Identity Resolution Makes Possible a 360 Degree Customer View

Identity Resolution, Name Matching, Record Management

identity resolution for 360 degree customer view

Getting to a 360 Degree View of Customers Is Much Harder Now

The arrival of technologies such as mobile devices, on-line customer support, social media platforms, discussion boards, etc. has multiplied enormously the number of ways in which customers can interact with a company in addition to the traditional physical brick-and-mortar store. This has greatly complicated the creation of a unified view of each customer. All the incoming data via multiple channels needs to be aggregated, merged, and de-conflicted – both the profile of the customer, starting with basics such as name, address, phone number, etc., but also such items as the nature of each individual customer’s interactions (e.g, purchase, support request) and its outcome (sale, problem solved, etc.). The favorite term for this is the “360 Degree Customer View.”

A Company Needs to Make Sure the 360 Degree Customer Data is Available via All Appropriate Corporate Systems

The company also has to ensure that this information is available equally to all the relevant corporate systems and is kept faithfully updated. The ultimate goal is to make sure that when a given customer makes contact after previous ones, all their information is available to the company staff or system to provide them with the best possible customer experiences. In the world of 360 degrees, for example, a customer should only have to provide requested information once and not have to painfully and repeatedly disclose it. That’s not the way to build customer engagement and loyalty. A customer should always have a feeling of predictability and ease in every interaction.

It Can Be Hard to Recognize that Multiple Customer Engagements Reflect the Same Unique Customer

A major challenge in realizing the 360 Degree View is determining, out of all the very large number of records resulting from customer-company interactions, which ones represent the same person. Records for the same customer may have some common field entries, such as a phone number, but the names may vary, e.g., “Steve Bennett,” “Stephen Bennett,” “S. A. Bennet,” and so forth.

Nicknames, typos, and word order variations are also a problem. Finally, since the same name may refer to many different people, making a match on just the name is not sufficient. There also has to be a way not only to match on single fields, but to combine the likelihood of a match for each field into a unified likelihood over all fields.

Identity Resolution Is a Critical Technology for Developing a 360 Degree Customer View

Fortunately there is a technology, Identity Resolution (aka Entity Resolution or Entity Matching), which can figure out which records represent the same customer. When a potential customer makes contact via some channel, Identity Resolution matches the record created by the information that the customer discloses against a master repository of the known customers. If a match is found having a high similarity score, the new record can be merged automatically with the already existing record. If there is no match, a new record is added to the repository.

To merge customer records accurately, there needs to be intelligent matching of each field: these include names but also other important attributes such as

  • Email
  • Phone number(s)
  • Address
  • Date of Birth
  • etc.

(Of course, some records may contain only a subset of these fields). These fields contain different types of items, and the matching needs to be able to handle all types.

For example,  a date like “5/21/1978”  is very close to “5/22/1978” in absolute terms and the matching consequently needs numerical knowledge, but the matching also has to be capable of deciding that “5/21/1978” could be considered very close to “5/21/1988,” as the one-digit difference, though ten years later, could be a simple typo. That type of very fine-grained knowledge needs to be built into the matcher.

By contrast, person names exhibit a much wider diversity of variations that requires a sophisticated matching algorithm. For example, in addition to the example given above of full name vs. initials, there are other common phenomena including:

  • Nicknames: William vs. Bill vs. Billy; Mikhail vs. Misha; Alexandra vs. Sasha
  • Name Order Variants: John Smith vs. Smith, John
  • Missing name elements: Ali al-Sistani vs. Ali Sistani (Arabic names have components that are frequently dropped in English.)
  • Names that sound the same but are spelled differently: Sean vs. Shawn vs. Shaun
  • and many others.

Identity Resolution Scales Well Too

Identity Resolution must also be highly scalable because consumer records are a Big Data problem. A big retailer may have hundreds of millions of records and every new customer contact has to be matched against that very large repository. This needs to be able to happen in real time during the contact, as a successful match may well provide the company rep with a critical piece of data that improves the quality of the interaction.

In sum, Identity Resolution is a powerful technology that contributes a lot to the challenges in getting a full view of customers that have been created by all the new communications channels we have today. It gives companies a much more complete picture of their customers and would help achieve best possible customer experience to enhance customer loyalty.