Audience source: contacts, leads and profiles

What it does

Real-time journeys can target three audience tables: Contact, Lead and unified customer profile. Which one a segment or journey uses determines what you can query, what you can send to, and which other segments you can pair it with. It is the first decision on any implementation and the one most likely to be made by accident.

Key facts

Audience configuration lives at Settings > Customer engagement > Audience configuration. It's environment-level, not per journey.
Contact and Lead arrive with default recipient fields already set for email and phone. Those values pre-populate the Send to field on message tiles.
The customer profile table has no defaults. You must nominate the email and phone attributes yourself before anything can be sent to profiles.
You can change the default recipient field per channel and add additional recipient fields. Anything you add there becomes selectable on email and text message tiles inside a journey.
Multiple email addresses can be configured on a contact's email channel from the February 2024 release onwards. Before that, one address per contact.
Leads are first-class in real-time journeys. You can build lead segments and run lead journeys without a parent contact — the old outbound requirement to have a contact behind every lead is gone.
Segment pairing is constrained by audience table. A journey entering on contacts only offers contact-based segments for exclusion; lead segments won't be listed.
Consent is evaluated on contact point consent for the recipient attribute in play, using compliance profile, purpose, optional topic and channel — the same evaluation the segment builder's consent criteria use.
Outbound marketing, which had its own contact-centric audience model and marketing lists, reached end of support on 30 June 2025 and the module was removed from all environments in May 2026. There's no dual-model decision left to make.
Sales-side marketing lists still exist as a Dataverse concept and can be referenced from real-time dynamic segments.

When to use / skip

Contact is the default and covers most B2C and much of B2B. Use Lead when the marketing motion genuinely runs on unqualified interest — top-of-funnel nurture, event follow-up, anything preceding qualification — and the client accepts that lead and contact are separate populations that will overlap. Use unified profile only where Customer Insights – Data is already part of the estate.

The judgment call worth making early: a lot of B2B clients ask for lead journeys because their sales process is lead-centric, then discover their marketing questions ("who's already a customer?") only answer on the contact side. If the interesting attributes live on account and contact, run contact-based journeys and treat leads as an entry point rather than an audience.

Mixing all three is possible and rarely wise. Each additional audience table doubles the suppression segments, doubles the consent design and doubles the reporting.

Configuration decisions

Which audience tables are in scope. Fewer is better, and this is hard to unpick once journeys and segments are built against it.
The default recipient field per channel per audience table, and whether marketers may override it on a tile.
For contacts, whether multiple email addresses are configured — for example a work and a personal address — and which is the default.
For profiles, which unified attributes carry email and phone.
Whether each suppression and exclusion segment needs a contact version and a lead version, given the picker filters by entry audience.
How a lead becoming a contact is handled in-journey: exit criteria, a separate qualification journey, or nothing.

Gotchas

Audience configuration is environment-wide. A change made to fix one journey applies to all of them, and there's no per-journey override of the defaults.
Adding a second recipient field to a channel makes it selectable on every tile, which means a marketer can pick the wrong address without anyone reviewing it.
Lead and contact records for the same human are independent audiences. Someone will receive the same campaign twice and it will be logged as a bug.
Exclusion segments not appearing in the picker is almost always an audience-table mismatch, not a permissions problem. People spend hours on this.
Consent is held against the contact point, so an address that appears on both a lead and a contact record inherits whatever consent exists for that point — which may surprise a client expecting record-level control.
Changing the audience table of an established journey isn't a setting; it's a rebuild. Decide it in design, not in UAT.

Consultant notes

Force the audience decision in the first design workshop and record it, per journey type. Everything downstream — segments, suppression, consent, reporting — hangs off it.
If the client is coming from outbound marketing, say plainly that outbound is gone and that lead targeting no longer needs a shadow contact. It usually removes a chunk of legacy data plumbing from the plan.
Build the contact and lead versions of core suppression segments at the same time, or you'll ship a journey with no exclusions applied because the segment wasn't visible.
Check the recipient field mapping against real data before go-live, particularly where the client has custom email columns. The defaults are only defaults.
Where deduplication between leads and contacts is poor, raise it as a data quality workstream early. It's not a Journeys problem and it will be blamed on Journeys.

Worth another look if per-journey audience configuration ever arrives, or if lead and contact consent handling changes.