Karen Burka, Senior Research Consultant at Third Door Media, and Kurt Hawks, CRO of FullContact, break down the foundations of modern Identity Resolution, and how it helps you create personalized customer experiences and comply with the increasing array of privacy legislation.
Summary
A two-part session pairing an analyst’s overview with our own practitioner detail.
Karen Burka opens by comparing several published definitions of identity resolution to draw out what they share: persistent recognition of a person across devices and touchpoints, in a way that is accurate, scalable and privacy compliant. She then breaks the capability into four parts, with privacy compliance at the center rather than bolted on at the end.
Our half takes the question people actually ask us, which is how this differs from tools they already own. The short answer is that a CRM, CDP or DMP is a storage layer for what you already know, and a delivery platform is a rules engine acting on it. Neither one connects the undeclared identifier to the declared one, and that is the specific job we do.
We close with a crawl, walk, run sequence and take questions.
What We Cover
- This is not a replacement for your stack, it is a layer underneath it. Storage systems hold what has been declared. Delivery systems act on what is held. We connect the identifiers that were never declared to the person they belong to, which makes both of the others work better.
- Nor is it deduplication or address cleansing. Those clean up one channel’s identifiers, matching two spellings of a name. We unify the different ways one person represents themselves across entirely different channels.
- Deterministic does not mean right. Directly observed data can still be wrong. Someone logging into an app with their partner’s email is a real observation of a false link. Deterministic data is a signal to weigh against other signals, not a guarantee, and we say so about our own data.
- No single identifier is enough on its own. Asked whether a hashed email makes a good persistent identifier, the honest answer is that the value is not in any one identifier but in tying several of them, plus other identifier types, to one person. Any one alone will fail eventually.
- Your customer file is probably smaller than your record count. In one engagement, a file of 153 million records resolved to 96 million actual people, with 15 million of those sitting in both the customer file and the prospect file. Those are existing customers being marketed to as strangers.
- Most of the people who do not log in are already your customers. In a second engagement, matching anonymous website visitors back to the customer file showed roughly three quarters of the recognized visitors were existing customers rather than prospects.
- Unify first, then integrate in real time, then orchestrate. Crawl is resolving fragmented customer and prospect data to find out whether you have a problem worth solving. Walk is recognizing people at the moment they interact. Run is feeding that into personalization and orchestration.