What do attention metrics measure, and how far should a planner trust them?
Part of Programmatic media planning
The short answer
Attention metrics estimate whether a person actually took in an ad and how deeply, using device signals, eye tracking, physiological response or panels. The IAB and MRC guidelines of November 2025 treat attention as a measure of persons that needs validated models and full disclosure, and they say attention alone should not be used to measure campaign outcomes.
Key takeaways
- The IAB and MRC Attention Measurement Guidelines, final version 1.0, were published in November 2025 and serve as the basis for MRC accreditation audits of attention services.
- The guidelines name four approaches: data signals, visual and audio tracking, physiological and neurological observation, and panel or survey methods.
- The MRC treats attention as a measure of persons, so it requires viewability or audibility, sophisticated invalid traffic filtration and signs a user is present.
- Judging a site on ad attention without measuring attention to its content should be avoided, according to the guidelines.
- The CIMM and IAB playbook warns against using attention scores as a pass or fail filter for publishers.
What are attention metrics in advertising?
Delivery metrics say an ad was served and could be seen. Attention metrics try to go one step further and estimate what happened in the person on the other side of the screen: whether they looked, for how long, how hard they had to think and whether they interacted. The IAB and MRC Attention Measurement Guidelines (opens in a new tab) use the word attention "to refer to the exposure, engagement, focus, and cognitive impact measured through these methods, regardless of the specific media channel or device."
Definition
- Attention metric
- A measure of how much of a person's visual, audio or mental focus an ad or its surrounding content received, reported at the level of content, placement or creative.
| Category | What the guidelines say it captures | Example inputs named in the guidelines |
|---|---|---|
| Exposure | The potential level of visual or audio focus on an ad or content | Viewability, time in view, audibility, audible time, closed captions |
| Cognitive load | The mental effort needed to process the ad or content; emotional response sits under it and is not required | Measures of how complex or easy the ad is to understand |
| User interaction and behavior | How frequently people engage with the ad or content | Clicks, swipes, multimedia interaction, scroll speed, direction changes and pauses |
The guidelines also split measurement by what it describes. Content attention concerns the page, site, app or channel around the ad. Placement attention concerns the slot: its location, size and format. Creative attention concerns the ad itself. The guidelines say placement and creative measurement "should be distinguished in reporting," because a placement's history and a specific creative's result answer different questions.
What did the IAB and MRC publish on attention measurement?
On November 11, 2025 the IAB published the final Attention Measurement Guidelines (opens in a new tab), version 1.0, developed with the Media Rating Council. The IAB says more than 200 experts from brands, agencies, publishers and measurement companies contributed, and that the guidelines "serve as the basis for MRC accreditation audits of attention measurement services." The MRC lists the document in its standards catalogue (opens in a new tab) as issued in November 2025. A companion playbook for marketers (opens in a new tab) from the Coalition for Innovative Media Measurement and the IAB came out at the same time.
| Approach | What it observes | What a buyer should know |
|---|---|---|
| Data signals | Device, placement and publisher signals such as time in view, scroll depth, media plays and interactions | Scales across impressions, in the playbook's words; focus is inferred from the signals |
| Visual and audio tracking | Eye tracking, gaze tracking, facial coding, presence monitoring and audio signals | Measures a person directly and needs that person's participation |
| Physiological and neurological observation | Responses such as heart rate and brain activity | Adds depth, in the playbook's words, and needs a participating person |
| Panel and survey methods | What people report about their engagement | Links outcomes back to perception; the sample must be representative or adjusted for bias |
The playbook sums up how the approaches fit together: "While signal models scale, visual or neurological tools add depth and panels link outcomes back to perception." For a score offered at impression scale, ask which approach produced it, whether it was calibrated against a direct method, and how recently.
How is attention different from viewability?
Viewability is a condition of the ad: enough pixels on screen for long enough. The MRC viewable impression guideline (opens in a new tab) sets that condition for display at 50 percent of pixels for at least one continuous second. Attention is a claim about a person. The MRC therefore treats attention as an audience measure, "a measure of persons, not devices, as persons generate attention," and requires the audience standards to be met first: viewability or audibility, full sophisticated invalid traffic filtration and an indication that a user is present.
| Question | Viewability | Attention |
|---|---|---|
| What is measured | Whether the ad had the chance to be seen | Whether and how deeply a person took it in |
| Unit | The impression on a device | The person, though no one has to be identified |
| Time threshold | One second for display, two continuous seconds for video | No minimum when reported outside the viewable base, which must be labeled as diagnostic |
| Presence of a person | Not required | Required, or modeled from data that accounts for it |
| Role in the guidelines | Entry condition and input | Built on top of the viewable base |
The guidelines allow attention to be reported without the viewable condition only under set conditions, and such figures must be presented separately and "distinguished in reporting as non-standard and diagnostic." They also strongly encourage reporting attentive time relative to viewable time. A report that shows a single attention number with neither of those details has left out the context a planner needs. The viewability guide covers the pixel and time rules in full.
Should attention metrics guide site selection?
They can inform it, within limits the standard sets. The guidelines say that "Evaluating content, properties, or inventory sources solely on ad attention without consideration of the underlying content attention should be avoided," and that the MRC requires content attention to be measured when attention is used to evaluate properties or channels. A site ranked on ad attention alone may be rewarded for its ad layout rather than for readers who engage with its articles.
The CIMM and IAB playbook calls one misuse the binary decision trap: using attention "as a simple binary filter to eliminate publishers, inventory, or creative variants." Its advice is to "Use attention scores as graduated indicators within a broader decision framework, not as elimination filters," and it warns that a narrow focus on higher scores can push aside placements that suit a campaign's reach goals or brand positioning.
Where an attention score can sit in a site review
- Decide fit first. Check context, editorial quality, suitability, the selling path and audience in the market before looking at any attention figure.
- Read the score at the right level. Confirm it describes content or placement attention for this site, not a network average or one creative.
- Compare like with like. The guidelines expect benchmarks by media, device, environment and format, so compare a site's display score with display benchmarks on the same device.
- Use it as a tiebreaker or a flag. A low score on a well fitting site is a prompt to look at the page layout, not an automatic removal.
- Record the reason. If a score changed a decision, write down which score, which version and why.
What questions should a buyer ask about an attention score?
The guidelines' disclosure section says a measurement organization's methodology "should be fully described and accessible to users of the data," including the sampling, data collection, adjustment, calculation and limitations. The modeling section adds that models should be validated at least annually with fresh deterministic data, and that sample based projections disclose their standard errors. Those requirements turn into practical questions.
Questions to put to an attention measurement provider
- Which of the four approaches produce this score, and which direct method was used to calibrate the model?
- Is the service accredited by the MRC for attention, under audit, or neither?
- Is the score built on MRC compliant viewable impressions, and is any non-viewable figure labeled as diagnostic?
- Does the score separate content, placement and creative attention?
- Which devices, formats and environments does it cover, and which are excluded?
- When was the model last validated, and on what data?
- What privacy rules and consent conditions apply to the data collection?
- Is attentive time reported relative to viewable time?
Accreditation is a specific status, not a description. The MRC accreditation guide explains what an audit covers and why a provider's claim to follow the guidelines is not the same as being accredited.
How does PlacementGraph check the page around an ad before it is bought?
What does a careful use of attention data look like on one plan?
Fictional example
A sunroom builder's spring display plan in West Virginia
An unnamed builder of sunrooms and screened porches serving Charleston, Huntington and Morgantown asks its agency for a spring display plan beside home improvement, regional lifestyle and home design coverage. The client's marketing lead has read about attention scores and asks the agency to drop any site that scores below an internal cutoff.
The planner first builds the list on fit: context, editorial quality, suitability, the selling path and audience in West Virginia. Then the agency asks the attention provider the checklist questions and learns that the score is a data signal model, reported at placement level, built on viewable impressions, with no separate content attention figure.
Following the guidelines, the agency does not use the score as a cutoff. Two well fitting regional sites with lower placement scores stay on the list, flagged for a layout review, and the agency explains to the client that the score describes ad slots, not the articles readers come for. After launch the agency reports attentive time beside viewable time and keeps outcome measurement separate.
Frequently asked questions
Does attention measurement require personal data about viewers?
The guidelines say meeting the MRC audience requirements does not mean identifying an individual or assigning demographic or behavioral characteristics. They exclude from scope any approach that does not follow applicable privacy rules or that fails to ensure user consent and transparency. Privacy practices still vary by provider, so ask how each one collects its data.
Can attention scores from two providers be compared directly?
Not safely without checking methods. Providers may use different approaches, calibration panels, thresholds and coverage. The guidelines call for benchmarks across media, device, environment and format, with performance criteria set before the study. Compare scores only when both describe the same level, such as placement, on the same device and format.
Is a higher attention score proof of more sales?
No. The guidelines say attention on its own should not be used as a measure of outcomes, and they strongly encourage rigorous empirical support and independent audit for any claimed link between attention and outcomes. They also note the effect of attention on outcomes may be diminishing, nonlinear or even negative for certain ads.
Do attention metrics apply to connected TV and audio?
The guidelines cover ad activity across media channels where attention can be measured, including audio without visual media. They note that environments less prone to interaction, such as TV and connected TV, may be expected to contribute less to interaction measures, so benchmarks should be set by environment rather than shared across them.
Sources
- IAB and MRC Attention Measurement Guidelines, version 1.0 (opens in a new tab)IAB and Media Rating Council, November 2025
- Attention Measurement: The Industry Framework for Measuring Attention (opens in a new tab)IAB, November 11, 2025
- Attention Measurement Playbook for Marketers (opens in a new tab)CIMM and IAB, November 2025
- Standards and Guidelines (opens in a new tab)Media Rating Council
- MRC Viewable Ad Impression Measurement Guidelines, version 2.0 (opens in a new tab)Media Rating Council, August 18, 2015
Researched and written by PlacementGraph Research. Read how we research, source and update these pages. Corrections: diana@placementgraph.com.