 'value' of time is equal across all media.
This is where I struggle. For example, a monthly magazine such as Vanity Fair may take 3 hours to read an issue. That is an average of around six minutes a day, whereas other media such as TV, radio and online can rack up 3 hours a day.
Also needing to be factored in is that the consumers 'portfolio' of content varies widely between media. For example, someone may read 2-3 printed papers in a week, maybe 4-5 magazines in a month, maybe 4-5 TV programmes in a day, and maybe a hundred or so webistes in a day.
When I was building such models in the late '90s I found that there were so many variables - of which time was my number one candidate after R&F - of which the majority were unmeasurable in a quantitative way that the model was both unwieldy and insufficient. In essence we had to add in so much subjectivity (which good strategists and researchers already do) that a purely quantitative model languished.
Having said that, I genuinely hope that this nut has finally been cracked!
As the original developer for RTF I have to agree with all that you say, but point out that RTF is only designed to address some of the issues you raise.
I entirely agree that time spent with the ad, or media vehicle, is extremely valuable and increasingly available for digital media. But it is expensive to collect and, as Barry Lowenthal comments, is not available for some big analogue channels.
Importantly, where it is available, it is usually only available in single media silos. That was one of the drivers for this development: to collect cross-media time spent data at a respondent level.
By developing a methodology to convert this data into R&F, we can provide invaluable assistance in the cross media channel planning/budget allocation process.
And, as a benefit, because the research is relatively limited in its scope, it is possible to undertake virtually everywhere – even where there is little or no conventional media research.
Of course, since we measure time spent at a medium level, this can be used to analyze the intensity of respondent’s media involvement which in itself can be used in the planning process.
All of which raises another set of questions. Say you know how many seconds a digital video commercial which lasts 30 seconds is displayed on a user's screen. For some users the amount of time may be 30 seconds---assuming that this means that the ad, istelf, is what is being "watched". For others, the amount of time may be anywhere from 1-29 seconds. Do these less than 100% "impressesions" count as equal to the full exposure ones in a reach and frequency calculation? Or do you discount some or all of these "exposures" in some way? And, if the latter option is the "solution" how do you express your reach and frequncy projections---by time spent level? That sounds like a very complicated and potentially misleading path to take. Far better, in my opinion, would be to count only those cases where the ad was shown in its entirety, and eliminate the others.
As for TV, radio and print media, there are no time spent measurements of commercial or ad exposure. Yes, I know that the electronic peoplemeter and PPMs appear to have this capability for TV and radio, however, in reality, they are not measuring commercial audience, but merely whether a set/receiver was tuned to a channel airing a commercial at a given moment in time. We could pretend that this doesn't matter and that all of the digital, TV and radio data is technically comparable----but we are probably just kiding ourselves.
Where does this lead us? Who knows----maybe nowhere.
You comment correctly that exposure to the ad is what is important and is basically not measured in most cases and where it is it will be in a non-comparable way across media.
The complexity and cost of doing this is why we did not attempt it. We have NOT looked at time spent with the ad but time spent with the medium in order to generate estimated R&F for a typical campaign at the planning stage. In this respect the technique is open to all the same failings that exist in traditional GRP based planning. However what we have achieved is a very efficient way to do cross media planning/budget allocation when full multi-media currency data is not available, and even where no currency data exists.
In simple terms the basis for the approach is the relative time spent by individual respondents both within a medium and across media. Utilising this information we can calculate probabilities for each respondent as to their likelihood of being exposed to a campaign of a given number of rating points
Dick, thanks for your reply, though I'm having trouble understanding how your model functions, as described. Setting that aside, as this is probably not the right forum for a technical discussion, have you attempted to validate your model using independent sources like MRI or Nielsen? Just curious.
This thread does indeed seem like it should go off-line for details which we are happy to discuss. The short answers are that our model can use any time-spent data from proprietary or subscribed surveys that can be cross-tabbed (i.e. for which respondent data is loaded—not summarized data). We have validated our methodology using data sources we are permitted to use. We will be releasing a white-paper shortly.