Last year, international food giant Kellanova (formerly part of Kellogg) aimed to boost sales of its Special K cereal in the UK amidst growing competition. Their hypothesis: target customers would purchase more Special K if the messaging focused on quality and value.
To test this, Kellanova turned to data clean rooms — a method of deep data analysis using advanced technology and artificial intelligence. The results were striking: sales increased 9% among price-sensitive shoppers and 36% among premium buyers. Encouraged by this success, Kellanova is now applying the technology across broader areas of its marketing strategy.
I spoke with Louise Cotterill, Global Senior Director of Insights and Intelligence at Kellanova, and Charisse Hughes, Chief Growth Officer, about how they’ve leveraged clean room technology and how it’s shaping future campaigns.
(This interview was edited for length and clarity and originally featured in the Forbes CMO newsletter.)
What challenges has Kellanova faced in accessing customer insights?
Cotterill: Like all consumer packaged goods (CPG) companies, we lack direct last-mile transaction data — so often, we are blind to who actually buys our products. Aggregate data gives us trends, but not granular insights. We wanted to better understand how consumers interact with our brands, products, and retail partners.
Clean rooms allowed us to do this. We licence data from partners — whether loyalty cards, Circana data, or retailers — and anonymise it using a hashing process. We can then layer this with location and lifestyle data, without knowing personal identities.
This gives us richer insights while respecting privacy. We’ve moved from analysing sample panels of a few thousand to millions of records, processed through AI to help us engage more meaningfully with consumers and drive commercial performance.
Why did you decide to implement clean room technology?
Cotterill: The key is to start with a business problem, not the technology itself. My role focuses on enhancing marketing through advanced analytics. Media spend is our largest investment — even a 1–2% gain in media efficiency significantly impacts our bottom line.
Clean rooms offered a privacy-first solution — we can anonymise data to the highest standards and use it meaningfully. But we didn’t start perfectly. It wasn’t a plug-and-play solution. Leadership believed in the potential, and we kept refining it.
Now, we can measure impact and embed these processes into our AI and data science programmes — something repeatable and scalable.
Is this a permanent platform for Kellanova? How is it structured?
Cotterill: The clean room is managed by a neutral third party, ensuring anonymisation and data protection. External data sets are combined within this analytical environment, which our in-house data scientists can access.
Crucially, data itself won’t be the differentiator — models will. Everyone has access to similar third-party data. The real value comes from how we build models to identify and engage with high-value consumers.
Our internal teams build these models in close collaboration with marketing, media, and creative teams — which was key to the Special K success. This collaborative approach ensures we control audience building while working with partners to bring campaigns to life.
Was there internal resistance to such a tech-heavy initiative?
Cotterill: Because we were solving a clear business problem, buy-in was easier than simply chasing a new shiny tech trend. Still, we didn’t succeed fully at first — we tried to do too much, which made results hard to interpret.
The takeaway: keep it simple and align expectations across the organisation. Set clear success metrics. In Special K’s case, we aimed to drive sales — and exceeded expectations:
+9% in price-sensitive shoppers, +36% in premium buyers.
Hughes: This kind of initiative does require cultural adaptation. You're sometimes targeting smaller, more valuable segments, which can increase costs (e.g. CPMs), but you ultimately achieve better ROI.
We adopted a test-and-learn mindset — with quick, inexpensive failures helping us improve. Building cross-functional trust and alignment is critical. We ensured leadership was involved, including the CEO and CFO, to support this journey.
What lessons have you learned from using clean rooms?
Cotterill: Initially, we tried to do too much — the "kid in a sweet shop" effect. We now understand the importance of simplifying tests and managing organisational readiness.
It’s also vital to align upfront on what success looks like. With clean rooms, for the first time as CPG marketers, we can link specific ad exposure to real sales outcomes — a huge shift.
We’re now embedding this capability across many campaigns and markets, and every iteration of the models makes them smarter and more impactful.
Hughes: It’s about creating organisational buy-in. Different teams see the problem from different perspectives. Getting everyone aligned on a single hypothesis was crucial.
How does this compare to traditional measurement approaches?
Cotterill: The granularity is unprecedented. Panels are useful, but many marketing insights happen at the margins — which are often lost in traditional panel data. Clean rooms give us visibility into things like geographical sales patterns or nuanced consumer behaviours.
Critically, we can now prove marketing impact with real data, not just modelled projections. That’s transformative — and it’s hard to go back once you’ve seen this level of precision.
Hughes: This gives us the ability to demonstrate both sales and brand impact — crucial in justifying marketing investment. We now have the data to prove marketing’s contribution to business growth.
What’s next for clean rooms at Kellanova?
Cotterill: We’re rolling it out across many campaigns and markets. Beyond targeting and measurement, there’s potential to use it for breaking down organisational silos — supporting sales, pricing, and innovation.
The key is not to get too complex too quickly. For now, we’re embedding the basics deeply — but we see tremendous potential to expand its use.
Any advice for marketers considering clean room technology?
Cotterill: Make sure you start with a clear business brief. Clean rooms are powerful but should be used to solve real problems — not as an isolated tech experiment.
When properly integrated, they offer more control and precision than other tools. But without focus, they can easily become underused investments.
