Skip to main content
Adrià García
← Back to the simulators

Identity Service · Customer Journey Analytics

Graph or stitching

Laura leaves a cart on her phone and buys days later from the home laptop, which Marcos also uses. Experience Platform decides who she is so it can act on her. Customer Journey Analytics decides who she was so it can measure. They do not always agree.

Scenario

The cart is anonymous on the phone. The purchase is made logged in on the laptop.

Does she also log in on the phone?

If not, the phone only links through the newsletter email.

The CRM carries Laura's email
Identity Graph Linking Rules

With CRMID as a unique namespace, one graph cannot hold two customers.

CJA package
Stitching

Persistent ID: ECID. Person ID: CRMID.

Illustrative scenario ↑ Back to the controls
Example assumptions

Example people and identifiers. Lookback by CJA package: 7, 14 or 30 days on Select, Prime and Ultimate. Graph-based stitching requires Prime or higher.

Experience Platform · Profile and AJO

To activate: identity graph and Profile

Forward-looking and near real time. Profile uses the graph to gather fragments and merges them under its merge policy. Audiences, AJO and Destinations use that merged profile.

Customer Journey Analytics · connection

To measure: the person ID on each event

After the fact and per dataset. Replay can restate history inside the lookback window.

Architecture

Where it fits in the platform

The graph says who someone is, to activate, going forward and in near real time. Stitching says who they were, to measure, after the fact and per dataset. Both should use the same person key.

Comparison between the identity graph and Customer Journey Analytics stitching, top to bottom in five stages. Data: web events from the Web SDK with the ECID, plus the CRMID on login, and a CRM in batch with CRMID and email. Data Lake: events and records land in their datasets, with the primary identity flagged as primary=true. Identity: the Data Lake passes identities to Identity Service, which links them in a graph with its Linking Rules. Separately, the event datasets reach the CJA connection, which sets one person ID per dataset. Resolution, to activate: the graph reaches Profile, which merges in real time according to its merge policy, and the merged profile goes out to AJO and Destinations to act. Audiences are evaluated on that merged profile under its merge policy, which uses the graph to gather related identities. Identity settings cannot be switched off without resetting the sandbox. Resolution, to measure, by two routes. Field-based stitching, from CJA Select, takes the persistent ID and the person ID from the same dataset. Graph-based stitching, from CJA Prime, takes the persistent ID, looks it up in the graph and gets back one namespace. It ignores timestamps and inherits graph quality. Use: both deliver one person ID per row to Workspace, for People and attribution, and replay restates history inside the lookback window. The shared key: the top-priority unique namespace in the graph should be the connection person ID. CJA audiences published to Profile send that person ID, and if no profile has that identity, a new one is created.

The recommendation

One person key, decided once, in the graph and in CJA

They are not two competing engines. The graph decides who a customer is in order to act. Stitching decides who a visitor was in order to measure. They work well when both use the same key.

  1. 01

    Linking Rules before data arrives

    The customer ID as the unique, top-priority namespace. Test it with Graph Simulation in development, because identity settings cannot be switched off without resetting the sandbox.

  2. 02

    That same namespace as person ID

    On every dataset in the CJA connection. That way they join, and the audiences CJA publishes to Profile land on profiles that already exist.

  3. 03

    Field-based by default

    It works from CJA Select if the key is on authenticated events. Graph-based, from Prime, aligns people with Profile but inherits any flaw in the graph.

  4. 04

    Watch shared devices

    Field-based splits people by time on the device. Graph-based ignores timestamps and assigns the whole device to one person.

Architecture note · 08 The person you measure is not the person you activate The identity graph decides whom to activate now. Analytics stitching rebuilds who it was. They are two answers, and each is designed on its own. Read the note →