People-based marketing is a term used in every corner of the martech landscape, but here’s the rub: People-based marketing requires people-based identity, and not everyone has figured that out yet. They’re starting to, though, and 2019 will be the year when this understanding reaches critical mass. Marketers across the board will realize that it is only by connecting data to real people that true people-based marketing can take place.
In the past few years, martech leaders have leveraged breakthroughs in technology to reach new heights in delivering highly personalized, valuable, and immediate experiences. This has drastically raised consumer expectations for interactions with every brand with which they engage. The reality is, today’s consumers expect individualized interactions—not just personalized, individualized—that optimize the shopping experience by showing them what they want when they want it, while saving them time and money. And it’s easier than ever for them to take their business elsewhere if they don’t get what they expect.
People-based identification is the key to delivering individualized interactions. Simply put, it is the ability to recognize consumers as the people they are, whether they’re online, offline, mobile, in-store, and wherever else they’re interacting with your brand. It enables marketers to move beyond anonymous cookies and third-party data that guess at a shopper’s identity to build a holistic understanding of each consumer. Too much of what’s called people-based marketing today uses cookie-based guessing strategies, with the wrong messages going to the wrong people.
Identity resolution is the key
Identity resolution is what links someone interacting with your brand to a universal identifier that’s directly linked to that person, such as an email address (instead of a browser, device, etc.). This is typically accomplished with an identity graph. Both “universal identifier” (or “universal ID”) and “identity graph” (or “ID graph”) are terms you should expect to hear more of as people-based identity becomes mainstream.
A universal ID is a unique, cross-device identifier. The walled gardens like Amazon have built-in universal identifiers because users log in, but most retailers don’t have that luxury. Fortunately, the email address has emerged to serve very well in this function. It is often easy to capture a web visitor’s email address, whether or not they log in. And, because it’s typically linked to home addresses, social media accounts, apps, loyalty programs, in-store kiosks, etc., it’s also an ideal connector to other identifiers.
An ID graph is a database that houses all identifiers known to correlate with an individual consumer and connects them to a universal ID. The data used to form an ID graph is extremely important and should be deterministic in nature. “Deterministic” and “probabilistic” are additional terms that you’ll hear a lot more of. Basically, deterministic data is judged to be 100% accurate because it self-authenticates, meaning it can be clearly linked to the universal ID upon capture. Probabilistic data cannot be judged 100% accurate because algorithms are used to produce a probable match.
That’s not to say that probabilistic data doesn’t have its uses when used judiciously. While it’s important to keep the ID graph based on deterministic data, probabilistic methods can be added outside the ID graph to expand marketing reach when there is little cost or consequence if messaging goes to the wrong person (such as traditional basic ad retargeting).
When talking about deterministic and probabilistic data, you’ll inevitably hear the terms first, second, and third-party data. First-party data is a brand’s own proprietary data collected via their digital or offline channels (i.e. websites, apps, social, in-store). When collected and managed properly, this data is deterministic and the most valuable for populating an ID graph. Second-party data, loosely defined, is when a brand has access to another brand’s first-party data. When used in the context of a verified identification network—another term you may hear more of—it refers to a network of first-party, deterministic data from multiple brands. In this case, brands do not have access to other brands’ data, but the aggregated data is used to improve identity resolution across all clients.
Third-party data is accessed via an outside company or platform. It is often aggregated from a variety of sources that may or may not use proper capture and management procedures, and in most cases, cannot be treated as deterministic. Unless you are 100% sure the data is deterministic, it should not be used to inform an ID graph or its polluting effects can be exponentially detrimental.
Who’s on first? What’s on third?
One thing should be increasingly clear. As people-based identity becomes the new gold standard, first-party data is incredibly valuable—even more so than in the past. Every retailer should be making their utmost effort to ensure that first-party data capture and management are at maximum quality and capacity. The multiple facets of first-party identification can be tricky though, because in most cases, data is captured through some type of website tagging system, and inevitably, tags break. To keep your ID graph accurately populated, be sure that data capture is audited on a regular basis.
Beyond its importance for establishing identity, first-party data also makes it much easier to address expanding privacy regulations—because the further you are from the capture point of data, the more challenging it is to ensure that these requirements are met. With first-party data, you set the standards for how data is captured and used. In contrast, third-party data is typically collected from multiple sources and handed over multiple times, making it nearly impossible to ensure that privacy standards are adhered to.
It’s a lot to digest, but here’s a good way to think of it. If people-based identity is the new lifeblood of eCommerce, then the identity graph is its beating heart. Captured identity data becomes the blood cells, delivering oxygen to the heart to keep the system alive. First- and second-party data keep the heart healthy and strong, while third-party data is anemic and can ultimately lead to system failure.