e approaches to merging profiles for resolving identity

Due to the fragmentation of digital channels, marketers are now receiving more signals from their customers than ever before. To deal with this, you need to have a clear identity resolution approach. How do they match identifiers? How confident are the matchmakers? A clearly defined strategy will lead to a better customer experience and more effective communication.

The first step is to adopt as a framework in order to develop an identity resolution strategy. You can choose to focus on the device first, the person first, or combine both. It is important that marketers have confidence in the identifiers used to maintain a customer view throughout the digital journey.

Probabilistic vs. deterministic matching

A single customer can have multiple profiles in your database if they interact with your company through different digital channels. To merge these profiles, you need to find a way to match devices, digital accounts, and other identifiers.

Two main methods are available.

Deterministic Matching. You will only merge profiles when the match is 100% certain. This happens most often when a common identifier appears in several profiles. If your customer placed an order with an email address, then the postal address and phone number that was found in the order can tell you what other profiles the customer is associated with. These common identifiers allow for a deterministic matching.

The email address can be included in a profile of a client, and the data from that customer can then be combined with information separately stored by the email service provider.

Probabilistic Matching AI uses behavioral data and other signals to determine the likelihood of separate customer interactions being from the same person — without using a common identifier.

The MarTech Conference featured Greg Krehbiel of The Krehbiel Group as a consultant.

Krehbiel argues that even deterministic matches aren’t 100% reliable. He gave an example that is by no means uncommon. Krehbiel’s mother asked her sister to assist with the Christmas shopping. The sister used her laptop and his mother’s credit card to buy presents on Amazon. Deterministic matching could lead to the conclusion the mother used the daughter’s computer.

Create a single source for truth

When merging profiles of customers and matching identifiers, there is always a certain amount of judgment involved. It is important to have a single source for all your customer data. This will ensure that the results are as accurate as they can be.

Krehbiel said, “You want a single record of a customer to the extent you can and that means merging other records.” “This means that a single system must be the source of truth, no matter what it is,” said Krehbiel.

Consider all possible uses for a channel, and then decide where to consolidate the data. It will avoid duplication and conflicting sources of truth.

Krehbiel asked: “If someone changes the email address in [the CRM], will that overwrite things in the ESP?”

Merging profiles may not be the best solution in some situations. For example, some customers prefer to use separate emails for personal and work purposes. In this case, you shouldn’t merge the emails. Your organization should instead think of this client as a multifaceted individual with multiple emails.

Penske Media’s CDP helps advertisers reach digital users

Scale of Confidence

Both deterministic and probabilistic matches depend on the accuracy of the data when merging customer profiles and resolving their identity.

Marketers should rate their level of confidence on a sliding-scale based on the use cases – how they plan to interact with customers.

The calculation is necessary because the assumption made when merging multiple profiles can be shaky.

Krehbiel said, “There are always edge cases.” “Sometimes, one person can have multiple email addresses. But sometimes, one email address can be used by multiple people.” I also know of some families who only have one email for their entire family. In general, these are rare cases and you shouldn’t worry about them. It’s not a big deal if it causes problems. It might be — you just have to consider your use cases.”

This confidence scale can be used to determine the likelihood of getting a data point wrong. A wrong postal address, for example, might result in a waste of postage on a direct-mail piece. If it is part of the customer’s account and they see that you have entered the wrong address, then they may be more concerned about their privacy and how your business manages data.

It depends again on the use cases that are specific to your company. This is another example: if your company delivers food and a customer has a severe peanut allergy, you will need to provide accurate information.

These use cases will help you make confident decisions when matching identifiers and merging profiles. This will ultimately allow your team to deliver the best experience to your customers.

Register here for The MarTech Conference.

The post 3 ways to merge profiles when resolving Identities first appeared on MarTech.