I’ve used some version of the “It’s The Data, Stupid” line at least a half dozen times in the past year to make the point in columns or public speaking that audience data has become the real economy driving Madison Avenue’s audience-buying systems, and someday, most of its media-buying. But I’ve never used it as appropriately as I am in today’s RTBlog, because that line originated, you know, with some political media operatives. And while Targeted Victory’s Michael Beach may seem as far removed as anyone could be from the Clinton campaign’s James Carville, they do have a lot of things in common, especially how they use data to target and influence voters.
The reason for today’s “Golden Age” modifier, is because today’s story about how Targeted Victory is using data targeting Wall Street traders solidifies how creative and innovative audience targeters have gotten. And while political media operatives like Beach have long left no stone unturned in pursuit of better data science, I think conventional agencies and brands are moving in that direction, too -- especially in their use of first-party data, and in the ways they integrate it with other audience attributes to fine-tune and focus in on the best possible audience mix they can.
It is such a radical shift in mindsets from the classic ways Madison Avenue operatives have thought about targeting audiences -- via carefully constructed sample-based research like Nielsen’s, MRI, Scarborough and the like -- but if you ask me, it’s the absolute logical progression the industry has to move in, because there just is too much good audience attribute data out there to be ignored.
It will take a while for old-line strategists and classic media researchers to embrace it, if ever, but the change will come from sources willing to embrace a much more pragmatic approach: you know -- results. And there are likely few categories as immediately results-driven as political campaigns. Movie openings maybe. Retail sales events. Automotive pushes. And, of course, almost any performance-based product or category that cares more about conversions than it does about the statistical validity of projected audiences reached.
Apologies if I sound like a heretic, and don’t get me wrong. I’m not saying we should abandon conventional sample-based research. It’s still great for planning, for thinking and strategizing, and even for making some types of media buys, and for posting them. I’m just saying more of the media economy will be based on actual consumer data in the future -- because, well, brands are targeting actual consumers, not samples, in-tab or otherwise.
Right about now, I know some of my research intelligence friends are frothing, and I fully expect some of them to post comments below. I welcome the discussion. It’s good for the industry. But just because we’ve done things one way for a century or so doesn’t mean we’ll continue to do it the same way for the next century or so.
I saw just that sort of reaction the other day when we published a story about
AT&T AdWorks releasing new data via its Bueprint TV audience targeting system that will enable advertisers and agencies to buy TV shows based on their composition of mobile users, including their
devices, operating systems and data usage plans. Some of the reactions were, well, reactionary -- usual suspects grousing about the research validity of using data that way. The point here isn’t
that one approach necessarily replaces the other. It’s that brands and their agents need to use every source of data in their power to figure out how to break through to their most valuable
audiences in the most effective way. Call it research. Call it data. Call it whatever you want. When it’s done right, I call it something that works.
"Audience targeting" image from Shutterstock.
I have been in marketing research for 20 years. Your article is on point except that it doesn't go far enough. I shall froth at researchers who will froth at what you've write. Fielding sample based research is like driving a Model T with George Jetson. It's what we know, but that doesn't make it the best way. I'm posting a link to your article on my web page. Thank you.
Joe, I assume that I'm one of those usual suspects who is overly concerned about the validity of some of the "big data" systems, which purport to measure "audiences". If so, so be it. It's all well and good to champion systems that produce audience data from gigantic samples----on the grounds that this is a more accurate way to go than making planning or buying decisions based on projected audiences per traditional surveys with seemingly microscopic samples. And it's all very good to talk about interfacing third party marketing data with the "big data" audience findings. But how far does this get you when the audience data, itself----no matter how large the sample----provides an often misleading picture due to its basic reliance on set usage, not viewing metrics? Do we blithely skip over that little detail? And how many product categories are represented by third party marketing sources? Aside from automotive and many packaged goods, there isn't much third party data to meld in with big data audience findings. Also, when we talk about interfacing product use information from one source with household set usage from another source, you have to utilize some form of ascription----or simulation----- technique, which, in itself, is rife with "issues". You are no longer using pure single source data. Among the other questions, how does such a system handle the TV upfront, when there is only dated information on last season's shows and nothing about next year's programs? And what happens when the TV networks, syndicators and cable channels refuse to bow to buyer pressure and guarantee audience delivery only by broad demos as before? Finally, it is likely that TV programs with high proportions of youngish, affluent audiences will be the desired "media vehicles" of most advertisers who use the new "audience buying" systems. In other words, if a network has 30 primetime shows, most buyers will want only 10-15 of them----invariably the same holdover shows plus an occasional newbie that everyone is hot for. Once they see this, the networks will greatly increase their CPMs on the "desired" shows----which nullifies any edge the buyer gets by singling them out---and the sellers may also increase the amount of commercial clutter on such shows, to sell more GRPs at a premium---which also poses ad impact problems to consider. In other words it's not a given that "audience buying" will deliver anything like the promised dividends. While I'm one of the harshest critics of the current and woefully outmoded TV audience targeting system, until some of the basic issues inherent in the "big data/audience buying" concept, as it has been promoted so far, are addressed in an acceptable manner, it's merely an interesting notion that requires a good deal of testing to determine whether it is a viable alternative.
Joe, great article and timing. Buying better audiences is the future. Julia (pro) points out some of the research pushback. Ed (con) also let's us know nothing is perfect. My coffee appointment this morning reminded me about the question, I always loved, which was----"Do you want progress or do you want perfect?". I most often choose progress.
Skip, I'm not advocating "perfect" since I know from long experience that it is unattainable and one has to make compromises to get something done. What I'm reacting to is the deluge of hype about "big data", "audience buying", programmatic buying, etc. where everyone is jumping on the bandwagon without considering the inevitable trade-offs and issues. What surprises me is the continued silence of most users----planners, buyers and sellers---of the current and, I reiterate, totally outmoded targeting and buying systems. What do they have to say about all of this? Apparently nothing!S o all I'm doing is calling attention to the other side of the argument if nothing else, to give a few of the bandwagon jumpers something to think about. I'm all for progress, providing it's really an upgrade.
@Skip: Or don't let perfection be the enemy of good enough. What I like about data based audience targeting is its pragmatism - and its science. It is different kind of science than the historic sample-based research used by the ad industry, but no less valid when used appropriately. My sense is it is still most often used appropriately when measured against actual results, you know, real human behaviors like engaging with an ad, opting into something, or buying a product - not ratings estimates for audience exposure.
@Ed: Your POV is extremely well articulated, and I don't disagree with anything as it relates to how media - especially TV - has been planned, bought and sold in the past. I am mainly talking about how data is being used to target audiences in media where you can actually target users, and a little bit about the future of media that currently cannot - like most of television. That said, I do think the kind of consumer behavior data being applied by organizations like Targeted Victory or AT&T AdWorks is innovative when used appropriately for planning and indexing the composition value of media (like TV) that currently don't have survey-based data on the audience's brands are actually trying to reach. I believe that's the main way advertisers are using it now - to index the audience comp value of the show to their actual targets. I have no idea if deals are being negotiated on this basis, and I agree with you that it will be a long time, if ever, before it impacts things like the upfront. But there already are organizations putting their money where their mouths are - companies like Simulmedia - that are willing to hedge on the basis of actual consumer behavior data in order to make a market. Time will prove if they are right.
@Joe - While there are some very interesting opportunities coming up, there's one huge leap that I don't hear anyone talking about. Namely, the data we CAN gather has poorly defined connection with the ways we define audiences. No problem at demographics. But demographics are pretty meaningless. Yes, we can measure certain behaviors. But to draw conclusions about what those behaviors means introduces potentially huge error. (A friend's software company focused on people who were searching and reading a lot about their software. Turned out they weren't prospects - they were job hunters trying to learn to sound smart about that software system.) Look at it this way: Client/agency describes the target market as X/Y/Z. But the data that's available about media audiences only reveals statistics about 1/2/3. In the agency discussions I've been in, there's no acknowledgement of the need for a leap of faith to bridge that gap. Leading me to concur with someone I know who observed: With all the new audience tracking, companies will finally have all the data they need to completely destroy their brands.
I stand with Ed and Doug on these matters. Thank you both.
The industry could use a refresher course in logical fallacies.
For that matter, I stand with Plato, Socrates & Aristotle.
It's the "Golden Age" of data only because vested interests want it to be -- and don't know how to handle the data sandstorm they have created to obscure or destroy our vision and to take the paint of the (media) vehicles we rely on to communicate with consumers. Onwards & upwards.