‘The yin and yang of magic and logic’: who really owns the data?

But shared accountability isn’t always straightforward. “This shared accountability often comes with shared confusion,” says Nives Zver, head of strategy and insights at Marble LDN. “We both have access to live data, but sometimes the interpretation is different. The facts are facts – but each party has a very different understanding of what that actually means.”

The tension highlights why agencies and brands must “read from the same manual” when interpreting insights and act on them in real time rather than retrofitting months later.

The trifecta: machine, talent, data

Success lies in balancing three forces, says Tina Allan, global chief data and intelligence officer at FCB: “I’m seeing it almost as a trifecta of machine, talent and data.” Clients are investing rapidly in all three. The agency’s role is to build trust, connect the dots and act as “a Sherpa” through the complexity.

“They need someone to guide them because it’s very complicated,” she says. That Sherpa role matters because, as Allan points out, the landscape is littered with over-purchased tech and under-used systems. Agencies must help brands map the gaps, simplify workflows and focus on outcomes, not just dashboards.

New skills, new agency power duos

If data is to be owned, it also has to be understood. So agencies are rethinking the skills they need at the client interface – and there’s a new agency power duo in town: the data analyst and the business expert.

“It’s this yin and yang of magic and logic,” says Allan. “And we’re massively investing in talent and aptitude across all of this.”

Classic account-handling roles must evolve into hybrids, adds Zver: “You’re not just replying to freelancers or pushing documents around any more. You have to use AI to do the admin work, but also understand the why behind it and specialize in something deeper.” Whether it’s CRM automation or decoding social algorithms, that’s how you future-proof your role.

And of course, in the world of generic AI, creativity matters more than ever: “Everything I do and everything technology can do will be the same for everybody,” adds Darbon. “The differentiator is how you creatively bring a brand into that infrastructure.”

Redefining measurement

Measurement, too, is caught in this culture shift. Allan notes that clients now demand more than “showing a lift; we need to understand who, the why and the where, all together, in real time.”

Darbon points to the potential of one-to-one CRM measurement powered by machine learning: “Rather than reporting on a campaign, I can start to understand how cohorts are performing and the contribution of everything we do to the overall experience.”

Zver, meanwhile, highlights the predictive gap. “We’ve added more data points and AI helps us process them faster. But we’re still not using predictive AI enough to model outcomes – we’re too often just reporting what happened.”

The key is incrementality testing. “Cost per incremental acquisition has become my safe haven,” says Ebrahim Bakhtar, senior analytics director at PMG. “Geo-lift holdouts and incrementality testing give us actual truth amid attribution models that often over-promise.”

Why the ‘why’ matters

So what’s the answer? Across the discussion, three themes stood out:

• Clarity of ownership: Brands must lead, but agencies play a crucial guiding role.

• Culture over tech: Pipelines are easy; changing mindsets takes time but creates lasting impact.

• Curiosity and caution: Data can answer many questions, but not all – so always ask why before acting.

Data ownership isn’t about who holds the dashboards. It’s about who takes responsibility for asking the right questions – and having the culture, talent and courage to act on the answers.