News UK turns The Times’ first-party data into synthetic audiences for advertisers 

Behind the scenes, the system, powered by Electric Twin, has been in use for months across The Times and Sunday Times, informing subscription strategy, content decisions and product development. The Times now has 659,000 digital subscribers, up 7% in the past year, per the publisher. 

“High-net-worth individuals and business decision makers tend to be really time poor,” said Caroline Tredget, commercial director, The Times and The Sunday Times. “We can emulate those using Times Exploration and do lots of questions with that panel, which makes it much easier if you’re an advertiser targeting those segments. They’re quite hard‑to‑reach people,” she said.

Six months ago, The Times pulled together its previously scattered parenting coverage into a dedicated parenting channel aimed at younger, often female, readers. Before launch, the marketing team used Electric Twin to explore what topics this audience felt were missing, how the section should be framed and named, and even how to brand it visually – effectively using synthetic audiences as a shortcut to the kind of qualitative research that would normally take weeks, while also informing product decisions. 

Meanwhile, The Times had long suspected widespread password sharing and “bonus accounts” (letting readers share with three friends and family members) were already on the roadmap, but synthetic audiences helped push it to the top of the priority list and shape the naming and comms. Rolled out in 2025 to all premium subscribers, those who use them now show higher perceived value and better retention than comparable subscribers who don’t, per the publisher.

The publisher also plans to extend the approach across other brands, including The Sun and its broadcast brands, but is first rolling it out to advertisers.

“It’s a significant cost reduction,” said Charlie Celino, commercial director of News UK. “CMOs are fighting for every single pound at the moment and trying to show ROI. To reach those high-net-worth individuals or business decision makers is really hard and expensive.” Once you’ve built a strong, reliable dataset, layering in Times ExplorAItion allows the team to expand what they can do with that audience, he added.

Celino also stressed that the learning can only be as good as the data put in. “Without humans at the root, the data wouldn’t be feasible – we’d be modeling on models on models and getting into weird algorithms. At the heart of this are the data sets we’ve collected… but fundamentally, the humans are the building blocks, the foundations that enable this product to really come to life,” he added.

Media planners are grappling with a landscape that’s splintered across platforms and formats, where the old certainties of reach and frequency no longer cut through. Andy Collins, head of planning at Wavemaker, describes a planning environment defined by fragmentation and data overload, where the problem is no longer access to numbers, but turning that firehose into something genuinely useful for clients.

At the same time, clients have grown weary of top-line audience claims. Brands increasingly expect partners to dig deeper into behaviour and attention – not just exposure, he noted. Economic volatility and budget pressure are also pushing marketers to demand proof of performance within weeks. 

“Across the board, we’re seeing clients feeling the pinch, being a little bit more cautious, a little bit more nervous,” he said. “There’s always talk of the challenge of short-termism in the industry, but it just feels like that’s ramping up a bit.”

Agencies have also experimented with their own synthetic audiences, including Wavemaker. Collins said the promise lies in turning sprawling first-party data into a living, breathing picture of how people actually behave. The attraction isn’t more dashboards, but a constantly refreshed, data-fuelled focus group that can surface patterns traditional research struggles to capture.

“Potentially you’ve got an amazing dynamic focus group [with the News UK tool] you can tap into day in, day out, that’s fueled with loads of data – and actually, how do you mine for those interesting insights, and how do you find the stuff that becomes the competitive advantage for our clients?,” he said. Instead of relying on quarterly surveys or static audience reports, synthetic audiences built on a publisher’s cross-platform data can help planners see how people move across touchpoints – from print to apps to podcasts to YouTube – and, crucially, where attention really accumulates.

But Collins is clear that none of this is a magic bullet. The usefulness of synthetic audiences depends heavily on the foundations. He warns that it’s easy to spin up shallow synthetic focus groups that say a lot but mean little. “You could go and build a focus group in five minutes based on nothing and it would spit out things, and you kind of go, what’s that worth?”

Applying LLM logic to a high-quality first-party dataset means an audience panel can now support many more research conversations without panel fatigue, stressed Gabriel Dorosz, Ad Initiative lead at INMA.

“The real question is which modeled approaches publishers can adopt credibly as platform signals and referral traffic degrade,” said Dorosz. “Synthetic research is one of the more interesting developments I’ve seen, particularly for publishers with high-quality first-party datasets. News UK is one of the most advanced publishers in the first-party data space, and what they’re doing with Electric Twin on top of their Nucleus infrastructure is a really interesting and innovative example of that.”