Incrementality at Scale: Closing the Loop Without Moving Data
Incrementality has become a board-level priority. It is now even more important for those signing off on marketing budgets to understand what outcomes advertising actually drives, not just where it reaches. For CMOs, this presents a significant challenge, as most measurement methods aren’t suited to this.
Recent research from the Advertising Research Foundation (ARF) found that advanced incrementality measurement remains widely underused. While established methods like marketing mix modelling (MMM) and multi-touch attribution (MTA) have some value, they are far from perfect. MMM typically requires months of data to produce reliable outputs, while MTA is increasingly difficult to implement due to stricter privacy legislation and greater fragmentation.
Most significantly, these techniques either tell marketers what contributed to outcomes or which channels and touchpoints got credit for conversions. One of the most important questions – whether a specific campaign caused a specific person to buy – remains unanswered.
The data broadcasters already hold
Broadcasters and publishers hold verified exposure data about who saw which ad and on which platform. Retailers and D2C brands hold verified transaction data on what was bought and by whom. Brought together, those datasets allow you to compare customers who were exposed to a campaign against those who weren't, and measure the difference in purchase behavior directly. This is genuine incrementality measurement, telling you what actually drove outcomes, not modelling what probably did.
The problem, historically, has been getting those datasets to work together. Data sharing agreements take time to negotiate, and moving data between organizations creates legal and regulatory risk. Moreover, this requirement to share data goes against the grain of consumer privacy and the protection of commercially sensitive data. For businesses in regulated industries, this risk is especially acute. These practical barriers have meant that most brands continue to rely on aggregated outputs and slow feedback loops.
Keeping data where it belongs
Decentralized data collaboration changes this situation. Rather than bringing datasets together in a central location, decentralization allows each party to keep its data within its own cloud environment. Analysis runs across those separate environments without any raw data being moved. The broadcaster's exposure data stays with the broadcaster. The retailer's / D2C brand’s transaction data stays with them. The insights move, but the data does not.
This approach means a brand can compare exposed versus control groups against real purchase data without any party having to share, move, or centralize data. And this model has already shown results. In the UK, broadcasters work with retail partners to prove sales lift for advertisers. A campaign runs on a broadcaster's platform, exposure data is matched against checkout data, and the result shows whether viewers were more likely to buy the product than those who weren't exposed. Deliveroo and Channel 4 used the same approach and attributed a 20% lift in app signups directly to an advanced TV campaign.
The challenge of doing this at scale
Running this kind of analysis with one broadcaster and one retailer is achievable. But scaling it across multiple partners simultaneously is much harder. Each new data partnership has traditionally required its own legal framework and data sharing agreement, with a separate technical integration on top. A brand trying to build a cross-channel view of incremental impact would need to manage dozens of these in parallel; the cost and time involved make it impractical for most.
There's also a consistency issue to solve. If the methodology for matching exposure to outcome data differs between partners, the results aren't comparable. Private data networks address both of these issues. A private data network is a controlled, multi-party collaboration environment where a single curator sets the rules and establishes a consistent analytical framework. An advertiser can invite its media partners to join its network rather than requiring standalone integration with each partner. What once took months to set up with a single partner can now scale across many in just days.
Better decisions through verified sales data
Media plans are still largely built on panel data. Panels tell you which audiences watch which channels and consume which content at a demographic level. While this is broadly useful, there are serious limitations. Panel data tends to over-represent audiences willing to share their viewing behavior, which skews toward the major platforms, and media budgets follow in that direction.
However, when you replace panel-based assumptions with verified, individual-level signals from media partners and retailers, the picture changes. An advertiser can transparently see which media partners it has the greatest audience overlap with before a campaign even starts to improve efficiency and reduce waste. It can learn in-flight which placements and creatives perform best, not just that campaigns reached large audiences.
Modelling first-party seeds against first-party media owner data produces better high-quality audiences that translates into better marketing performance compared with approaches that rely on stale or aggregated data. Feedback speed also increases; MMM might take months to show that a particular channel was underperforming, but an incrementality test running against live transaction data produces the same insight in weeks or even days.
The loop closes
An advertiser that can see in real time what’s driving incremental sales can reallocate budget toward what's working before a campaign ends. That reallocation informs the next planning cycle and shapes the next set of recommendations. Over time, the media plan becomes a virtuous closed-loop cycle that intelligently optimizes for what actually drives purchase behavior for the brand's specific customers, not what a panel suggests about broad audience groups.
For broadcasters, this represents a direct commercial opportunity. Surfacing verified audience data into a federated measurement framework gives advertisers a reason to invest more confidently in premium video allowing broadcasters to protect and grow revenue streams. The incrementality enables marketers to accurately make faster, evidence-based decisions to more effectively drive growth.
The technology exists today, and it’s empowering CMOs to prove campaign performance in a world where every advertising dollar is under scrutiny. Perhaps most importantly, incrementality underpinned by first-party data collaboration is allowing brands and broadcasters to measure without ever sharing or centralising data. A win for consumer privacy and a win for advertisers looking for faster insights and safer measurement.
[Editor's note: This is a contributed article from InfoSum. Streaming Media accepts vendor bylines based solely on their value to our readers.]
Related Articles
How are streaming content companies doing? Publishers, creators, and aggregators have traditionally made the majority of their revenue on subscriptions, so why does everyone want to talk about advertising? Ad-supported content is more affordable. Because most streaming consumers have maxed out their subscriptions, leaving the SVOD market saturated, advertising is more likely to be the dominant incremental revenue driver for streaming over the next several years, based on how media companies are talking about themselves and guiding investors.
26 Mar 2026
Today's audiences have made their expectations clear: irrelevant ads are no longer tolerated. They're skipped, ignored, or avoided altogether through paid, ad-free subscriptions. As more viewers opt for environments where advertising can be bypassed, and audiences are more in control than ever, advertisers must ensure their messages are timely, targeted, and additive to the experience—not a disruption to it.
12 Aug 2025
TV is being digitized at an interesting time on the digital advertising timeline. Just as advertisers are able to target audiences on the big screen, their favorite approach - third party cookies - are going away. This is inconvenient, but David Naffis, GM of CTV at Kargo, argues that it's also a catalyst to force advertisers to test and perfect other targeting methodologies on CTV. One of the most promising is contextual targeting.
19 Mar 2024