The W3C’s proposed “Attribution Level 1” browser standard deserves far more scrutiny from the advertising and measurement community than it has so far received. The public comment period remains open through June 10.
At a moment when advertisers face enormous pressure to justify spending and many publishers outside the largest platforms are struggling economically, industry standards bodies should be encouraging more rigorous causal measurement, not reinforcing decades of confusion between observational attribution and advertising effectiveness.
The proposal is framed primarily as a privacy-preserving replacement for older forms of cross-site tracking and attribution in a post-cookie environment. But embedded within the specification is a much larger and more consequential assumption: that attribution systems are valid mechanisms for determining advertising effectiveness.
Attribution is not effectiveness
The draft repeatedly describes attribution as a way to identify “effective advertising,” determine “what ads perform best,” understand “which creative works best” and help advertisers spend more on “effective advertising” and less on “ineffective advertising.”
That framing reflects one of the advertising industry’s oldest and most persistent measurement problems: the tendency to mistake observational pathway analysis for causal evidence of incremental business impact.
Attribution systems generally do not estimate a counterfactual. They observe advertising exposures and subsequent consumer behavior, then allocate credit according to deterministic or statistical rules. That can produce useful operational reporting. Reach and frequency management and campaign diagnostics all have practical value.
As consultant Brian May recently observed in response to the W3C debate, “These signals need to be viewed as inputs to a more holistic analysis, not as an end in themselves.”
Former IAB Canada president Chris Williams made the point even more bluntly in response to the debate on LinkedIn: “The whole concept of attribution as defined is so fundamentally flawed it should be deprecated. MTA should stand for ‘analysis,’ not attribution.”
Their concern reflects a deeper structural problem in modern digital advertising systems. Platforms increasingly optimize ad delivery toward users already likely to convert. Consumers already in-market naturally generate more searches, retailer visits, social engagement, commerce activity and measurable lower-funnel signals. Attribution systems, therefore, risk confusing underlying purchase propensity with advertising persuasion.
The result is a structural bias toward channels positioned closest to observable conversion activity, including search, retail media, retargeting and click-oriented social advertising. These systems are exceptionally effective at harvesting existing demand signals.
Meanwhile, media channels often responsible for creating demand rather than intercepting it – television, audio, sponsorships, out-of-home, premium video and broader brand advertising – become structurally undercredited because their effects are probabilistic, delayed, indirect or difficult to capture through clickstream observation.
If attribution systems systematically overcredit lower-funnel media environments already optimized around purchase intent, advertisers risk steering billions of dollars toward channels that harvest existing demand rather than create new demand.
And it matters especially for the future economics of the open web.
The platform advantage
The largest platforms already possess overwhelming advantages in authenticated identity, commerce visibility, logged-in environments, AI optimization systems, lower-funnel behavioral data and massive economic clout. Many publishers outside the largest walled gardens do not.
As browser standards increasingly institutionalize attribution-centric measurement frameworks, the industry risks further concentrating measurement credibility, optimization power and advertising budgets into the same handful of dominant platforms.
A framework promoted partly in the name of privacy may also increase pressure toward more first-party identity harvesting, more authenticated ecosystems, more clean-room infrastructure and greater dependence on proprietary identity graphs.
Attribution and pathway reporting are not the same thing as causal measurement of advertising effectiveness. Although recent revisions to the W3C’s proposed framework acknowledge important limitations of attribution and the value of randomized control trials, the current draft still repeatedly implies that attribution can identify advertising effectiveness.
If the specification were narrowed to describe attribution as one observational input among many into broader marketing analysis, the proposal would be far less concerning. But language suggesting that attribution systems can determine “effective advertising,” “what ads perform best” or “which creative works best” should be removed or substantially qualified.
The advertising industry already possesses far stronger methodologies for measuring causal impact, including large-scale randomized geo experiments and other incrementality-testing frameworks specifically designed to estimate counterfactual outcomes.
Those are the approaches industry bodies should be encouraging advertisers to adopt.
Once weak measurement assumptions become embedded into technical standards and industry infrastructure, they acquire a legitimacy that can shape advertiser behavior, media allocation and competitive dynamics for years.
“Data-Driven Thinking” is written by members of the media community and contains fresh ideas on the digital revolution in media.
Follow Central Control and AdExchanger on LinkedIn.
For more articles featuring Rick Bruner, click here.
