Identity Graph, Identifier, & Identity, Explained

Marketing professionals and other business users are grappling with a lot of meaty topics nowadays — none more so than the ever-changing concept of customer identity.

It’s a difficult and important topic because identity is uniquely tied to so many meaty topics – notably customer data, consumer privacy, digital experience, and measurement.

"Identity" raises existential questions, both theoretical and practical, for business users. And answering them goes right to the foundational decisions every company must make.

Key takeaways

• Identity refers to an individual person, while an identifier is information used to recognize that person, such as an email address, device ID, cookie ID, or customer record.
• An identity graph associates multiple identities and identifiers with one master ID, but it does not consolidate the underlying customer data into a persistent profile.
• Identity graphs can help marketers recognize relationships across digital and addressable channels, while customer profiles provide a persistent repository for unified first-party customer data.
• The right identity approach depends on how the data will be used, the types of identifiers involved, and the level of matching precision required.

Identity vs. identifier: What’s the difference?

In the context of marketing and customer data, there are two terms that are often used interchangeably, but shouldn’t be confused with one another:

• Identity is conceptually simple. Identity refers to an individual person, and it can evolve and change over time, much like the person herself. For example, though you remain the same human being, your identity as a student is replaced by employee when you enter the workforce. Your identity as a value-shopper may change if you get a big raise. The fact that you are one, distinct person never changes. Attributes that make up your distinct self, though, absolutely do. Thus, identity is fluid and ever-changing, but inextricably linked to a single person.
• An identifier, on the other hand, is a piece of information that helps you recognize a distinct person. Depending on the context, an identifier might be an anonymous cookie ID, a device ID, an email address, or a customer record. Each of these can be useful in their way, but they are decidedly not all created equal in terms of marketing utility, and certainly not when it comes to privacy.

Using identifiers, marketing technologies like the customer data platform (CDP) give you a mechanism to create ways to maintain identity within a system.

With customer profiles, business technology users gain a mechanism to maintain a single repository of information about an individual person.

Different kinds of systems support different profile types and structures. And each has its own strengths and weaknesses when it comes to marketing utility.

For example, a few profile flavors include:

• Anonymous vs. known profiles
• Cookie-based (a.k.a. "client-side") profiles, which are usually deleted within days vs. persistent profiles (a.k.a. "server-side"), which store data indefinitely
• Pre-defined attributes, in which the system determines what types of data it will store in the profile
• Extensible data storage profiles, in which business users can choose what data and in what format the data is stored — structured or unstructured data, text or whole number, synthetic or calculated profiles

Most companies today have dozens of disparate profiles associated with a single person, largely because of fragmented and siloed data management across the organization.

But the quest for a single customer view, which sparked the meteoric rise of customer data platforms, led business users to reconcile these profiles into one that represents an individual and their 'complete' identity.

Then there's the identity graph.

What is an identity graph?

An identity graph is a moment-in-time view of customer identities and the relationships between them across disparate systems, channels, and data sources.

Although the terms are often used interchangeably, an identity graph and a customer profile are distinct concepts. Both are constructs of customer data aggregated from disparate systems, channels, and sources.

While an identity graph provides a snapshot of the current, raw state of your customer data, this snapshot isn’t stored beyond that moment in time. With an identity graph, the data itself continues to reside in separate systems and can be queried to create a current view of the data, including to account for any changes.

In other words, while a customer profile allows you to resolve multiple identities into a single, consolidated identity, an identity graph allows you to associate, but not consolidate, multiple identities with one master ID.

For first-party data to be unified and actionable across disparate systems and for different programs (e.g., advertising, direct mail), the best home for a single source of identity truth is a profile, as it can be readily utilized by an identity graph for certain use cases.

Identity graph example

Consider a customer who first visits an e-commerce site without logging in. At that point, the brand may recognize the person only through an anonymous device or browser ID. Later, the customer makes a purchase using an email address, creates an account with a customer ID, and joins a loyalty program that assigns another identifier. Information associated with that customer may now exist across the website, e-commerce platform, CRM, loyalty program, and other systems.

An identity graph can establish relationships among those identifiers and associate them with the same individual or master ID. When the graph is queried, a marketer can see that the anonymous website visitor, purchaser, account holder, and loyalty member likely represent the same customer rather than four unrelated identities.

The important distinction is what happens to the underlying data. An identity graph connects these identities while the customer data continues to reside in its original systems. It can create a current view of those relationships as identifiers and data change, but it does not consolidate all of that information into a persistent customer profile. A customer profile, by comparison, resolves and stores the customer data as a unified record that can be used across marketing and other programs.

The role of customer identity graphs and identifiers in modern marketing

The rise of digital channels where consumers engage with brands — on websites, mobile apps, social platforms, etc. — drove the creation of solutions where marketers can recognize people by all their different proxies.

Then, marketers' wanted to stitch all of the proxies together into one because it improves business results. Identity is an essential pillar of marketing today.

(Even TV, the most mass media of all, isn’t immune, between addressable TV and OTT.)

Customer identity plays a variety of roles within marketing, contingent upon:

• The type of identifier used to distinguish one person from another. A cookie ID and an email address are fundamentally different identifiers, both functionally and in terms of privacy and value to the marketer.
• Where the data is collected and how it will be used. First-party data and third-party data both use identity but cannot always be used in tandem (e.g. DMPs have to use pseudonymous identifiers and anonymization techniques to prevent co-mingling data collected for different purposes). Compliance with data collection laws and regulations such as GDPR also influence this.
• Matching techniques and use cases. Fuzzy or probabilistic matching is, by definition, an imperfect science. Match rates between one dataset and another vary widely and can be difficult to test if the matching dataset is a third-party black box. Deterministic matches are more precise but happen less frequently because of a higher threshold for the data. Each has its purpose but should be considered in the context of how you will use the data. You can tolerate less precision for a massive acquisition ad campaign. Imprecision, though, would spoil a campaign that targets a unique attribute of a small set of individuals with a particularly high value and specific set of traits.

Identifying the right identity solution

I spoke during a recent LiveRamp RampUp panel discussion, “Three Flavors of Identity Solutions.” The question at the heart of the session: Do marketers go with a CDP, an onboarding solution, or a graph (i.e., an identity or device graph)?

My answer? It depends on what you’re trying to do, but here's a good rule of thumb:

• Identity or device graph: Anonymous or anonymized, graphs focus on tracking across digital and addressable channels.
• Onboarding solution: Onboarding is most effective for improving advertising campaigns by starting with first-party data and matching based on attributes available in paid channels.
• Customer data platform: With persistent profiles, CDPs unify all first-party customer data and associated identifiers, which you can then activate in a variety of marketing applications — for measurement, campaign orchestration, personalization, et cetera.

Watch our on-demand webinar on customer identity to learn how you can improve your cross-channel engagement efforts with a pure-play CDP like BlueConic.

Frequently asked questions

What is the difference between an identity graph and a knowledge graph?

An identity graph is designed specifically to connect identifiers that may belong to the same person, household, or device. For example, it might associate an email address, customer ID, cookie ID, and device ID with one master identity.

A knowledge graph has a much broader purpose. It organizes entities, such as people, products, companies, locations, or concepts, and maps the relationships between them so that people or systems can understand how those entities are connected. In other words, an identity graph focuses on resolving who someone is, while a knowledge graph can represent many different types of entities and relationships.

What is the difference between an identity spine and an identity graph?

The terminology can vary between identity providers. Generally, an identity spine refers to the persistent identifiers or identity framework used to anchor and connect customer identities over time, while an identity graph represents the broader network of relationships among those identifiers. Some vendors use the terms almost interchangeably, so the exact distinction can depend on the technology or provider being discussed.

What is the difference between an identity graph and a customer profile?

An identity graph associates multiple identities or identifiers with the same person or master ID while the underlying customer data can remain in separate systems. A customer profile goes further by consolidating customer information into a persistent record that can serve as a unified source of customer data. This distinction makes customer profiles particularly useful when first-party data needs to be continuously available for personalization, measurement, campaign orchestration, and other marketing activities.

‍