Subscribe to our newsletter
Get the latest insights on identity, data, and audience activation.
Most businesses already know a lot about their customers. Sales and service history lives in a CRM or DMS, the website records sessions, and the email platform tracks clicks. Each system identifies people its own way, so the same household can appear as a customer record, a string of anonymous sessions, and an email address, with nothing connecting them.
An identity graph is what connects them. It's also where every identity resolution platform starts, which makes it worth understanding before evaluating one.
What Is an Identity Graph?
An identity graph is the infrastructure that links the identifiers a business sees for the same household or person, along with the relationships between them. Those identifiers include email addresses, phone numbers, devices, browsers, IP addresses, and customer records.
Identity resolution is the process of applying the graph to decide which activity belongs together. When you're evaluating vendors, the useful questions are what the graph connects, how those connections are made, and what level the platform resolves to.
What Is an Example of an Identity Graph?
Take one household shopping for its next vehicle over a few weeks. Across the dealership’s website, email platform, and DMS, its activity looks like this:
- A customer record from a past service visit, with an email address
- A click on a service reminder sent to that email, made on a phone
- Inventory browsing on that phone over the household's home network
- More browsing from a laptop on the same network
An identity graph stores each of these identifiers: the email address, the phone, the laptop, the home network, and the customer record. It also stores the links between them. The email click connects the email address to the phone. The shared network connects the phone to the laptop. The customer record connects the email address to a known customer.
Read together, those links describe a past service customer who's shopping again. Read separately, they're a service record, two anonymous visitors, and an email click.
How Is an Identity Graph Built?
An identity graph is built from observations. Each time two identifiers show up together, the graph records the connection. The signals that feed it typically include:
- Website behavior, such as pages viewed, return visits, and session patterns
- Device and browser information
- IP and network data
- Cookies and other first-party site identifiers
- Email engagement, including tracked clicks
- Known identifiers a shopper provides through a form, login, or call
Some connections are direct. A login, a form submission, or a tracked email click ties a device to a known identifier with high confidence, which is why click-based and visit-based resolution produce different levels of certainty. Most website visits come without a direct identifier, so much of an identity graph's value comes from indirect signals: devices, networks, and browsing patterns that keep appearing together.
How Does an Identity Graph Relate to Identity Resolution?
An identity graph is one of three layers, and each answers a different question.
Identity resolution uses deterministic matching when a direct identifier is available and probabilistic matching when it isn't, evaluating multiple signals and relationships to decide whether activity belongs together. Resolution is also where quality gets decided, because it sets how much confidence a connection needs before it counts. That's the balance between accuracy and reach.
What Can You Do With an Identity Graph?
An identity graph lets marketing systems treat scattered activity as one household or customer. That makes a few things possible:
- Consistent experiences across devices and channels. A shopper who browses on a phone and comes back on a laptop sees content that reflects what they've already looked at.
- Audiences built on behavior. Connected activity can be grouped by interest and stage of the buying journey, such as households comparing SUVs or households that keep returning to the same model.
- Recognition of existing customers. When anonymous activity connects to a customer record, a past buyer or service customer who's shopping again shows up as a returning customer.
- Less repetitive outreach. Recognizing the same household across channels keeps it from getting the same message three different ways.
- Clearer measurement. Connecting visits shows how interest built before a conversion and which channels were part of that journey.
How Launch Labs Uses Identity Graphs
All Launch Labs products are built on the Launch Labs identity graph. The products use its connections to resolve anonymous website activity, build and activate first-party audiences, and tie verified households to confirmed sales and service.
Resolution settings are validated against known identities, so reach and accuracy can be balanced to the goal, whether that's broad audience insight or precise retargeting. The resulting audiences can be activated through the advertising platforms marketers and their agencies already use, including Google, Meta, and TikTok.
Book a demo to see how much more of your website traffic you could identify and reach.
Frequently asked questions
What's the difference between an identity graph and a CDP?
A CDP and an identity graph do different jobs. A customer data platform manages customer data: it collects records from sources like a CRM, email platform, and website, organizes them into profiles, and sends that data to tools like ad platforms and email systems. An identity graph is a connection layer that records how identifiers relate to one another.
Many CDPs match people on their own when someone fills out a form, logs in, or clicks a tracked email, which works well for known customers. Most website visitors never take those actions. An identity graph that connects devices, networks, and browsing patterns can recognize more of them before their activity reaches the CDP. The CDP remains where customer data is managed and activated, and the identity graph extends how many of the people behind that data it can recognize.
Does an identity graph identify individual people?
An identity graph connects identifiers. Whether that leads to an individual or a household depends on how a platform applies it during identity resolution. Some platforms resolve to individuals, others to households.
In many consumer categories, vehicles included, the household is the more practical unit, because purchase decisions often involve several people and devices. Launch Labs works at the household or individual level depending on the product and use case.
How accurate is an identity graph?
It depends on how connections are made and how much confidence a platform requires before treating activity as related. Direct identifiers like logins and tracked email clicks produce high-confidence matches, while connections based on devices, networks, and browsing patterns involve more judgment. Platforms that require more confidence reach fewer people more precisely, and those that require less reach more people with more uncertainty. That's the balance between accuracy and reach every platform has to strike. The best way to judge accuracy is to test results against identities you already know.
Are identity graphs privacy compliant?
They can be, but it depends on how the data is collected, stored, and used. Questions worth asking a vendor include where identifiers come from, whether personal data is hashed or encrypted, how consumer opt-outs are handled, and which privacy regulations the platform is built to meet.
