Marketers know that consumers shift devices throughout the day, engaging with their brands by visiting a website, watching a video, downloading an app, making a purchase, or completing any other measurable conversion event. To reach these consumers wherever they are to drive brand interactions, marketers have long relied on deterministic, or login-based, approaches to reach consumers across devices.
But that approach is changing as a new method makes inroads—and allows marketers greater flexibility.
Deterministic has long been associated with the large social networks and Internet properties. These networks use consumer-provided data, including personal information such as phone numbers, home addresses, birth dates, family members' names, interests, sexual orientation, political views and even medical history and credit card numbers. Solutions that rely on this kind of data have been thought to be the most accurate, so marketers bet on the deterministic model.
They started with Google, until Facebook came along and it became a two-horse race for where to place the majority of digital marketing dollars. It's almost a rehash of the beloved old IT department line, "Nobody ever got fired for buying IBM," with the modern marketing version being, "Nobody ever got fired for buying ads on Facebook and Google." But now there's a viable alternative technology that's becoming the catalyst for shifting mindsets on this front.
This method is probabilistic, or prediction-based, technology. It involves collecting nonpermanent, user-resettable identifiers such as browser cookies and device IDs, and using big data applications and machine learning to correlate that information to predict device ownership, demographic information, interests and other attributes.
For years, marketers believed the myth that probabilistic identity was 50% to 70% accurate at best, and deemed deterministic, login-based solutions preferable. Until now, they had no way of comparing both solutions, so fear, uncertainty and doubt persisted. Recently, Nielsen tested the accuracy of the two leading probabilistic solutions, with both scores topping 90%, and Drawbridge coming in at 97.3%.
With two probabilistic heavyweights now weighing in with strong accuracy scores, the accuracy argument is losing steam. As the accuracy gap is being bridged between deterministic and probabilistic, this leaves cross-device scale, or potential audience reach, as the main differentiating factor between the two methods. But even the perceived gap in scale is narrowing—and so is the difference in the number of addressable consumers for each method.
Facebook is the undisputed leader in active cross-device consumer reach with over 1 billion users logging in across devices on a regular basis. Google is also a large player, as are Yahoo! and Twitter, each with hundreds of millions of recognized cross-device consumers. Probabilistic solutions easily reach these levels—and now we know they can do it with deterministic-like accuracy. Because probabilistic models are built on data signals from ad requests across the Web, and aren't limited to closed, walled gardens, the potential scale is much larger than any single deterministic environment.
With probabilistic accuracy being benchmarked, it's hoped that mindsets that shun probabilistic methods across the board will start to change. An example of that mindset is when Jon Hook, former head of mobile at Mediacom, said, "I spend a lot of time trying to answer the question: 'Why aren't clients spending on cross-screen advertising?' You're still struggling to nail cross-screen unless you are Facebook or Google. Otherwise, it can end up seeming like a very expensive game of 'Guess Who?'." I think we need a shift in thinking here, and I hope these recent developments start changing these mindsets. Cross-screen is not a struggle for technology providers—it's ready for mainstream brand budgets.
Speaking of data being made available, probabilistic identity solutions enable marketers to leverage their own customer data in addition to identity data to get a complete picture of their customers across devices—bridging desktop data to mobile devices or mobile data to desktop devices. In this sense, probabilistic identity has the ability to be democratized across the Web on just about any connected device. And probabilistic identity solutions can solve the dilemma of cross-device attribution—for both online and offline conversion events.
About the Author
Brian Ferrario is VP-marketing at Drawbridge. Prior to joining Drawbridge, Brian was VP-marketing at Sociomantic, a leading programmatic advertising technology company, where he was part of the executive team that helped the company be acquired by dunnhumby, a Tesco company, the world's second largest retailer. He previously joined digital advertising technology company Rocket Fuel in 2008 as employee No. 13 and was instrumental in helping the company achieve record category growth; numerous innovation, cultural and revenue awards; and huge traction in the market that helped lead to a successful initial public offering.
About the Sponsor
Drawbridge is the industry's leading cross-device technology company that enables brands to have seamless conversations with consumers across their connected devices, including desktops, smartphones, tablets and connected TVs. By leveraging its Connected Consumer Graph, which includes more than 1 billion consumers across more than 3 billion devices to date, the company is able to gain insights and a much deeper understanding of consumer behavior to drive better results for advertisers—from creating brand awareness to driving incremental sales. For more information, visit Drawbridge.