Anonymous and anonymized do not mean the same thing in customer data
– and this is not a semantics lesson 🤓. A few weeks back, I was on the phone with a prospective customer who has been leading a task force at his company for a fairly massive marketing technology project. The goal of the call was to provide some clarity around the differences between CDPs and DMPs, and this “anonymous vs. anonymized” discussion was one of the points that came up. I thought it’d be worth zooming in on it a bit.
This difference between the anonymous and the anonymized is at the very root of what distinguishes a customer data platform and a data management platform. It is through the explanation of this point that it is revealed why a DMP will never be able to get you close to a unified customer view.
Anonymous is defined as “without any name acknowledged” or “of unknown name; whose name is withheld.” At any given time, an individual’s anonymous status can change to known – with a name acknowledged or provided. In customer data and marketing, a person is often anonymous and may be able to do a number of things to change their status from anonymous to known. For example, someone browsing a website without having logged in or providing some identifying piece of information is browsing anonymously. Their behavior is still theirs, and the information may still be associated with that person.
Anonymized, on the other hand, is data in a transformed state. The acknowledged name or other identifier has been deliberately removed. An explicit choice has been made to scrub identity away – the only way for data management platforms to do what they do without violating any privacy regulations. A developer who works on a leading DMP explained that the tool “can’t augment a profile that’s addressable (i.e. identifiable) with data that is collected anonymously” or vice versa. The key differentiator between how a DMP and CDP handle first party data collection and merging with 3rd party data: this may only occur when the data doesn’t have an identifier or when that identifier has been removed. While this has highly useful applications, it’s also a considerable hole if you’re trying to build up a database of first party customer data.
Marketers need addressable, identifiable people who actually interact with their brand. In order to make use of this customer data across the marketing technology stack, these details are critical if not required. Anonymized is good enough for the kind of probabilistic groupings that are the basis of reach; it’s decidedly not when you aspire to create an accurate, deterministic dataset about your customers.
Customer Data Platforms deal intentionally in first party, identifiable or potentially identifiable data collection.
Data Management Platforms deal in audience creation based on scrubbed, aggregated and/or non-identifiable data.
Hope that helps!
P.S. In the interest of intellectual honesty, I have to admit the anonymized isn’t actually a word.