Given the press coverage and hallway conversations at Ad:Tech and the iMedia Summits, behavioral targeting clearly has come of age and perhaps may, in fact, be the "next big thing" in Internet advertising.
It seems every week that established companies offering behavioral targeting sign major new clients, and lots of other companies are nibbling around the edges announcing some version of their own behavioral targeting. Increasingly, agency RFPs ask for behavioral targeting as case studies show that it works for publishers and advertisers.
Behavioral targeting, in case you have been lost on a deserted island for the past year, is combining known demographic data provided by users (generally through registration, but also from online subscriptions, contests, etc.) with navigation -- or behavior, if you will -- to target ads. The demographic data yields something about a site visitor's gender, age, ZIP code -- and if the publisher is lucky, household income. The navigation behavior tells us about their interests. For example if someone clicks on car classifieds, or repeatedly reads stories like new car reviews in the automotive section of a site or responds to ad units for new cars, it is a safe bet he is an active, in-market car buyer.
Everyone who meets similar criteria is put into an audience segment and can be served an ad (no matter where they are on a publisher's site). Behavioral marketing provider TACODA Systems' studies have shown that users respond to these ads at a better rate than the current gold standard, contextually-placed ads, and up to 2,500 percent better than ROS banners.
Behavioral targeting gives online publishers the information they need to sell ads to media buyers who are addicted to age, gender, zip and content environment as the basis for buying ads in traditional offline media. No more "hits" or "pages served" metrics that they little understand and trust even less.
But don't neglect offline data
But behavioral targeting is but one tool in an arsenal of information that publishers could use to help sell ads. Since Interep Interactive's parent company is one of the world's largest radio rep firms, we see a good deal of radio audience data and have begun to sell packages that include both online inventory and radio spots. They happen to be two media where user consumption syncs up particularly well.
Let's look at how offline and online user data can work together.
Let's pick an audience segment that represents lots of change with lots of new purchases attendant to that change: expecting parents or those who have just had babies. It's not just a diaper maker and baby shampoo marketer's dream segment. It also is of interest to car makers -- time to get rid of that two-seater Porsche and get the easy-access mini van with the built-in child safety seat -- or real estate broker -- time to get out of that one bedroom apartment for a home with a yard and a dog.
From Mediamark Research's radio audience data I can find out household members who are pregnant, number of children currently living at home, and households which have used disposable diapers/underpants (and if they were Huggies or Luvs) in the last six months -- according to what kind of radio format they listen to most.
From Scarborough Research I can find out about lifestyle changes (birth of a child) in the past year -- and tie it to specific markets and ranked order of radio station preference. Although the data are updated only twice a year, they have a history of being accurate, and are trusted by media buyers.
Behavioral data, by contrast, is new and at times misunderstood by media buyers who may not have worked with it before.
The offline data sets give me a very good market-by-market view of where new families are occurring in significant numbers. What if I overlay that information with behavioral data that can also be analyzed on a market-by-market basis? Now, I have trusted syndicated offline data and very fresh behavioral data from people visiting the parenting sections or reading stories about new advances in pre- and post-natal care etc. The radio data gives me a forecast of the size of the market and the behavioral segmentation gives me a shot to show they are there and online.
Are we mixing apples with oranges? The radio data tracks radio listeners -- the behavioral data tracks online users. While this could make a case for an on-air/online package, it really gives us, as media sellers, a more thorough roadmap of the potential market. By combining old-media data with new-media data, it might make a more convincing proposal or simply offer a way to benchmark the performance of one medium or the other.
The future of advertising is in audience data. Let's think of creative ways to use all of it to our advantage.
Adam Guild is president of Interep Interactive.
To illustrate the hit or miss nature of behavioral targeting as commonly practiced, consider the often-cited example of the Car Shopper, whom we'll define as one who has visited content about new car makes and models at a certain level of recency and frequency. (We'll look at user retargeting another time, but the same principles apply.)
Car Shopper browses several pages looking at new four-door intermediate- and full-size domestic sedans priced under $30,000 on a site that participates in a behavioral targeting network. A data tag on the pages browsed allows the site or a network to drop a bit of code on her browser so that Car Shopper can be identified later as a candidate to receive an appropriate ad for a mid-level four-door sedan.
Car Shopper browses more pages outside the network and eventually lands on a site that has given the network permission to deliver targeted ads when possible. Unfortunately, when she lands on a page that has an available, targetable ad position, the network has in its database ads for convertibles, imports, SUVs, trucks and luxury sedans, but none that match her specific segment.
What happens next? Because this is an inexact science, the network might serve a luxury sedan ad that is out of Car Shopper's price range or an ad for a different body type. More often the network serves an untargeted default ad, a low-priced run of network (RON) ad or public service announcement (PSA).
Another common result is that the network passes the ad call back to the site, which then delivers its own default ad, usually a low-priced RON, house ad or PSA, or another untargeted ad that was sold by the publisher and given a low priority in its ad server.
The point is that instead of a valuable, well-targeted ad being served to an interested user, both the ad inventory and a potential customer go to waste. Adding insult to injury, the publisher usually has to pay for an additional ad delivery for the untargeted default ad.
Multiply this by millions of such lost opportunities every day across dozens of segments and hundreds of sites and you get a portrait of disappointment: a buyer fumbles to explain why an underdelivering campaign has to be cancelled; fingers point and voices rise between an unhappy seller and his "told-you-so" colleagues in operations when the publisher asks why she has to eat the cost of failed expectations.
What if the advertiser controls the targeting across sites using its own ad server? The result may be less waste for the advertiser, but for the publisher it means smaller volume, greater difficulty and more expense to absorb.
BT today: hit or miss
Paradise found: targeting without waste
Requirement 1: an open marketplace
Requirement 2: an open technology platform
Requirement 3: an open mind
No one ever said this would be easy, but there is a better way. In a perfect world, advertisers have access to an extraordinary volume of targetable inventory. Any network or publisher making ad space available for these high-value ads has access to an equally extraordinary number of advertisers, campaigns and ad creative units. In this world of abundance, advertisers almost always hit their target, and networks and publishers almost always deliver a targeted ad without defaults.
This is Eden, so advertisers have access to inventory they can cherry-pick all day long, serving ads only when they have precisely the right ad for the right targetable browser.
There are fewer bright red cherries than dull or overripe ones, so campaigns end up smaller and prices somewhat higher. But because the whole orchard is so much larger, even the lower campaign volumes may exceed what's available today with single networks running individual farms. And if pricing is auction-based, there's no behavioral premium, either. You bid higher when you know that a user is worth more to you as a Car Shopper, a customer, or other segment you seek, but you only pay a bit more than the next highest bidder, who may not recognize the same value you do.
There's no waste, no defaults to kill off publishers' inventory that can be used for something better, and both advertiser performance and publisher effective CPMs meet or beat expectations.
Where in the vastness of the World Wide Web does this parallel universe of endless inventory and advertising exist? How can the process be enabled without any single party controlling all of it?
We need a playground big enough to accommodate anyone with inventory, campaigns or technology for a marketplace where publishers, advertisers, networks and technology providers can all mix it up and compete fairly and in the open.
When you select a network for your behaviorally targeted campaign, by definition you limit opportunities to reach a target population to that network's base of identifiable visitors on that network's group of participating sites, using that network's technology exclusively.
An open marketplace, by contrast, is one where no single provider sits in the middle of every transaction. Instead, many networks, publishers, advertisers and technology providers all link together in order to give each one equal access to the combined pool.
The largest networks today provide reach and targeting, but only within each one's closed universe, and limited to its selected advertisers and publishers. They straddle and take a piece of every transaction, not revealing true pricing or identifying the parties on either side.
Networks and technology providers offer their own supply (inventory) and demand (advertising), their own targeting technologies (behavioral, contextual, demographic, et cetera), sales relationships, and so on. They must balance campaign requirements with the difficulties inherent in forecasting availability of targetable segment populations. In other words, they can't pick just the bright red cherries; campaigns would be too small to justify the effort and expense.
This is why large networks such as MSN offer a limited number of targetable segments which they use to extend reach to consumers of their most popular content areas. Neither Google nor MSN offers ready access to their platforms by third-party technology providers. Yahoo! -- a customer of and investor in Right Media -- is currently exploring opportunities to target advertising to their users outside its network.
Compare the approach taken by the evolving media exchanges. On the Right Media Exchange, (the company I work for) more than 60 networks and a half-dozen different behavioral targeting providers all compete openly across three to four billion impressions per day. AdECN announced it intends to offer an exchange in the United States pending success of tests in China, and AdBrite offers an "online advertising marketplace" with about 500-million impressions per day.
Access to massive amounts of inventory from thousands of participants is essential. What if you have ad space but you are limited to only one choice for targeting? The provider may offer a strong solution for targeting based on prior site visits, but not for retargeting against customer data.
Open APIs (application programming interface) make it possible for any provider of behavioral targeting technology to hook into the platform to offer its own flavor to member networks, advertisers and publishers. Instead of being limited to one network's inventory, segmentology and technology, an exchange advertiser can choose among an array of competing networks and technology partners, greatly increasing access to targetable inventory as well as the best targeting tool for the job at hand.
APIs are a hallmark of the open marketplace in which all parties can bring their best to bear on creating success on a level playing field. If a technology provider has a better way, it benefits all participants if they plug it in and offer it to all.
When you enter a marketplace, you see all the vendor stands and know who's offering what. You enter with an idea of what you want to buy, how much you're willing to pay, and how and from whom you're most comfortable buying. You go in armed with data, ready to bargain, and the more you know the better able you are to strike the best deal possible.
You know you're in a true community-based online media exchange when you can identify your counterparts, select the ones you'll do business with, and know that they'll play by the same rules as you. All participants start with equivalent information and on equal footing.
Just as many marketers have begun to embrace consumer-led conversations about their brands, so networks, advertisers and publishers are learning that there's much to gain from opening up their supply and demand to a huge community. Add a dynamic pricing mechanism and now you have access to inventory, multiple technologies, and prices set by market supply and demand, not by third parties that hide their inventory sources and pricing formulas.
Advertisers expect to be able to incorporate behavioral targeting in one form or another in their media plans. The only thing holding BT back from realizing its full potential is a new target range: one that provides access to extraordinary scale on a common platform with selectable partners and transparent pricing.
Can your creative team really scale for programmatic?
If we're talking about direct response creatives, the answer is probably, "Yes." Direct response advertisers have had a long-running relationship with digital targeting technologies. And for the most part, creative has kept pace. But the demands direct response places on the creative team are mostly about hitting a small target with a number of angles, rather than thinking outside of the box to create something wholly original. If we're talking about brand advertising, it may be a different story.
"Display creative is becoming more automated -- much like media buying is," says Sarah Sikowitz Hite, VP, group media director at 360i. "Dynamic creative allows for creative updates based on the target audience -- all delivered in real time. For creative teams, this means templates that technology can update, rather than creating a number of different versions."
But while Sikowitz Hite believes it is possible for creative teams to keep up with automation, she's not sure targeted messaging is a panacea.
"It's important for both brands and media buyers to think about the incremental lift that targeted or personalized creative will deliver," Sikowitz Hite says. "For direct response brands, specifically those in the retail or travel space, dynamic creative can drive high ROI. However, for brands focused on branding and awareness, tailoring creative to narrow micro-targets may not deliver incrementally better results."
But Brian Carley, executive creative director at Rokkan, an agency that handles creative for a lot of targeted campaigns for major brands, sees it differently.
"The idea that you can match the right message to the right eyeball at the right time definitely does present some creative limits in terms of how much one agency or even a team of agencies can generate," says Carley. "I'm just not sure we've really hit those limits yet."
That may be because few brands are really going as far on the right person, right message, right time promise as targeting technology might allow. But remember, programmatic buying is still picking up steam. Media buying technology is improving, the data sets are becoming increasingly rich, and the marketing spend is growing. The question is, "Where does that leave the creative?"
"I don't think it is a great use of time to attempt to craft a million personal headlines and notes for a database of a million people," says Jason Norcross, executive creative director and partner at 72andSunny. "Rather, I think it's smartest to organize those people into specific groups (like recent users, lapsed users, etc.) and serve them a more general message (like, 'We've just upgraded our product, give it another try')."
Segmenting those audiences and connecting with them at the right time in appropriate channels, and perhaps even a few surprising channels, is the creative act, according to Norcross.
But even segmenting makes Carley somewhat suspicious. For one thing, he suspects there's also a limit to just how much consumers really want to hear from a brand. Aggressively pushing tailored messages to segmented audiences runs the risk of numbing consumers, says Carley, or worse, turning them off to the brand entirely.
But Carley's bigger concern is that the targeting that goes into programmatic buys really isn't all that targeted.
"You see a lot of media plans that look suspiciously like all the other media plans out there," he says. "The idea is to be very granular in your media plan and strategy, but a lot of what we see feels very cookie cutter."
Avoiding the cookie cutter
In a perfect world, each team that works on the campaign is looped in early and often.
"Early communication is key," says Sikowitz Hite. "Having kick-off meetings with both creative and media in the room -- even across agencies -- allows for work to happen concurrently. Additionally, clarity on specs and rich media partners upfront can stop a lot of the chicken and egg dynamic that often occurs between creative and media; media can plan around what it knows will be built, and creatives can build what it knows will run."
Norcross agrees. When good, early communication happens, the client usually gets the best result.
"We have had the most success by getting all partners -- media, PR, social, creative -- aligned on a brief with our clients very early in the process," says Norcross. "Then we all come together and share ideas. Get everyone focused on solving the problem, not filling buckets and checking boxes."
Norcross says he's a big proponent of what he calls "better beginnings." It starts with the best possible articulation of the client's problem or opportunity, better briefs, and better meetings, especially in the early going.
"Then you can better identify the right thing to do and actually get to newer, more interesting and, potentially, more innovative solutions," he says.
But while the process Norcross describes is certainly possible, it doesn't always work out that way. Sometimes the problem is that the media is bought way in advance, says Carley.
"That means you're trying to fit the creative to the media," he says. "That happens a lot, and it can work great, but the tradeoff is that your ideas need to conform to the media buy, which means you may not be able to go as big as you like."
Another common problem: clients that see situations with multiple agencies as an opportunity for a creative bakeoff.
"That can be exciting, and it can lead to a big idea, but pitting multiple agencies against each other isn't good for collaboration, and usually that's what the brand says it wants because we're creating so much creative for so many channels and scenarios, which is one of the reasons multiple agencies were hired in the first place," says Carley.
Are silos still the problem?
Even if a brand is genuine when it says it wants to collaborate across agencies, there's always that pesky problem of coordinating between teams.
"Silos still exist to some extent," says Carley. "But I'd say that the lines between teams have become really blurred in the last few years, and that's a very good thing."
But blurred lines between teams aren't the same thing as having integrated teams. Time and skill sets remain limiting factors. Each team has a full-time job, so while it's possible to loop them in with the other players, it's not necessarily feasible to cross-train those teams with any real depth.
As Carley puts it with a chuckle, "Creative directors all know the basics of what the media buyer does, but there's a lot of drop-off after that."
Don Draper or Harry Crane?
It's easy to think of the rise in automated buying as a kind of battle between Don Draper and Harry Crane -- or maybe Harry Crane 2.0. But it's not a zero sum game, at least not when you're talking about brand advertising.
"Custom creative executions are meant to live alongside specific content and, at their best, align the brand with the content in a unique way," says Sikowitz Hite. "Those programs are not going away. It's up to media and creative, along with the premium publisher community, to keep bringing those partnerships to life through compelling storytelling and new ideas."
Unfortunately, at the high end, all the challenges of aligning media and creative remain. While at the low end, it's clearly a media buyer's world.
But the way Norcross sees it, the agency's challenge is to weave together the custom creative that embodies the big idea with targeted opportunities that extend the message's reach to the largest possible audience.
"A broader audience might not see the custom campaign, but the hot molten core of your audience will experience it in a more meaningful way," says Norcross. "It'll feel bigger. And that is meaningful scale. Then you can use other sorts of tactics -- media and otherwise -- to create on-ramps to that custom campaign for a broader audience."
Michael Estrin is a freelance writer.
On Twitter? Follow iMedia Connection at @iMediaTweet.
"Two businessmen playing tug of war" image via Shutterstock.