Recognizing people online has only become more complicated over time. If we look back just 10 years ago, the most common marketing channel was direct mail, where marketers identified customers and prospects through mailing addresses and sent them the latest campaign, product or (physical) newsletter.
As the internet grew in prevalence, more marketing channels developed. Email gave marketers the option to stay connected with consumers daily, and smartphones eventually brought us in-app messaging and more opportunities for exposure.
Now, everyone has multiple email addresses, devices and browsers. Consumers buy things on any one of their various phones, computers or tablets, and devices like Amazon’s Alexa or Xbox create new unexpected engagement opportunities that are shaping the future of consumer behavior. And smart appliances are expected to see strong financial growth as companies like LG and Samsung introduce products allowing consumers to speak to their appliance and order groceries.
The amount of consumer touchpoints keeps expanding, and it’s only going to get harder to find and recognize each person across all of their channels and devices.
Dive deeper with the guide: 5 building blocks of identity management
Every single point where you could interact with a customer needs to be tied into one holistic view of that person through proper identity management. Yet this proves to be the most important—and most challenging—aspect of successful omnichannel marketing.
How you can recognize people online
When marketers talk about recognition in digital marketing, it’s often with a very basic understanding. The key question is: “When I see a cookie or device in the digital world, how much historical information do I have on that person?”
Although there are many solutions in the market that attempt to create recognition across devices using emails, logins or IP addresses, none of them are a substitute for matching your first-party data to an established and verified network of real people, always in a privacy-protected environment. Real, verifiable information about a person is the best anchor for matching.
Matching methods for building customer profiles
There are two common methods used to identify customers in the digital world:
- Probabilistic matching, where an attribute or multiple attributes are given a score to make an educated guess about a person’s identity. For example, if you see that two device IDs are on the same IP address, using probabilistic matching you might assume this is the same person and combine the profiles of each device. Your guess could be right, but it could be a family member or friend using the home wifi. Even though probabilistic matching may be the easiest option to generate scale, it will not be the most accurate, which can lead to targeting, personalization and performance issues.
- Deterministic matching uses personally identifiable information (PII) to identify an authenticated user of a given device. Using real names, addresses, transactions, hashed emails and customer IDs—after being scrubbed clean of PII—you know for certain who this person is. This is the most accurate way to match real people to their online activities because it is rooted in the person’s real information—who they are, where they live and what they buy—but it removes the identifiable information before being matched to that person’s privacy-protected online profile.
When it comes to matching, most marketers only think about one number—the match rate—but recognition goes far beyond that single metric. A provider could offer a high match rate, but that doesn’t mean you can accurately communicate with your customers.
Depending on how a provider matches your customers and prospects to online IDs, you could be sacrificing reach, accuracy or persistency just to achieve higher match rates.
Deterministic matching can help solve many of these issues; however, very few companies have the capabilities to execute this way. And of the ones that can, fewer still anonymize the person’s identity for a privacy-first approach.
Avoiding repetitive conversations
Recognizing your customers impacts not only your marketing performance but also the customer experience. Think about Dory from Finding Nemo. Dory meets Marlin for the first time, introduces herself and has a conversation with Marlin.
Two minutes later, Dory doesn’t remember meeting Marlin, and they have the same conversation all over again. This happens all the time in digital marketing; just because you can match a single cookie doesn’t mean that you’ll be able to have an ongoing conversation with your customers that builds over time with every interaction.
Recognition is the fundamental first step in digital marketing because it builds the foundation for a successful campaign and performance. Good identity recognition ensures you know who you’re talking to, across channels and devices, without losing sight of each person over time.
Ultimately, identity recognition impacts how you:
- Find the right people to message. You should be confident in who you are messaging. Are they the target audience you were hoping to reach? Are you contacting your existing customers when you wanted to reach new prospects? Even if you can identify your customers, it might not always be the right audience, time or place to message them.
- Determine what message is best to send them. The message you send to each of your customers should be unique and personalized to the right content, format, time of delivery and preferred device. For example, are you sending the wrong product offers to the wrong people? Only if you can accurately build a holistic profile across devices will your message be relevant enough for your audience to respond.
- Measure your delivery with accuracy. Accurately identifying your customers is a critical step in measuring program success. If not, what is the real cost of messaging those people? Knowing that your customer saw an advertisement on one device but bought on another can dramatically change how you view the impact of marketing campaigns.
Marketers often chase the cheaper or quicker solution, but there are a lot of opportunity costs with that approach. If you work with a provider that focuses on probabilistic matching, you’re not getting the full view of each person.
And—like Dory—you’re probably restarting the conversation over and over with the same person—or worse—speaking to the wrong person entirely.