What Is an Identity Graph? A Healthcare Marketer's Guide
What an identity graph is, how it links people, devices, and households, deterministic vs probabilistic links, and its uses and limits in health.
The short answer
An identity graph is a database that links the identifiers belonging to the same person or household, such as names, postal addresses, email addresses, phone numbers, device IDs, and cookies or mobile IDs, so they can be recognized as one entity. Healthcare marketers use identity graphs to match audiences to media, cap frequency, and connect exposure to outcomes, within privacy rules. Graphs differ in coverage, accuracy, freshness, and how they source data.
Almost every modern audience, from an HCP list matched to programmatic to a DTC segment on CTV, passes through an identity graph. Most marketers never see it. Understanding what it does helps explain match rates, reach numbers, and measurement results.
The basic idea
People and households leave many identifiers:
| Identifier type | Examples |
|---|---|
| Offline | Name, postal address, phone |
| Online, personal | Hashed email, login IDs |
| Device | Mobile advertising IDs, CTV device IDs |
| Network | IP address |
| Professional | NPI for HCPs, business email |
An identity graph stores which identifiers belong together. When a brand sends a list of HCPs or customers, the graph finds the matching digital identifiers so media platforms can reach them.
Deterministic and probabilistic links
- Deterministic: based on a direct connection, such as the same hashed email used on two devices.
- Probabilistic: inferred from patterns, such as devices seen on the same network.
Most graphs use both. Deterministic links are more reliable but cover fewer people. See deterministic vs. probabilistic HCP identity.
Consumer, business, and HCP graphs
- Consumer graphs link people and households.
- Business graphs link professionals to companies, work emails, and work devices. See evaluating a business identity graph.
- HCP graphs link NPIs to professional and personal digital identifiers.
Connecting consumer and HCP graphs raises specific issues. See where consumer and HCP graphs connect.
How graphs are used in health marketing
- Onboarding. Turning an NPI list or CRM file into addressable media audiences.
- Frequency management. Recognizing the same person across devices.
- Measurement. Linking exposure to outcomes, often through de-identified tokens.
- Suppression. Excluding people who should not receive messages, such as existing patients in some programs.
What makes a graph good
| Quality | What to ask |
|---|---|
| Coverage | What share of my list can you match? |
| Accuracy | How do you validate links? Against what truth data? |
| Freshness | How often are links refreshed and stale ones removed? |
| Sourcing | Where does the data come from, and under what consent? |
| Privacy controls | How are sensitive categories handled? |
The healthcare identity graph evaluation gives a full checklist.
Limits
- Graphs are never complete. Match rates below 100% are normal.
- Higher match rates are not always better if they come from weaker links. See why HCP match rates differ.
- Data ages. People move, change jobs, and replace devices.
- Health data requires careful handling under HIPAA where applicable, FTC rules, and state laws.
Common mistakes
- Choosing a graph only on match rate.
- Not asking how links are validated.
- Treating a graph's audience as exact.
- Moving health inferences across devices without privacy review.
Practical takeaway
When a partner quotes a match rate, ask for it split by deterministic and probabilistic links and by how recently the links were verified. That breakdown tells you more about quality than the headline number.
Frequently asked questions
What does an identity graph do?
It connects different identifiers that belong to the same person or household so that data and media can be matched across channels and devices.
What is a third-party identity graph?
A graph built and licensed by a data company, as opposed to a first-party graph a brand builds from its own customer data.
Are identity graphs legal for health marketing?
They can be used within privacy law and industry codes. Health data and inferences require extra care: consent, sensitive-data rules, and state health privacy laws may apply.
Sources
- NIST Privacy Framework
- Network Advertising Initiative
- FTC, Health Breach Notification Rule: The Basics for Business
- Truthset, Data Ratings
External guidance and platform documentation change. Links were current at publication; check them again before relying on them for a decision.
Editorial note. Analysis and frameworks are the author's own and do not represent Acxiom or any current or former employer, client, or named platform. Examples labeled hypothetical or illustrative are not results from real campaigns. Nothing here is legal, regulatory, or medical advice.
New pharma programmatic breakdowns, occasionally
One email when I publish something worth reading. Benchmarks, measurement teardowns, and case studies with the caveats attached. No cadence promises, no reselling your address.
Unsubscribe any time. See the privacy policy.
Working through this decision on a real plan?
I work on health and pharma data, identity, and activation, after five years running HCP and DTC programmatic agency-side. Happy to talk through how this applies to your situation.