Paul Cushman likes to start his talks to marketers by asking for a show of hands for how many people checked their email that day on their desktop or laptop. Pretty much everyone raises their hands. Then he asks how many did so on their mobile device. Same response.
For Cushman, the former head of Yahoo’s mobile sales and now managing director of Cross Screen Consultants, that’s as good an argument for cross-device targeting as you’ll find anywhere. Cushman has helped persuade marketers ranging from General Motors to Williams-Sonoma to spend $250,000 and up to get a clearer picture of how customers and potential customers are behaving across their desktop, smartphones and tablets.
Doing so is trickier than you might think. Until recently, most searches occurred on desktop. Using cookies – lines of code that marketers use to identify individual users – it’s fairly easy to see how consumers behave. If a customer’s in the market for a minivan, for instance, it’s likely that she’s done some searches for a Honda Odyssey or a Toyota Sienna, and automakers and dealers can send her retargeting ads.
On mobile, it’s a different story. Cookies don’t work nearly as well, so marketers often have no idea that you are the same person, even though you switch off between your MacBook Air, iPad and a Samsung Galaxy S5.
Helping marketers create a composite sketch is big business for Facebook, but also for more than a dozen firms offering "probabilistic" solutions tapping big data. The only hitch is that the Federal Trade Commission might ruin the party by making everyone pull back in the name of consumer privacy.
Why this information is important
Back to the web surfer trying to get information about a minivan. If Toyota knew that user’s online identity, then the company or local dealers could send her an offer on the model when she is near a dealership.
That’s what at least one major carmaker in France is currently doing. "Being able to connect location and behavior on different devices is critically important, especially for someone like an automaker, where the price tag for each item is so high and so valuable," says Michael Moreau, Chief Solutions Officer and GM Interchange at Krux, a vendor that works with the unnamed carmaker. (Unfortunately, vendors are fairly paranoid about using real brand names.)
Chuck Moxley, CMO for 4Info, another firm offering cross-device tracking services, says that the purchase cycle for buying a car lasts weeks, so "our point of view on it is you want to reach them on the desktop when they’re actively searching, but you also want to reach them when they’re in lean-back mode on their iPad while they’re watching TV and at the same time researching cars."
Ian Dailey, senior product marketing manager at RocketFuel (a vendor offering cross-device targeting), says a campaign using desktop and mobile together can increase conversions by as much as 30% versus just one or the other. Dailey often poses a question for marketers: "For the same amount of money, do you want 30% more conversions?"
In addition to ad campaigns and direct-response reminders, marketers can also use cross-device tracking for attribution. If you run a banner campaign on desktop, it’s valuable to know whether that same consumer searched for your product later on her phone or, even better, bought it.
Some firms, like Persio, close the gap on the latter with mobile coupons. In Persio’s case, a retailer can guide a consumer to download a coupon on desktop which she can send to her phone. This not only connects desktop browsing behavior with a phone number, but reports when the coupon is activated and a sale occurs.
Perhaps the most compelling reason for cross-device is that users tend to search on mobile and buy on their desktops. If you’re sinking a lot of money into mobile advertising, that’s important information.
Deterministic vs. probabilistic
There are two primary ways of linking desktop and mobile use. One is when an app gets a user to log in on both platforms. When the app is Facebook, that’s a huge advantage. With 1.4 billion users, Facebook has a huge slice of the market, and its users frequently switch back and forth between desktop and mobile.
Facebook recently exploited this cross-device identification with Atlas, a network of affiliated sites and apps that is able to target on desktop and mobile. Facebook’s not the only probabilistic vendor, but it’s currently the biggest since Google hasn’t offered its own solution (except for cross-device attribution) and Apple isn’t likely to. Still, there are lots of other players, including Yahoo, Spotify and the New York Times.
If you’re a New York Times subscriber, for instance, and log on to both the website and mobile apps, then Krux tracks you.
4Info, another firm that offers cross-device tracking, takes a different approach: The company taps data from Nielsen Catalina Solutions and cross-references it. Moxley says the company can link 95% of mobile users to their address and their other devices. "We know who we serve an ad to, and we can pass that on to Nielsen Catalina Solutions and look at actual sales by those people during that week or the week after," he says, noting that the company doesn’t use the consumers’ names for targeting.
Another approach is to look at various data, including Wi-Fi networks, location, device types, operating systems and many other data points to make an educated guess about who the user is.
There are at least a dozen firms currently offering probabilistic tracking including Tapad, RocketFuel, Krux and Adelphic. Probabilistic firms say they offer a more expansive view of the entire market beyond the walled gardens of Facebook, Yahoo, eBay and others. (Complicating the picture a bit, probabilistic firms often use some deterministic data in their equations.)
"We believe that deterministic methods are the best, they’re simply more accurate," says Max Kalehoff, CMO of SocialCode, a social advertising managed services provider. "We refer to networks like Facebook as ‘identity-based networks’ because they’re allowing advertisers to target, reach and measure against known audiences, not theoretical audiences."
There’s a saying in ad tech: You’ve really made it when you get the FTC interested. If that’s the case, cross-device has hit the big time. In November, the agency plans to host a workshop about cross-device tracking, looking at, among other things, the privacy and security risks associated with the practice.
One idea, posed by Jonathan Mayer, a Stanford University grad student and lawyer who has studied cross-device tracking, thinks the practice should be "opt-in," meaning consumers would automatically be opted out unless they decide otherwise.
On the other hand, consumers are getting free apps and services in exchange for looking at ads. A similar arrangement has supported broadcast TV for 70 or so years. "We believe consumers believe there’s a bit of a tradeoff here," says 4Info’s Moxley. "They’re going to get ads for products that they actually buy instead of incontinence products or pay-day loans that they have no interest in. Those are the guys who are advertising in a non-targeted way. So if they opt out, they’ll start hating the ads again."