Behavioural measures from Journeys interaction data (preview)

What it does

Interaction data produced by Customer Insights - Journeys — email opens, clicks, form submissions, event check-ins — becomes available directly in the measure and segment builders under a Behavioral tab, without building an ingestion pipeline for it. You aggregate it into a measure and then filter on that measure in segment rules.

Key facts

Preview. Not intended for production use, and the functionality may be restricted or change.
Both Customer Insights - Journeys and Customer Insights - Data must be deployed in the same environment.
Contacts or leads from Journeys must already be unified into the customer profiles, otherwise there is nothing to attach interactions to.
The interaction types appear on the Behavioral tab in both the segment builder and the measure builder.
Email interactions cover Consent not given, Blocked, Bounced, Delivered, Link Clicked, Opened and Sent.
Push notification interactions cover Link Clicked, Not Sent, Opened and Sent; text message interactions cover Bounced, Delivered, Link Clicked, Not Sent and Sent.
Event and form interactions cover Marketing Event Check-in, Event Registration Created and Canceled, Form Submitted and Form Visited.
Each interaction type carries fields such as Journey, Email, Email address, Email client, Timestamp, Profile ID and Profile type.
Aggregations offered over behavioural data are count, sum, average, min and max, with optional filters and time windows.
Processing is batch only, tied to the system refresh. There is no real-time behavioural evaluation here.
Only interactions linked to unified customer profiles are visible. Unmatched interaction rows are not reachable.
Interaction data does not appear on the customer profile timeline in Customer Insights - Data.
Initial ingestion takes from around thirty minutes to several hours on large environments.

When to use / skip

This is worth having on any project where Journeys and Data sit in the same environment and the client wants engagement-based segmentation — re-engage anyone who has not opened an email in ninety days, or promote people who clicked three times last month. Getting that data across without this feature means an export, a pipeline and a set of custom tables, which is real effort.

Skip it where the client's engagement data lives somewhere else entirely, or where they need real-time reaction to a click, because batch processing on the system refresh cadence will not deliver that. And be careful about building a client's core segmentation strategy on a preview feature — it is fine for a proof of value, harder to defend as the backbone of a production activation programme until it reaches general availability.

Configuration decisions

Whether Journeys and Data genuinely share an environment, since the prerequisite is hard rather than a preference.
Which behavioural counts become named measures versus being expressed inline as segment conditions each time.
The time windows on each behavioural measure — a rolling ninety days behaves very differently from all time, and the client will assume rolling.
How engagement measures are defined per channel, given that email, push, SMS and forms all expose different interaction sets.
Whether the refresh cadence for these measures matches the campaign cycle the marketing team works to.
Whether to depend on this in preview at all, or to model the same thing from an exported dataset until it goes GA.

Gotchas

Preview status is the main one. Do not put it in a statement of work as a committed capability without flagging it, and check the current release wave before you promise a GA date.
Only interactions tied to a unified profile show up. Poor contact or lead matching during unification quietly shrinks every behavioural measure, and the number just looks low rather than wrong.
Nothing appears on the profile timeline, so a marketer looking at an individual customer sees the aggregate measure but no supporting detail. Expect that question.
Batch processing means a behavioural segment is as fresh as the last system refresh, which is not what "opened an email" suggests to a marketer.
First ingestion can take hours on a large environment, so build the demo the day before, not during the workshop.
Journeys reports its own engagement figures over the same events, and they will not match the measure exactly because of matching, time windows and refresh timing. Get ahead of that conversation.

Consultant notes

Check unification match rates for Journeys contacts and leads before building anything behavioural. If matching is weak, fix that first — every measure here inherits it.
Demo one behavioural measure feeding one segment end to end. The link from "opened an email" to "in this audience" is what sells it.
Be explicit with the client that this is preview, in writing, and agree what the fallback is if it changes.
Set expectations on freshness in the first workshop. Marketers hear "behavioural" and think real time.
Before go-live, compare one behavioural measure against the equivalent figure in Journeys and document why the two differ, so nobody raises it as a defect later.

Worth another look once this leaves preview, or if real-time evaluation of interaction data arrives.