Best identity graph for cross-device tracking in martech

A detective trying to solve a case with three different witnesses, each describing a different person, is basically what marketers are up against every day. One customer might show up as three different visitor profiles: an anonymous session on a phone during a commute, a return visit on a work laptop, and a purchase on a tablet that night. Without a way to connect those dots across devices, your data collection treats one shopper like three strangers, and every user in your funnel starts to look like more people than actually exist.

When you can't tell that a mobile ad and a desktop visit came from the same person, you end up crediting the wrong channel, misreading your customer journey, and making budget calls based on a distorted picture of your data. Getting cross-device identity right is a business problem before it's a technical one, and understanding how identity resolution tools connect user activity is the key to fixing it before you invest.

Key takeaways

• An identity graph connects a person's activity across devices and touchpoints into a single, unified profile.
• Deterministic matching uses verified identifiers like a hashed email or login, while probabilistic matching infers a match from signals like IP addresses and device fingerprinting.
• LiveRamp, Segment, Cometly, Wunderkind, mParticle, Tealium, and AppsFlyer each solve cross-device identity differently, and the right pick depends on your use case.
• Third-party cookies and iOS restrictions have made match rates harder to maintain, even for deterministic tools.
• Privacy regulations are getting more specific about how identity data can be collected, stored, and used across devices.
• An identity graph tells you who a customer is. It doesn't tell you whether your marketing spend is actually working.
• Prescient's marketing mix modeling (MMM) can work alongside any identity graph or independently of one, depending on what you need.

What an identity graph actually does

An identity graph is a data structure that links identifiers, like a hashed email, a device ID, or a login, to build a unified profile of a single user across every device they touch. Think of it as a map connecting all the different keys a person leaves behind (a cookie here, a login there) into one identity, often built on top of graph databases designed to handle that many connections at once.

\

\

For cross-device tracking specifically, the graph's job is to recognize that the user browsing on desktop this morning is the same person who converted on their phone this afternoon. Get this right, and your customer journey reporting reflects reality across every device in your funnel. Get it wrong, and you're left with unified user profiles that are really just guesses, and a customer identity that shifts depending on which device someone happens to be holding.

Deterministic vs. probabilistic matching

Every identity graph relies on one of two approaches to connect a user's activity across devices:

Deterministic matching uses verified, first-party identifiers, like a hashed email, a phone number, or a login, as the key that confirms two sessions belong to the same person. This kind of identity matching is highly accurate, but it only works when someone has actually logged in or shared traditional identifiers with you.

Probabilistic matching looks at signals like IP addresses, browser type, and device fingerprinting to infer that a different device likely belongs to the same user. It scales further because it doesn't require a login, but this identity matching approach trades away some accuracy to get there.

Most mature martech stacks end up blending both. Deterministic data anchors the accurate core of the graph, and probabilistic data extends it into anonymous sessions that haven't converted yet. Even the same ad shown across multiple devices can look like it reached three different people without one of these approaches connecting the identity data behind it.

Top identity graph solutions to consider

Here's a rundown of the tools most commonly discussed for cross-device tracking and identity resolution, along with the specific use cases each one tends to fit best. Each vendor connects data a little differently, so the right choice depends on how your team already collects and uses customer data.

LiveRamp

How it works: Often described as the gold standard for enterprise identity resolution, LiveRamp builds an authenticated identity graph from first-party data like hashed emails and CRM records, connecting both online and offline data to a persistent identifier that follows a person across devices without leaning on third-party cookies.

Best for: Enterprise teams that need privacy-safe data collaboration across ad platforms and want the most established graph for advertising and reach.

Segment

How it works: Segment sits at the data layer of your stack, collecting user interactions and custom events from every source and resolving them into a single profile that other tools can pull from.

Best for: Engineering and data teams that need a flexible foundation to route unified customer data to hundreds of downstream marketing and analytics tools.

Cometly

How it works: Cometly uses server-side tracking to capture the customer journey across devices without relying on browser cookies, then feeds that enriched data back to ad platforms like Meta and Google.

Best for: Performance marketing teams that want cross-device attribution tied directly to conversion events, so they can optimize campaigns and refine ad targeting with cleaner data.

Wunderkind

How it works: Wunderkind builds identity profiles from behavioral signals across a large network of sites, using its own identity matching to recognize logged-in and returning visitors without depending on cookies.

Best for: Ecommerce and retail brands looking to boost logged-in identification rates for on-site personalization and more targeted email campaigns.

mParticle

How it works: Similar to Segment, mParticle connects first-party data across sources into unified profiles, with server-side collection and a particular focus on giving teams control over how identifiers are structured and governed.

Best for: Teams that want CDP-style identity resolution but need more flexibility in how customer data is modeled and audited.

Tealium

How it works: Tealium pairs tag management with a customer data platform, using server-side data collection to give you control over how identifiers are captured alongside real-time identity resolution in one system.

Best for: Organizations that want tracking implementation control bundled with cross-device profile building.

AppsFlyer

How it works: AppsFlyer specializes in mobile attribution, using deterministic identifiers when available and probabilistic matching to fill gaps, with server-side deep linking to preserve campaign context when a customer moves from one device to another.

Best for: Mobile-first or app-heavy brands that need to connect an ad click on one device to an app install or purchase on a different device.

Details you need to know

The tool comparisons above cover how each platform works, but there's a bigger picture most rundowns leave out when it comes to cross-device identity resolution. Here's where your marketing strategies could hit friction if you go into cross device tracking blind.

The real cost and timeline

Identity graph solutions rarely come with a simple, practical way to estimate what implementation actually takes. Enterprise pricing is often quote-based, and the true cost includes CRM data hygiene, engineering time to integrate identifiers, and ongoing maintenance as identifiers change or expire. Before you commit, ask for a realistic timeline, not just a features list.

Match rates keep shrinking

iOS restrictions and the decline of third-party cookies haven't stopped since these tools launched, and they aren't likely to reverse. Even deterministic matching depends on people logging in or consenting to be tracked, so as fewer people do that at the household level, the accurate core of any graph gets smaller. A vendor's device fingerprinting or probabilistic matching can help fill some of that gap, but shrinking match rates also mean shakier conversion rates in whatever marketing attribution model sits downstream. It's worth asking directly how a vendor's match rates have trended over the past year rather than taking a static accuracy claim at face value.

Privacy compliance is getting more specific

Regulations aren't just tightening in general terms anymore. Expanding definitions of third-party data sharing under state privacy laws now cover cross-context behavioral advertising directly, and other regulations abroad are introducing documentation requirements for systems that use behavioral data to make automated decisions. None of this means you should avoid identity resolution altogether, but it does mean privacy compliance needs a seat at the table when you're evaluating vendors.

A better way to evaluate vendors

Yes, you can and should ask how many devices a graph recognizes when vetting potential vendors, but don't forget to dig deeper. You'll have a smoother road ahead if you also ask:

• How is consumer data retained, and for how long?
• What happens to a user's profile when they opt out?
• How deep is the integration with the specific use cases in your stack, like your CRM or ad platforms, and can it help your marketing team target customers without duplicating work across tools?
• Can the vendor show you real match rate trends, not just a headline accuracy number?

Any analytics platform can claim accurate data. Fewer can show you the trend line behind that claim.

Where identity graphs fall short for marketing decisions

Knowing that the same user showed up on three devices isn't the same as knowing whether that combination of touchpoints was worth the spend. An identity graph can tell you a customer's identity and connect their cross-device journeys into one profile. It can't tell you if your marketing budget would have performed better shifted toward a different channel entirely.

This is where a lot of marketing teams get stuck. They invest in cross-device identity resolution expecting it to answer budget questions, and it simply wasn't built to do that. Identity graphs are a data collection and personalization layer for understanding a user's identity. They're not a measurement layer that connects spend to actual business outcomes, and users deserve a marketing team that can tell the difference.

Where Prescient comes in

Identity graphs solve a real problem: knowing that the person on mobile and the person on desktop are the same customer, through identity matching built on first-party data. But knowing who someone is across devices still leaves the bigger question unanswered, which is whether your marketing spend actually drove that customer to convert, and where your next dollar would work harder. That's a measurement question, and it takes a model built to isolate incremental impact and connect it to a clear budget decision. Prescient's marketing mix modeling (MMM) does exactly that, layering in saturation curves and confidence scores so you know not just who converted, but which spend actually caused it.

You don't have to choose one or the other. Prescient runs independently of any identity resolution tool, so if you're already using LiveRamp, Segment, or another graph for cross-device tracking and personalization, you can layer Prescient on top to connect that work to a clear view of how to shift budget for maximum efficiency. Book a demo to see how the Prescient platform helps you plan where to put your next dollar.

FAQs

What's the difference between an identity graph and a customer data platform?

An identity graph is the underlying structure that connects identifiers like hashed emails, device IDs, and logins into a single customer identity for cross-device tracking. A customer data platform, or CDP, is a broader tool that collects data from many sources, often uses an identity graph internally to power cross-device identity resolution, and then routes those unified profiles out to other marketing and analytics tools used for cross-device attribution. In practice, many CDPs and identity graphs overlap, but the graph is the connective tissue underneath the platform, and it's what allows a CDP to connect data across every device a customer uses.

Do identity graphs still work as third-party cookies disappear?

Yes, though the mix has shifted for cross-device tracking overall. Deterministic matching, built on hashed emails and logins, isn't dependent on cookies to begin with, which is part of why vendors describe it as more future proof. Probabilistic matching and server-side data collection lean more heavily on device fingerprinting and other signals to fill gaps left behind as cookies and mobile ad identifiers become less available.

How much does an identity graph solution typically cost?

Costs vary widely and most vendors quote enterprise pricing based on data volume and use case rather than publishing a flat rate. Beyond the platform fee, factor in the cost of preparing clean, well-structured first-party data, plus server-side engineering time to optimize how identifiers integrate across your CRM, website, and ad platforms.

Can small or mid-sized brands use identity graphs, or are they only for enterprise?

Some tools are built with enterprise scale in mind, but smaller brands have options too. Platforms with tiered pricing or narrower use cases, like retail-focused personalization tools or CDPs with self-serve plans, can make cross-device identity resolution accessible without the full enterprise price tag, and can still connect online data across multiple devices well enough to support basic cross-device attribution.