Audio Still Struggles For Credit In Marketing Mix Models, IAB Finds.

Marketing mix models may be getting more attention than ever, but the buy-side is signaling a blunt reality for audio: the models still don’t see enough of the channel to value it properly. But a new report from the Interactive Advertising Bureau says audio isn’t unique. “Alarmingly, no media channel is fully represented in marketing mix models (MMM),” it says. 

The IAB says among MMM users who say they have “line of sight” into specific channels, audio-adjacent formats land squarely on the underrepresentation list. Four in ten (39%) say podcasts are underrepresented in models, and more than a third (36%) say streaming audio isn’t getting a fair shake. The problem is even more pronounced for broadcast as 46% of MMM users say AM/FM radio, along with several other traditional media, also don’t get what they should in ad models.

The report argues the consequence is not academic — it changes where dollars go and the return on investment. “The result is a myopic view of performance that reduces confidence in MMM outputs and reinforces bias toward more measurable channels (display, search, social) rather than those driving true incremental impact,” IAB says.

For audio sellers, that’s the familiar squeeze. If the model is built on inputs that are thinner, slower, or harder to reconcile, audio can wind up under-credited even when it is contributing incremental reach or lift.

The IAB report — based on a survey of more than 400 senior planning and analytics decision-makers at U.S. brands and agencies — frames the foundational issues in plain terms. “Marketing mix models fail to reflect modern media consumption,” it says. The report concludes that reality leads teams to waste time stitching together fragmented data instead of generating insights. Despite widespread adoption, the report says up to 75% of buy-side users believe leading advanced measurement approaches still underperform on rigor, timeliness, trust, and efficiency.

IAB’s diagnosis is that advanced measurement is widely used, but too often not “decision-grade.” For audio, the most damaging piece of that shortfall is coverage — because MMM is still leaving meaningful chunks of the media mix partially modeled or inconsistently represented.

The State of Data report was released at the IAB Annual Leadership Meeting where CEO David Cohen announced the launch of Project Eidos, a coordinated, industry-wide effort to fundamentally modernize advertising and marketing measurement. “It’s time to address the foundational issues that have quietly undermined measurement for years,” said Cohen.

A potential fix lies in artificial intelligence. IAB estimates AI-driven improvements in advanced measurement could help unlock $32 billion in value over the next one to two years. The report says improvements could help unlock $26 billion in media investment and $6 billion in productivity value. “AI can enable marketers to run models two to three times more often than they do today,” it says, suggesting it allows marketers to update their models monthly.

That tempo shift matters for audio because MMM is often treated like an annual verdict. If the industry moves to something closer to monthly modeling, audio sellers could get more frequent opportunities to validate performance — especially when schedules, promotions, or flighting patterns create lift that can disappear in longer modeling windows.

But AI isn’t a magic wand. VP Angelina Eng says there are major concerns related to legal and security risks, accuracy, and data quality, and past workarounds and band-aid approaches have allowed the underlying issues to become worse over time. “We need to tackle problems that are systemic and foundational,” she said.

Project Eidos is an attempt to move past channel-by-channel patches and toward interoperable measurement language and standards. For audio, the most important test is whether the initiative produces the unglamorous pieces that make MMM treat the channel as “countable” with consistent inputs, clearer exposure-to-outcome connections, and common definitions that allow it to show up in models more consistently.

Download IAB’s “State of Data 2026” report HERE.