The MRC viewability standard turned 10 years old this year, and like most decade-old measurement frameworks in digital advertising, it has accumulated the gap between what it was designed to measure and what the industry actually needs. The standard — 50% of pixels in view for one continuous second for display, two seconds for video — was an important baseline when it was introduced. It separated ads that had literally no chance of being seen from ads that had some chance. That was valuable work in 2013.

In 2022, viewability has become a hygiene metric rather than a performance signal. Most programmatic inventory purchased through quality supply chains consistently clears the MRC threshold. Clearing the threshold tells you the ad was theoretically seeable. It tells you nothing about whether anyone actually paid attention to it. The industry is now having the conversation it should have been having for several years: attention metrics, not viewability, should anchor display and video measurement.

What Attention Metrics Actually Measure

The attention measurement ecosystem has fragmented into several methodological approaches, and it is worth understanding what each actually captures before evaluating vendor claims.

Eye-tracking panels, anchored by companies like Lumen Research, use webcam-based gaze tracking on opt-in panels to measure where on a screen users are actually looking during ad exposure. The data is precise and behaviorally grounded — if your ad was on screen and a user’s gaze never intersected with it, that is genuinely useful signal. The limitation is panel size; eye-tracking studies are inherently sample-based rather than census-level.

Proxy signal models, the approach used by Adelaide, construct attention probability scores from a combination of observable signals: time-in-view beyond the MRC threshold, screen real estate (full-page versus banner), scroll velocity, device type, and publisher context quality scores. Adelaide’s AU metric is not measuring attention directly — it is measuring the conditions that correlate with attentive exposure based on panel-calibrated research. The advantage is that it applies to all impressions at scale, not just a calibration sample.

Behavioral signals — scroll depth, hover time, video completion, interaction rates — are a third approach, embedded in platforms like DoubleVerify and IAS as they build out their measurement suites. These are observable at census scale and do not require panel calibration, but they are noisier than eye-tracking as direct proxies for attention.

DoubleVerify and IAS Are Pivoting From Brand Safety

The competitive dynamic between DoubleVerify and Integral Ad Science has historically been about brand safety and viewability. Both companies went public in 2021 — DV in April, IAS in June — and both spent significant portions of their S-1 roadshow stories positioning themselves as the verification layer for quality digital advertising. That story has served them well.

The pivot to attention metrics represents a strategic bet that quality measurement is evolving from “did the ad run in a brand-safe environment” to “did the ad have a chance to generate a real impression on a real human.” DoubleVerify announced its attention measurement initiative in late 2021, and the company is building toward integrating attention scores into its DV Authentic Brand Suitability framework. IAS has been developing similar capabilities with a focus on attention-based optimization signals for DSP integration.

The platform risk for both companies is that if attention measurement becomes the dominant currency, the companies best positioned to benefit are those with the largest panel calibration infrastructure and the most sophisticated proxy signal models — not necessarily the ones that built their business on brand safety verification. This is why the partnership between Adelaide and GroupM announced in 2021 is strategically interesting: GroupM is validating a specialized attention vendor rather than waiting for its existing verification partners to build out equivalent capabilities.

The Case Against Viewability as a Proxy for Value

The empirical case for replacing or supplementing viewability with attention metrics rests on a straightforward observation: viewability and attention correlate weakly. A half-page interstitial that occupies 100% of the viewport and holds position for 10 seconds on a slow-loading page is 100% viewable. A user who is actively scrolling past that interstitial while glancing at their phone screen is likely paying minimal attention. An ad placed above the fold in a premium editorial environment with a slow-loading creative asset may technically miss the two-second video viewability standard while actually capturing significant attention from users who notice the brand frame before completion.

Research from Lumen has consistently shown that time-in-view beyond the MRC threshold is one of the strongest predictors of attention — but the marginal value of additional time-in-view decreases after roughly five seconds. The shape of the attention curve matters for creative strategy: a 15-second video that holds attention for the first five seconds may deliver comparable attentive impressions to a 30-second video that loses attention at second six.

For advertisers, the practical implication is that optimizing purely for viewability compliance leads to inventory selection decisions that are not necessarily optimized for attention probability. A publisher with a 95% viewability rate is not necessarily delivering 95% of impressions with meaningful user attention. The gap between the two numbers is what attention metrics are designed to quantify.

What It Would Take to Make Attention a Trading Currency

The barrier to making attention a primary trading metric is not technological — the measurement methods exist, the proxy signal models are validated well enough to be directionally useful, and the panel research base is growing. The barrier is standardization.

Viewability became a trading currency because the MRC ratified a specific, binary technical definition. An impression either meets the standard or it does not. Attention probability is continuous, vendor-specific, and methodologically diverse. Two attention vendors measuring the same impression can produce meaningfully different scores based on their weighting of different signals. Until there is a common framework — either MRC-ratified or adopted by major agency holding groups as a condition of spend — attention metrics will supplement viewability rather than replace it.

The MRC’s work on attention measurement standards is in early stages. GroupM’s commitment to buying on attention-based metrics for certain inventory classes is moving faster than any standards body. The practical path to attention becoming a trading metric is probably agency holding group adoption ahead of formal standardization — similar to how viewability itself became a standard in practice before the MRC framework was complete.

For media buyers right now, the actionable position is to start incorporating attention metrics as a secondary scoring dimension in inventory evaluation, run head-to-head tests between attention-optimized and viewability-optimized campaigns against brand lift and recall metrics, and begin the conversation with publishers about attention-based guarantees as a premium tier above standard viewability compliance.

FAQ

What is the MRC viewability standard and what does it require? The Media Rating Council standard requires that 50% of an ad’s pixels be in the user’s viewport for a minimum of one continuous second for display ads, and two continuous seconds for video ads. An ad that meets this threshold is classified as “viewable”; one that does not is classified as “non-viewable.” The standard was designed to filter out ads with zero chance of human exposure, not to measure actual attention.

What is the difference between Adelaide’s AU metric and Lumen’s eye-tracking approach? Lumen uses webcam-based gaze tracking on opt-in panels to directly measure where users look, providing empirical attention data calibrated to specific creatives and placements. Adelaide builds proxy scores from observable behavioral and contextual signals — time-in-view, page position, scroll velocity, device type — calibrated against panel research. Lumen provides ground truth at panel scale; Adelaide applies calibrated predictions at census scale.

Are attention metrics currently usable for programmatic buying? Attention scores from vendors like Adelaide and IAS can be integrated with DSPs for pre-bid decisioning in some inventory environments, and can be applied post-campaign for measurement and optimization feedback. Full pre-bid attention-based trading at scale in open programmatic is still limited, but deals with publishers who can provide attention measurement integration are increasingly available through private marketplace structures.

Will attention metrics ever fully replace viewability? Replacement is unlikely in the near term — viewability has regulatory and contractual embedding that makes it sticky. The more probable evolution is a two-tier measurement framework where viewability remains the baseline hygiene standard and attention metrics provide differentiated value measurement above that baseline. Premium inventory guarantees may eventually shift from viewability floors to attention probability floors.