Marketers Rethink Measurement As AI Reshapes The Data Landscape

Artificial Intelligence

Getty Images - BlackJack3D

A shift is underway in marketing measurement, accelerated by the collapse of the deterministic model and the rise of AI-powered ad platforms. Deterministic models, built on cookies, click tracking and last-touch attribution, underpinned digital marketing for well over a decade.

Following its relative demise in recent years, these models have been replaced by fragmented signals, customer journeys that span channels that didn’t exist when measurement frameworks were originally designed, and AI platforms operate as black boxes.

Simultaneously, the stakes for marketers, particularly CMOs, to stay on top of measurement remain high as CFOs and CEOs apply budgetary pressure and need proof of returns on investment. This is an uphill battle for many marketers who don’t have the best marketing reporting tools or analytics software. According to NIQ’s 2026 CMO Outlook, 84% of CMOs cite marketing ROI as their primary metric for budget allocation. Evolving with changes in measurement and consumer behaviors is the only way for marketers to survive today’s scattered data landscape.

The issue of signal loss

It is well-established that for years, cookie depreciation, browser restrictions and tightening privacy regulations eroded the click-based attribution model marketers relied upon. Signal loss is increasing measurement complexity, as consumer purchasing journeys splinter across an ever-growing number of platforms, made even less traceable by AI.

According to a recent survey from my company, Prosper Insights & Analytics, 54.3% of AI users now use generative AI to search the internet, surpassing every other individual use case, including research, writing assistance, and personal assistance. A further 30.5% are specifically using generative AI for shopping for products and services. For marketers, this poses a formidable problem as traditional measurement tools were not built to track purchasing in this way.

Fredrik Skantze, CEO of Funnel, a marketing intelligence platform processing around 11% of global digital advertising spend, says that to make up for this signal loss, marketers need to do two things: “use all the data that’s available and use algorithms, technology, and AI.” SEO and GEO changes affect not just the advertiser, but also the ad platforms themselves, he says. “Google and Meta have less visibility over what happens after the click with the same clarity they once did.”

Accounting for measurement changes

The toolkit for measuring where marketing spend goes looks markedly different today from five years ago. The industry treated marketing mix modeling (MMM) and multi-touch attribution (MTA) as competing approaches, a debate that has largely been retired. An emerging consensus is triangulation: running MMM, MTA, and incrementality testing in parallel, with each method checking and informing the others.

Skantze explains the reasoning: “Measurement is really hard. There are so many variables, and you can’t control for all of them, but by using all the available data and applying different methods, including using a machine learning framework to triangulate between them, marketers can get the best estimate.”

MMM mechanics have also changed, where before it was a slow, consultative process that meant modeling projects were delivered months after the decisions it was meant to inform. SaaS-based tools now refresh models daily, allowing operational teams to catch underperforming channels before significant budget is wasted. Resultantly, marketing campaigns can become more attuned to consumer engagement and purchasing trends.

Executive Data Advisor, Dr Tim Wiegels, however, warns that while measurement tools have become more accessible, measuring based on inconsistent data is a “a very expensive way to get wrong answers with more confidence.” Moreover, “the companies that will actually benefit are the ones who already have clean data flows, consistent definitions, and trustworthy tracking” before bringing in MMM, attribution models and platform reporting.

Consumer behavior shift

AI visibility is now vital for marketers, with brands competing for the trust of AI agents, which have been thrust into end-to-end purchasing. According to McKinsey, AI agents will encompass anywhere between 15% and 20% of the e-commerce market by 2030, handling discovery and transactions autonomously across the retail ecosystem.

For marketers, this means agents selecting products, recommending brands, or completing a purchase on a user’s behalf. According to a recent Prosper Insights & Analytics survey, between 15% and 27% of Gen-Z, Millennials and Gen-X are using agentic AI for booking travel, buying groceries, and paying bills.

For Toby Coulthard, CPO of the AI-based platform for enterprise marketing messaging Jacquard, AI-driven purchasing isn’t creating entirely new behaviors, but reinforcing existing trends and exposing weaknesses. “Consumers,” he says, “whether interacting directly or through AI agents, have developed sophisticated filters for inauthentic communication. When an AI agent evaluates brand messaging on behalf of a consumer, it’s looking for genuine value signals, not manipulative tactics.”

At the same time, as János Moldvay, Chief Data Science Officer at Funnel, notes, “brands that aren’t parsable by AI, lacking clean structured data and metadata, will struggle. Site visits may decline, but conversion quality could rise significantly for those who get it right.” He adds that a key challenge is that most of today’s measurement frameworks weren’t built for a world where an agent completes the journey without a human click.

Although traditional search will not disappear overnight, despite the balance shifting, consumer journeys are becoming shorter, faster, and harder to follow, making accurate measurement a crucial advantage for marketers and brands looking to retain their customers.

Data foundations and the significance of governance

Ultimately, measurement tools that proactively work across measurement infrastructure, calibrating between MTA, MMM, and platform data simultaneously, without requiring a data scientist to interpret the outputs, are required for blended human and AI consumer journeys.

There’s a real risk in this transition, though. AI systems make more decisions, automate more workflows, and move at speeds no human can match, the stakes for trusted measurement get higher. Without clear governance and continuous, ideally automated validation, marketers risk mistaking AI confidence for truth, letting automation run unchecked while losing sight of what’s actually driving results.

Solid data foundations, unified data, transparent measurement, and first-party signals are essential for marketers to navigate more advanced systems.

Disclosure: The consumer sentiment study referenced above was conducted by my company, Prosper Insights & Analytics. This is the same dataset used by the National Retail Federation, and available from Amazon Web Services, Bloomberg, and the London Stock Exchange Group for economic benchmarking.