AI could free up $30 billion for marketers, if we solve measurement first | MarTech

Your analytics team spends hours connecting the dots between your offline and online campaigns. Your attribution approach is primarily a last-touch approach, or when it’s more sophisticated, it’s a black box that you can’t quite explain to stakeholders.

You wonder if your marketing mix model (MMM) is delivering the right recommendations. You trust your incrementality tests, but structuring and analyzing them takes a lot of effort. In the meantime, you wonder: are we investing in the right channels? Do we optimize towards what actually produces results, or only what is easy to measure?

If that sounds familiar, you’re not alone. According to the IAB’s State of Data 2026 report60%-75% of marketers say their measurement approaches fall short in coverage, consistency, timeliness and trust. Not a single respondent said their MMM includes all paid media channels. Your CTV investment? Probably underrepresented. The same goes for retail media, gaming, creator content and audio.

Here’s what happens: If you can’t easily measure a channel, you invest less in it or skip it altogether. You call it smart allocation. But in reality, measurement bias dictates your strategy.

You’re optimizing for the wrong thing

Your models likely rely on platform-level or last-minute attribution. Your dollars will continue to flow to lower-funnel channels that are easy to follow, even if you suspect they aren’t the most influential. That mid-funnel brand campaign? The podcast sponsorship? They are undervalued because your measurement cannot see them clearly.

Here’s the more complicated truth: your models are confusing correlation with causation. If a channel is present in the conversion, it doesn’t mean it caused the result. Without incrementality tests or causal frameworks, you optimize based on chance instead of contribution.

I’ve seen planning teams default to what works last quarter, not because they believe it’s right, but because the outcomes indicate so. Strategy becomes a function of what you can measure, not what the right approach should be.

Dig Deeper: Are you struggling with marketing measurement? You’re not alone.

The AI ​​opportunity you’re not ready for yet

You’ve heard the saying: AI can fix measurements. There is some truth in it. The IAB report estimates that AI-powered improvements could deliver $14.5 billion to $26.3 billion in media investments and $6.2 billion in productivity gains within two years – nearly $30 billion on the table.

But here’s the catch: AI only works if you feed it clean, standardized data. Most organizations don’t have that. Taxonomies are inconsistent and data definitions vary across platforms. Therefore, you cannot reliably link exposure to outcomes.

AI already does some data preparation work. Soon it will be about tuning models, analyzing lift tests and tuning the results of the measurement methods. However, without the right foundation, you’ll end up automating the same problems you have today.

That’s where IAB’s Project Eidos comes into the picture. The name Eidos comes from the Greek verb ‘to see’, which underlines the initiative’s goal of creating visibility and coherence in a fragmented measurement landscape. Through Project Eidos, IAB is building the foundational elements that AI needs: standardized taxonomies and classifications, a unified framework that links exposure and behavior to outcomes, and modernized specifications for MMM.

If this works, the payout is real. You can allocate budget to channels in which you have underinvested. Your team could spend almost 10% of their time on data preparation and strategy.

Dig deeper: the smarter approach to marketing measurement

Infrastructure is the bottleneck

The friction you feel is not just about technology or methodology. It’s operational. The quality of the data is inconsistent. Workflows are manual. Teams operate in silos. You’re probably using processes that were built for rigid cycles, not the fluid, fast pace your business needs today.

If your infrastructure is broken, AI will expose these problems faster and on a larger scale.

You also have legitimate concerns: legal and security risks, model accuracy, data quality. Failure to address these will make measurements harder to trust, less inclusive of all media, and slower to update. That creates a feedback loop that kills the value of AI before you can scale it.

Of the IABs surveyed, 40% of brand agency contracts already contain AI-related clauses, including transparency requirements, accountability frameworks, performance expectations and efficiency standards. Within two years that will increase to 70 to 80 percent.

You need to show not only that your models work, but also that they meet new accountability standards.

What we should actually do

Fixing measurements is not about buying another tool. It is a structural shift that requires planning, analysis, data, legal and operations to work together. This is what we need:

• Build automated, repeatable workflows to measure more often and reduce manual work.
• Improve data quality and standardize access across channels and platforms. Models need consistent input, not patchwork.
• Align teams around shared KPIs instead of disparate dashboards that fragment decision-making.
• Make measurement a tool for optimization, not just validation. Use insights to inform planning, not just report what happened.

None of this is new, but AI is now making it impossible to ignore these long-standing problems and calling for immediate solutions. Without a solid foundation, the $30 billion industrial opportunity remains out of reach.

The technology exists and initiatives like Project Eidos are starting to build the frameworks. To enable smarter budgets and massive productivity gains, we need more than just tools. We need a collective commitment to push platform partners toward these standards.

Stop patching the past. Let’s rebuild the foundation and put that $30 billion to the right places.

Dig Deeper: 5 Ways to Improve Marketing Measurement in 2026