The IAB’s State of Data report explains why trust remains fragile as AI reshapes measurement.
Advanced measurement is at the heart of modern media planning. Brands rely on it to justify spend, guide budgets, and explain results to management. The latest IAB State of Data report shows that confidence across the buy side in these results is waning. The findings show that while advanced measurements are widely used, beliefs about the numbers vary, making it difficult for many teams to defend them internally.
This situation is important because AI is now entering the measurement process quickly and at scale. The report notes that as AI becomes more integrated into measurement workflows, models are expected to run more frequently, take less time, and reduce manual data preparation. While these changes promise efficiency, the findings also show that speed alone won’t address the trust gap teams are already experiencing. Faster output tends to raise expectations without resolving long-standing concerns about data quality, coverage, and consistency.
IAB’s newly launched Project Eidos is positioned as a response to this situation. As the report outlines, the measurements did not decline overnight. Data deteriorated over time as it became fragmented across platforms, channels multiplied, and teams relied on disconnected systems to keep up. “Advanced measurement is widely used across the industry, but it still falls short of its core promise,” said Interactive Advertising CEO David Cohen, who announced the initiative. “The days of single-channel fixes and one-off frameworks are behind us,” he said, adding that the industry now needs to “address the fundamental issues that have quietly undermined measurement for years.”
Measurement is common, but reliability is not
The report shows that adoption has not led to trust. 60% to 75% of buy-side marketers say that today’s leading measurement approaches are underperforming in terms of rigor, timeliness, reliability, and efficiency. Incremental testing, attribution analysis, and marketing mix modeling are all actively used, but few teams rely on one method without qualification. This finding suggests that more and more effort is being spent reconciling conflicting outputs between systems before making decisions.
Marketing mix modeling illustrates this challenge most clearly. The study found that no media channel is fully represented in MMM today. Games, commerce media, and creator-driven formats are most often cited as undervalued, along with gaps in areas such as CTV and audio. The report notes that this creates a distorted view of performance, favoring channels with established data over channels that have more impact on the business. As a result, budget decisions are often guided by what is easiest to measure, rather than what the team believes will produce results.
The report characterizes this as a structural issue rather than a tools gap. MMM and other advanced measurement systems were built for less complex media environments and expanded over time through stopgap solutions. These enhancements kept the system running, but reduced visibility into what was missing.
AI increases speed, but also increases exposure
Looking forward, this report shows broad expectations that AI will play a meaningful role in advanced measurement within the next one to two years. Respondents expect more frequent model updates, faster optimization cycles, and less time spent on manual data preparation. Planning teams in particular expect AI to save time on analysis and decision-making rather than cleaning up operations.
At the same time, the findings indicate growing concern. Approximately half of respondents said they expect significant challenges related to legal risk, data security, accuracy, and data quality to arise in the near future. Less than 40% report implementing or planning solutions to manage those risks. The report highlights this gap as a potential constraint on how confidently AI can be applied to advanced measurements, especially given the sensitivity of the data involved and the number of partners typically involved in the flow.
“Past workarounds and band-aid approaches have allowed the underlying problems to worsen over time,” Angelina Eng, deputy director of the IAB Measurement Center, said in the report’s closing letter, noting that AI has highlighted those weaknesses by increasing the frequency of result generation and review.
Want to know more? ask the drum
Contracts already enforce accountability
One of the strongest signals in the report is outside the model itself. Around 40% of contracts between brands and agencies and partners already include AI-related clauses covering governance, transparency and accountability, and the study predicts this number will double within the next two years.
The report suggests that this reflects the market’s reaction to uncertainty. In the absence of common standards, procurement teams set expectations per agreement and use contracts to define acceptable use, monitoring, and responsibilities. Accountability is manifested through legal language before being established through industry collaboration.
This trend affects the entire ecosystem. Government agencies are facing more detailed questions about how AI will be used within the measurement process. Platforms face increasing scrutiny regarding data access and verification. Measurement providers face scrutiny on how their output can be explained and trusted. Project Eidos entered this environment in an attempt to establish common ground before fragmentation becomes difficult to resolve.
What Project Eidos is trying to tackle
This report clearly shows that Eidos is not about choosing a single measurement method. Attribution, incrementality, and MMM each serve different purposes and answer different business questions. As our findings explain, the question lies in how disconnected the answer is across organizations and partners.
Eidos aims to establish shared definitions, consistent data structures, and systems that allow results to be compared without long-term coordination. The report describes this work as foundational and focused on helping teams understand why methodologies are inconsistent and where gaps in channel coverage exist.
Implications of the survey results for daily work
This report points to the need for brand leaders to have more direct internal conversations about what current measurement systems do and don’t represent well. Which channels are reflected in MMM, where partial information shapes decisions, and how often teams review results together are recurring themes in the findings.
This report suggests that alignment between measurement approaches is becoming a fundamental expectation for government agencies. It is increasingly necessary to regularly compare attribution, incrementality, and MMM output to explain results and plan with confidence.
For platforms and media owners, the findings show that transparency and standardized reporting are moving closer to commercial discussion. Access and consistency of data is beginning to impact not only industry conversations, but also purchasing decisions and partnerships.
Taken together, the report’s findings suggest that as AI increases the speed and visibility of measurement, long-standing weaknesses in data quality, coverage, and governance are becoming harder to ignore. Project Eidos reflects the IAB’s response to these situations outlined in the study, rather than assuming that faster tools alone can restore trust.
The measurements go so fast that they are useless. The report says it helps when teams trust what they’re seeing and understand its limitations. AI may support the results, but only if the underlying work is done first.
