One question still sits uncomfortably at the centre of most conversations about data: what actually drove growth? Not what drove clicks or impressions, or even platform-reported conversions, but real and incremental business impact.
As the volume of available data has increased, the level of genuine insight has remained stubbornly static — a clear sign that the problem isn’t data itself, but the way it is analysed.
Despite the dashboards, attribution models and post-campaign reports, much of the industry is still relying on correlation dressed up as causation. In a market where every pound is scrutinised, that simply isn’t good enough.
AI is starting to change that, but not in the way many headlines suggest. Its real impact isn’t faster reporting or more sophisticated dashboards; it’s a fundamental reframing of how we understand effectiveness and who gets to deliver it.
Marrying metrics and outcomes
For a long time, marketing has optimised for what is easiest to measure rather than what matters. Platform metrics tell a neat story, but they tend to over-credit what is visible and underplay what is influential. Attribution models attempt to fill in the gaps yet still struggle to isolate true incremental impact. The result is a version of performance that feels precise but often lacks commercial truth.
AI provides the tools to interrogate performance in new ways. Rather than asking what performed through a linear, short-term lens, we can begin to assess the indirect and longer-term impact on the bottom line. That is the difference between measurement and impact, and it is where the most meaningful progress is now being made.
The real power of AI, though, is in lowering the skill threshold required to ask interesting questions. The people with the technical expertise to join together complex datasets and build data ecosystems have rarely been the same people who engage most deeply with brand and marketing issues, and the reverse has also been true. AI collapses that divide. A planner can now run the kind of multivariate analysis that once required a dedicated data science team, and an analyst can engage directly with commercial questions without being bottlenecked by their tooling. The question is, which of those two camps will encroach the most on the other’s domain?
Atoms over bits
However the convergence plays out, location will be of central importance. For all the advances in digital measurement, true integration with non-digital channels has remained a blind spot. Consumers do not experience media in neatly defined channels — they move through places, environments and moments, where context shapes how media lands and how it influences behaviour. Most measurement frameworks flatten that complexity, which is why they so often fall short.
Cookie-level data is wide but not deep. It can describe a large volume of events and patterns, but it cannot provide cross-referenced, validated insight. The explosion of digital data over the last twenty years has led us to measure almost everything while drawing relatively little genuine understanding from that wide net of metrics.
More advanced location-based planning and measurement addresses that gap. Offline media has long compared exposed and control regions to isolate the true impact of activity in a way that reflects how people actually live and behave. Now, by blending first-party data with enriched location datasets and applying robust statistical methodologies, marketers can quantify the incremental contribution of different channels and combinations far more dynamically.
AI analysis is only ever as good as the data it is fed, however. When it doesn’t have enough, it will still attempt to answer, often in wildly incorrect ways. When it is guarded within the constraints of well-defined data fields and multiple triangulated signals, it excels at the kind of pattern recognition that would previously have required highly advanced expertise. The value here is not simply in understanding where media appeared, but in identifying where it genuinely drove outcomes, a distinction that allows for a far more grounded and commercially useful view of performance.
So long, silos
This shift also challenges how we determine the success of channels. Historically, post-campaign reviews have assessed performance in silos, with only limited reference to MMM insights. Increasingly, that approach feels outdated. What matters now is how different channels work together, in specific locations, to drive incremental growth.
MMMs have long acted as the method for this kind of analysis, and when done properly, they provide meaningful information to planners. The limitations, though, are real: namely, a dependence on frequentist logic (ie, thinking that because something happened before it will happen again) and a reliance on time-consuming and expensive specialist teams. AI can handle large, complex datasets and run multivariate analysis at scale, allowing us to move beyond simple comparisons and towards a more complete understanding of system-level performance. This reflects the reality of modern media, where outcomes are rarely driven by a single channel but by the interaction between many.
Fewer reporters, more interpreters
But the most important shift that AI will usher in may be the change in measurement’s role within organisations — from a retrospective exercise used to justify decisions after the fact to an active input of decision-making itself.
AI promises a direct link between insight and action. In traditional MMM or PCA-based workflows, human interpretation and insight production have been the bottleneck. Much like a fully automated trading platform, widespread AI adoption allows increasingly complex tasks to run hands-free, with the role of analysts and data scientists shifting towards providing signals and directing the process rather than managing the minutiae of day-to-day data operations.
This is a significant reframing for an industry that has too often treated the production of statistics as the end goal. Dashboards full of numbers, significance tests and confidence intervals have stood in for insight for years, with the work of translating them into commercial action quietly pushed downstream. What a business actually needs is a clear read on what to do next. Analysts who see their role as reporting rather than translating, who stop at the table of results, will find themselves increasingly on the outside of the conversations that matter.
When connected into broader measurement ecosystems, these insights can inform both planning and in-flight optimisation — guiding budget reallocation and providing a clear rationale for scaling investment in the areas proven to drive growth. In this context, measurement becomes less about reporting and more about enabling better commercial decisions.
Keeping the seat at the table
As AI makes it easier to uncover the true impact of media, the gap between perceived performance and actual contribution will become harder to ignore. At the same time, boards are becoming more rigorous in how they evaluate marketing investment, demanding clearer, more robust evidence of return.
The implication is that planning and measurement are no longer separate phases of a campaign process but parts of a single, dynamic ecosystem. The most effective strategies will be those designed with measurement built in from the outset, so that every decision can be tested, understood and optimised against real business outcomes.
For data and insight teams, this demands a transition from purely technical skills towards more creative ones. The question shifts from ‘how do we do this?’ to ‘what can we do to make this stronger?’. The teams that don’t make that shift will find themselves quietly displaced. Planners, newly equipped to run sophisticated analysis themselves, will not wait for insight to arrive from elsewhere if they can generate it at their desks. Analytics has historically justified its seat at the table through technical gatekeeping. That gate is opening. The seat will now be kept through creative, commercially fluent thinking, or it will not be kept at all.
Ultimately, this is about confidence – not built on proxies or assumptions but on evidence. AI-driven measurement is giving marketers a clearer view of where media is working, how channels interact and what actions will drive the greatest return.
In a landscape defined by complexity and accountability, that clarity is essential.
