Identity Graph

The SegmentStream Identity Graph is a core part of our attribution infrastructure. It connects fragmented user sessions across browsers, devices, and channels to enable more accurate multi-touch attribution (MTA) and a unified view of the customer journey.

By using a combination of identifiers collected on-site, the identity graph helps SegmentStream associate sessions with users across devices — even when they don’t log in — improving attribution, optimisation, and decision-making.

Why It Matters

Modern users interact with brands across multiple channels and devices. A typical journey may include:

Subscribing to a newsletter on mobile

Clicking a retargeting ad on desktop

Converting later on tablet

Without stitching these sessions together, each would be seen as a separate user, breaking attribution.

The identity graph enables SegmentStream to connect these interactions, reconstruct the full path to conversion, and attribute value accurately.

What Data Is Used to Stitch Sessions

SegmentStream combines the following identifiers to build and enrich the identity graph:

Identifier	Description
`user_id`	A persistent user identifier (e.g. when logged in) — the strongest signal.
`ip_address`	Used as a fallback when stronger IDs are missing; often combined with timestamp/location heuristics.
`click_id`	Unique ad click identifiers (e.g. `gclid`, `fbclid`, etc.) used to match sessions to ad interactions.
`email hash (ehash)`	A SHA-256 hashed version of the user's email collected on forms (e.g. newsletter signup, checkout). Enables matching to email campaigns and cross-device interactions.

💡 SegmentStream uses a confidence-weighted merging logic to stitch sessions based on multiple identifiers.

Use Cases

Cross-device attribution

Match a subscription on mobile to a purchase on desktop via ehash.

Campaign path stitching

Connect Meta, Google, and direct sessions leading to a conversion.

Better ML input

Feed complete user journeys into conversion modelling, improving performance.

Best Practices for Implementation

To get the most from the identity graph, implement the following:

Send user_id to your GA4 whenever a user logs in or authenticates: Additional GA4 settings

Preserve click_id parameters when redirecting from ad platforms by implementing: SegmentStream SDK

Implement ehash by following the guide: Cross-Device Tracking via Email Hash:

Capture and hash emails from forms (e.g. newsletter, checkout)
Add ehash to email links (e.g. https://site.com?ehash=...)