Measuring What Matters: Building a More Complete View of Impact

Marketers have never had more information available to them, yet greater access to data does not always produce greater confidence in decision-making. Across media, commerce, customer experience and brand measurement, organizations are confronting the same underlying problem: individual metrics can describe one part of performance while obscuring the larger system through which marketing creates value.

Recent ARF research across attention, commerce media, premium environments, artificial intelligence and data quality points toward a broader conclusion. The most effective measurement frameworks do not depend on finding a single definitive KPI. They connect different forms of evidence across the full sequence of marketing impact, from exposure and experience through memory, behavior and incremental business outcomes.

That sequence matters because different measures answer fundamentally different questions. Reach, frequency and viewability establish whether consumers had an opportunity to encounter a message. Attention helps indicate whether that opportunity was meaningfully registered. Brand measures capture changes in awareness, memory, perception or preference. Sales and conversions reflect behavioral response, while incrementality addresses whether those outcomes were genuinely caused by the marketing activity. Each measure has value, although none provides a complete account on its own. The rapid growth of attention measurement illustrates this clearly. The ARF’s Attention as a Metric study (Adams & Donato, 2025a) found that advertisers and agencies continue to cite data accuracy and the lack of standardization as major barriers to adoption. The ARF’s Attention Measurement Validation Initiative (ARF, 2023, 2024a, 2024b; Adams et al., 2026a) also showed that attention is defined and measured differently across providers, often in ways that reflect the tools and methodologies each company uses.

Those differences make interpretation essential. An ad can be viewable without attracting meaningful attention, yet attention alone cannot establish whether the advertising changed memory, perception or behavior. Its significance depends on the creative, the channel, the surrounding content, the audience and the campaign objective. Frequency introduces another dimension, since attention may build, decline or change across repeated exposures rather than remaining constant from one impression to the next.

Research into premium media and contextual environments (forthcoming) reinforces the importance of looking beyond isolated metrics. The same advertisement may perform differently depending on where it appears, how consumers experience the surrounding content and whether the environment supports trust, relevance or engagement. Context therefore becomes part of the mechanism through which advertising works, shaping how exposure is interpreted and whether it contributes to longer-term brand effects.

Commerce and retail media raise a different, but related, challenge. Sales, conversion and return on ad spend are among the most widely used measures because they are immediate, observable and closely tied to business outcomes. Yet attributed sales do not necessarily represent incremental sales. A purchase that follows an ad exposure may have occurred regardless, particularly when the consumer was already likely to buy.

ARF work in retail media networks (Adams & Donato, 2025b; Adams et al., 2026b) has highlighted substantial variation across networks in attribution windows, comparison groups, definitions and reporting practices. Results may appear comparable in a dashboard while resting on very different assumptions. Credible measurement therefore requires a clear understanding of how control groups were constructed, which time periods were included, whether pre-existing purchase propensity was addressed and whether the findings were validated.

The rapid adoption of artificial intelligence adds further urgency. Upcoming ARF research shows that AI is becoming increasingly embedded in measurement, attribution and analytics, even as these remain among the most difficult applications to validate (Adams & Donato, in press). Automation can increase speed and scale, while also creating an impression of precision when the underlying data, definitions or models remain uncertain.

A stronger framework begins with the decision the organization needs to make. From there, marketers can determine which evidence is required across opportunity, experience, brand response, behavior, causality and long-term value. They can also assess whether the data are sufficiently complete, the definitions consistent, the methods transparent and the conclusions credible enough to support action.

In a world of endless data, understanding real impact requires more than collecting additional metrics. It requires connecting them, interpreting them in context and recognizing the limits of what each one can explain.

 

References

Adams, T., & Donato, P. (2025a, July). Attention as a metric: Perspectives from advertisers and agencies. ARF.

Adams, T., & Donato, P. (2025b, July). Retail media networks: Navigating metrics, challenges and opportunities. ARF.

Adams, T., & Donato, P. (in press). Marketers’ use of artificial intelligence 2026. ARF.

Adams, T., Donato, P., & Zhang, S. (2026a, June). ARF Attention Measurement Validation Initiative: Phase 3. ARF.

Adams, T., Donato, P., & Zhang, S. (2026b, June). The ARF Retail Media Network Standardization Project: Capabilities, metrics and practical guidelines. ARF.

Advertising Research Foundation. (2023, July). ARF Attention Measurement Validation Initiative: Literature review. ARF.

Advertising Research Foundation. (2024a, February). ARF Attention Measurement Validation Initiative: Phase 1 report (updated). ARF.

Advertising Research Foundation. (2024b). ARF Attention Measurement Validation Initiative: Phase 2 report (2nd ed.). ARF.

 

Author

Tracy Adams, PhD

Senior Director of Research & Insights

ARF