One shift I’m watching in data collaboration: measurement is moving upstream.

When brands, publishers, and agencies collaborate on audiences, the question “how will we know this worked?” is increasingly part of the plan from the start. Planning, activation, and measurement begin to look like one connected workflow, designed together and owned together.

That’s my perspective, and it’s a pattern you can see in how collaboration platforms are being built and in how the industry is talking about clean rooms. Here’s what it looks like, what it asks of us, and three questions worth asking before your next collaborative campaign.

Adobe’s documentation for Real-Time CDP Collaboration describes three use cases. Discover finds target audiences, including the overlap between your audiences and a collaborator’s. Activate sends audiences to destinations for campaigns. Measure combines customer engagement from sources such as web impressions and shows how a campaign performed.

Listed that way, they read as a sequence. In my view, they work best as one design. The audience you discover shapes what you activate, and what you plan to measure shapes which audience is worth discovering in the first place. It’s the same idea I explored in From Knowing WHO, To Understanding HOW: an audience becomes more useful when you know what you want to learn from it.

One connected flowDiscover, activate, and measure work best when they are designed together.BrandPublisherAgencyPlanActivateMeasureBrands, publishers, and agencies plan, activate, and measure in one connected flow.DATA ON TREND

In April 2026, Adobe announced a set of updates to Real-Time CDP Collaboration. Several of them point at measurement:

  • Outcomes-based measurement workflows are generally available. Adobe describes overlaying publisher ad exposure signals with conversion signals to produce attribution insights, and monitoring campaigns while they run.
  • Agencies can use Collaboration on behalf of their clients for planning, activation, and measurement.
  • Invitations let a brand, publisher, agency, or data partner join a collaboration without a license of its own.
  • Amazon Marketing Cloud, powered by AWS Clean Rooms, connects directly from Collaboration.

The documentation on onboarding measurement data shows what setup involves. Conversion data is the supported measurement type today. You map fields to the match keys configured for your collaborator account, set a deduplication key so one conversion is counted once, and can configure up to three conversion events with up to five conditions each. Data is filtered for consent before upload. Once it’s processed, both the publisher and the advertiser can view campaign measurement reports, and the advertiser can create attribution reports.

Each of those settings is a decision made before launch. That’s what I mean by upstream. Availability varies by region, so check Adobe’s documentation for where Collaboration is offered today.

In September 2026, IAB published Data Clean Rooms in Commerce: Common Use Cases, a framework for deciding when a clean room is worth using. Its first screening question is “What business question are we trying to answer?” From there, it separates use cases a clean room typically enables, such as audience overlap and data enrichment, from use cases a clean room enhances. Attribution, incrementality, and reach and frequency sit in the second group. IAB notes they can often be performed with existing measurement, analytics, or reporting tools when the data resides in a single environment, and that a clean room creates additional value when ad exposure and purchase outcomes sit with different organizations.

I read that as another argument for designing measurement first. Start with the business question and where the data lives, and the right tool tends to follow.

When collaboration adds the most valueONE ENVIRONMENTExposure and outcomessit with one organizationExisting measurement toolsoften cover attribution and reach.ACROSS ORGANIZATIONSExposure and outcomes sitwith different organizationsCollaboration or a clean roomadds value by closing the loop.Based on IAB, Data Clean Rooms in Commerce: Common Use Cases (September 2026).DATA ON TREND
Before the audience is built1Which outcome arewe measuring?Exposure, conversion, orboth. Define what countsas one conversion.2Whose data closesthe loop?Bring that partner intothe plan from day one,with match keys and consent.3Who owns the read-out, and when?Name the owner, who seesthe report, and the dates:in flight and at close.Three questions for brands, publishers, and agencies to answer together.DATA ON TREND

Which outcome are we measuring? Exposure, conversion, or both. Agree on what counts as one conversion before the audience is built. In Collaboration, that’s your conversion event definitions and deduplication key. In any stack, it’s the shared definition each partner reports against.

Whose data closes the loop? Identify which partner holds exposure data and which holds outcomes. When they sit with different organizations, collaboration earns its place, and that partner belongs in the plan from day one, including agreement on match keys and consent.

Who owns the read-out, and when? Decide who builds the report, who sees it, and when you’ll look: in flight, at close, or both. Putting the read-out on the calendar early turns it into a shared milestone.

Measurement joins the planPlanMeasureActivatefrom day oneThe read-out is defined alongside the audience, then the campaign activates.DATA ON TREND

The clearest example of upstream measurement I know is a holdout. In a past role, I ran Adobe Target and Journey Optimizer side by side with a universal person-level holdout, which gave us a consistent baseline across channels. That baseline existed because it was designed before activation. The same thinking applies when partners collaborate: decide the comparison first, and the results have something to stand on.

I frame data work as a Virtuous Cycle: Learn → Think → Make → Measure. I wrote about it most recently in Context Is Everything. Upstream measurement is what keeps the cycle turning. What you measure in one campaign becomes what you learn going into the next, including which audiences, partners, and outcomes deserve another round.

  1. Pick one upcoming collaborative campaign. Write down the outcome, the conversion definition, and the deduplication rule before the audience is built.
  2. Map where exposure and outcome data live. One organization or several? That answer shapes whether collaboration or a clean room adds value.
  3. Put the read-out on the calendar. Name the owner, who sees the report, and the dates, in flight and at close.

Measurement planned up front makes the conversation after the campaign far more useful, for each partner in it.

Sources: Adobe, “Adobe expands offerings and partnerships for data collaboration and paid media use cases,” April 20, 2026. Adobe Experience League: Real-Time CDP Collaboration use cases; onboard measurement data. IAB, “Data Clean Rooms in Commerce: Common Use Cases,” September 2026.


If you have questions, or just want to talk analytics, data strategy, and anything in-between, reach out to me at Katie@DataOnTrend.com.

Subscribe

Sign up for a weekly newsletter with the latest blog posts and exclusive content!

By clicking this button you consent to receive emails from Data on Trend

Empowering Businesses with Data-Driven Insights. At Data on Trend, we help businesses unlock the full potential of Adobe Analytics, Adobe Customer Journey Analytics, Adobe Target, Adobe Real-Time CDP, and Adobe Experience Platform. Our expert insights, strategic frameworks, and hands-on guidance empower organizations to optimize marketing performance, enhance customer experiences, and drive operational efficiency. Explore our latest blog posts on marketing technology, data strategy, and industry trends to stay ahead in the ever-evolving digital landscape. Get in Touch: Have questions or need expert consulting? Reach out at katie@dataontrend.com. Follow Data on Trend on Instagram, LinkedIn and Facebook for real-time updates and industry trends!

© 2024 Data on Trend. All Rights Reserved.