How Identity Graph works

    Identity Graph

    See every customer as one person, across every screen, every channel, every visit. No more fragmented profiles. No more guessing who is who.

    Hash-firstNo raw PII stored
    Consent-awareEnforced at push time
    AuditableFull lineage chain
    MATCHINGCROSS-DEVICECONSENTSEGMENTSACTIVATION

    Most brands cannot answer a simple question: how many customers do we actually have?

    Where it breaks today
    • The same person counted three times across CRM, email and web
    • Anonymous traffic disconnected from known customers
    • Consent tracked in one system, activation happening in another
    • No way to prove which record fed which audience
    What changes with Identity Graph
    • One consent-aware profile per person, across every screen
    • Every link carries a confidence score you control
    • Hashed identifiers only, so no raw personal data at rest
    • Full lineage from signal to segment to platform push
    See it working

    Every touchpoint resolves into one person

    Email, device, CRM record, panel response and pixel event collapse into a single profile you can act on.

    • Deterministic and probabilistic matching, with the method shown
    • Consent state travels with the profile into every activation
    • Hashed identifiers only, never raw personal data at rest
    Identity Graph / Resolution
    Email hashDeviceCRM idPixelPanelConsent: on
    Interface shown for illustration
    What you get

    Six things your team stops doing by hand

    Each capability ships enabled. Nothing here is a services engagement.

    YOUR KNOWN CUSTOMERS, CONNECTED

    CRM, email and loyalty resolve to one person.

    • Duplicate profiles collapse into a single customer record
    • One customer count every team can agree on

    ANONYMOUS VISITORS, IDENTIFIED

    Connect the dots after the login ends.

    • Anonymous sessions link back to known profiles
    • Every link shows its confidence score and method

    PRIVACY RULES, BUILT IN

    Consent enforced at the identity level.

    • Regional consent rules applied before any push
    • Withdrawal propagates across every linked record

    EVERY SCREEN, ONE PERSON

    Mobile, desktop, CTV and in-store.

    • Cross-device interactions resolve to one identity
    • Household-level disambiguation on shared devices

    AUDIENCES READY TO ACTIVATE

    From identities to campaigns in clicks.

    • Behavioural, demographic and intent segments
    • Segments recalculate as new signals arrive

    ALWAYS FRESH, NEVER STALE

    Confidence managed over time.

    • Inactive profiles flagged, archived or refreshed
    • Retention policies you configure per market
    How it works

    Signal to activation in six governed steps

    Every step is logged, so you can show an auditor exactly how an audience was built.

    Signal Capture
    Matching
    Linking
    Consent Check
    Segment Build
    Activation
    Match MethodHash-first, no raw PII///Devices SupportedAll major///Consent SyncReal-time///Activation DestinationsMajor ad platforms
    Where teams use it

    Four jobs it pays for in the first quarter

    Cross-channel personalization

    One message per person across email, display, social and CTV instead of four disconnected conversations.

    Lookalike expansion

    Cleaner seed lists mean higher platform match rates and better prospecting performance.

    Suppression and frequency capping

    Stop paying to reach the same person twice on two devices.

    Churn prediction

    Spot the drop in engagement across touchpoints and trigger retention before the customer lapses.

    Questions buyers ask

    Do you store customer emails or phone numbers?

    No. Identifiers are lowercased, trimmed and SHA-256 hashed before they reach the graph. Raw personal data is never written to an identity table, and every row carries its consent flags.

    How do we know a match is real?

    Every link carries a confidence score and the signals behind it. You set the threshold at which a match is allowed to feed activation, and weaker links stay visible but unused.

    What happens when someone withdraws consent?

    Withdrawal propagates across every linked record and blocks that person from any future audience push. Revocation runs on a schedule and is recorded in the audit log.

    Can we push these audiences to ad platforms?

    Yes. Approved segments translate into platform specs for Meta, Google, TikTok and LinkedIn. Only identities with activation consent are eligible, and lookalike seeds need their own consent flag.

    See your own duplicate rate in the first session

    Bring a sample file and a platform connection. We show you how many profiles collapse into one person, and what that changes in your next campaign.

    BOOK A DEMO SEE PRICING