An identity graph contains linked records of identifiers and attributes that together describe a real individual. At its core, it stores the various digital and offline signals a person generates, such as email addresses, device IDs, and location data, and connects them into a unified profile. The sections below break down how that data is organized, what identifiers are stored, and what enrichment attributes can be appended.
How is data inside an identity graph organized?
Data inside an identity graph is organized as a network of nodes and edges, where each node represents an identifier or attribute and each edge represents a verified connection between them. Rather than storing data in flat rows and columns like a traditional database, an identity graph maps relationships, so every piece of information is linked to a real individual rather than to an isolated record or device.
This graph structure is what makes identity resolution possible at scale. When a person visits a website on their laptop, opens an email on their phone, and walks into a physical store, each interaction generates a different identifier. The identity graph holds all of those identifiers as separate nodes and uses probabilistic or deterministic matching logic to draw edges between them, confirming they belong to the same person.
Deterministic vs. probabilistic organization
The connections inside an identity graph are built using two complementary methods. Deterministic matching links identifiers based on confirmed, direct signals, such as a logged-in email address that appears across multiple sessions. These connections carry high confidence because the same credential was used. Probabilistic matching infers connections based on behavioral patterns, shared signals like IP addresses, and statistical likelihood. Both types of edges coexist inside the graph, typically tagged with a confidence score so downstream systems know how certain each link is.
The role of the persistent identity spine
At the center of the graph sits what is often called an identity spine, a stable, persistent identifier that anchors all the connected nodes. Individual identifiers like cookies or device IDs are temporary and change frequently, but the spine persists across those changes. This means the graph can continue to recognize a person even after they clear their cookies, switch devices, or update their email address, because the underlying relationships are preserved in the graph structure rather than dependent on any single identifier remaining constant.
What types of identifiers does an identity graph store?
An identity graph stores both online and offline identifiers, including email addresses, phone numbers, mobile device IDs, cookie IDs, IP addresses, and physical mailing addresses. These identifiers are the raw signals that the graph ingests and connects. They fall into two broad categories: authenticated identifiers that a person deliberately provides, and anonymous identifiers generated passively through digital behavior.
Authenticated identifiers
Authenticated identifiers are shared directly by the individual, usually during a login, form submission, or purchase. Common examples include:
- Email addresses (personal and professional)
- Phone numbers (mobile and landline)
- Physical mailing addresses
- Social media handles or profile URLs
Because the person actively provided these, they carry strong deterministic weight. A hashed email, for example, is one of the most reliable linking keys in modern identity resolution because it is stable, personally tied, and widely used across platforms.
Anonymous identifiers
Anonymous identifiers are generated without a person explicitly sharing their identity. They include mobile advertising IDs such as IDFAs and GAIDs, third-party cookie IDs, IP addresses, and browser fingerprints. These signals are valuable because they capture behavior from users who have not authenticated, which represents a significant share of any brand’s audience. The identity graph connects anonymous identifiers to authenticated ones wherever the data supports a confident match, extending recognition across sessions and devices.
What enrichment attributes can an identity graph append to a profile?
An identity graph can append hundreds of enrichment attributes to a profile, spanning personal demographics, professional details, lifestyle interests, and household characteristics. These attributes go beyond raw identifiers to describe who a person actually is, enabling brands to move from simple recognition to meaningful personalization. The depth of enrichment depends on the data sources connected to the graph and the quality of the matching layer beneath them.
Enrichment attributes generally fall into several categories that serve different use cases:
- Demographic attributes: Age range, gender, household income bracket, education level, and marital status
- Professional attributes: Job title, employer, industry, and seniority level
- Lifestyle and interest attributes: Hobbies, purchase intent signals, brand affinities, and content consumption patterns
- Location attributes: City, region, time zone, and geographic market segment
Each of these attribute types serves a distinct purpose. Demographic data helps with audience segmentation and campaign targeting. Professional data is particularly valuable in B2B contexts, where knowing a contact’s role and company size shapes both messaging and channel strategy. Lifestyle attributes power personalization engines by surfacing what a person actually cares about, rather than relying solely on past purchase history.
It is worth noting that enrichment attributes are not static. A well-maintained identity graph updates profiles as new signals arrive, so attributes like job title or location reflect the person’s current situation rather than data that was accurate two years ago. This freshness is critical for use cases like churn prevention, where acting on outdated information can mean reaching the wrong person at the wrong moment with the wrong message.
How FullContact helps with identity graph data
We built our Resolve platform on a true identity graph, one that has been developed over more than a decade and encompasses both online and offline data centered on real individuals. Here is what that means in practice for the businesses we work with:
- Real-time resolution: We match authenticated and anonymous identifiers in under 150 milliseconds, so recognition happens at the moment of interaction, not after the fact.
- Rich enrichment: We can append 900+ personal and professional attributes to new and existing customer records, giving teams the context they need to personalize at scale.
- Privacy-safe architecture: Our approach keeps your data secure; you access our identity graph without exposing your own customer data in the process.
- Persistent identity: Because our graph links identifiers at the individual level rather than the device level, profiles remain accurate even as cookies deprecate and device IDs change.
If you want to understand exactly what our identity graph can surface for your specific use case, feel free to contact us and we will walk you through it.