When Identity Disappears, What Happens to Measurement?
By Brian PozeskyLosing Identity Hurts Advertising. It May Break Measurement.
For the last 20+ years the advertising industry has been having the same conversation.
Cookies are disappearing. Mobile identifiers are restricted. Match rates are declining.
At this point it’s background noise. Everyone knows the mechanics. Safari and Firefox don’t have cookies. Chrome continues to threaten restrictions, pulling back and then going forward again. Apple’s App Tracking Transparency framework reduced access to mobile identifiers and on and on.
And over the last 15+ years the industry has been responding with a number of tactics to maintain the ability to identify consumers at a 1:1 level and continue the targeting and personalization that drives performance.
Identity graphs, clean rooms, retail media networks, and a growing ecosystem of probabilistic identity solutions have emerged to help marketers continue reaching consumers in a privacy constrained world.
But there is a second implication of identifier loss that we spend far less time discussing: If consumer identifiers are disappearing, measurement that depends on them is in big trouble. And that may have a bigger impact on how marketing decisions are made.
For more than a decade, digital measurement has largely relied on a simple idea. If an ad exposure can be linked to a person, and that person can be linked to a purchase, the incremental impact of media can be determined.
That logic helped power the rise of deterministic attribution models designed to connect exposure data with transaction data.
Today the environment looks very different.
According to IAB estimates, more than sixty percent of web traffic now occurs in environments where third party identifiers are unavailable. On mobile devices, the majority of users do not permit cross app tracking.
Even in the environments where identity exists, measurement visibility is incomplete.
Exposure to purchase match rates in retail media environments often land somewhere between twenty and forty percent depending on the retailer and identity framework involved.
Why 2026 Is the Breaking Point
In practical terms, that means the majority of purchases that occur after a campaign are never connected to the advertising exposure that may have influenced them. They simply never appear inside the measurement model.
Now consider the broader context of how products are actually purchased. Nielsen and Circana data consistently show that roughly eighty five to ninety percent of consumer packaged goods sales still occur in physical retail stores. Households shop across multiple retailers every month. Each retailer operates within its own data environment, often with limited interoperability.
So by relying on identifiers to link to current aggregated measurement solutions the industry is attempting to measure market level commercial outcomes using datasets that only capture a portion of consumer behavior.
When measurement depends on identifiable consumers, it can only measure the outcomes attached to those identities and projection to the whole universe of buyers becomes a required need.
Sales that cannot be matched disappear from the dataset. Cross retailer halo effects become invisible. Offline lift is understated. Market level impact is fragmented into smaller pieces that are easier to observe but harder to interpret.
The result is a form of measurement that can look extremely precise while actually being incomplete.
None of this means identity based marketing is inherently flawed. Identity remains extremely valuable for targeting, frequency management, and optimization within many digital environments.
But as identifiers become scarcer and ecosystems become more fragmented, relying on user level matching as the primary way to prove incrementality becomes increasingly fragile.
More than 60% of web traffic occurs without third-party identifiers.
Exposure-to-purchase match rates often land between 20–40%
85–90% of CPG sales still occur in physical retail.
Fortunately the measurement toolbox available today is far more advanced than it was even a few years ago.
Advances in machine learning and causal inference modeling have made it possible to analyze complex commercial systems in ways that were previously impractical. Instead of relying exclusively on deterministic user matches, modern measurement approaches can evaluate patterns, retailer environments, media exposure levels and time in order to isolate the true causal signal behind sales outcomes.
The question shifts from asking whether a specific person saw an ad and later made a purchase to asking something more fundamental.
Did media exposure actually cause sales to increase?
That may sound like a subtle distinction, but it represents a meaningful shift in how marketing performance is understood.
McKinsey has written extensively about the growing importance of causal measurement frameworks as marketing ecosystems become more complex and privacy constraints reduce deterministic visibility. The challenge is not simply linking exposures to outcomes. It is understanding what would have happened in the absence of media.
Answering that question requires analyzing the full system rather than relying on a narrow slice of identifiable consumer behavior.
In that sense the future of measurement is not about abandoning identity entirely. Identity will continue to play an important role in many aspects of advertising.
But it is unlikely to remain the structural foundation that everything else depends on.
As privacy expectations evolve and retail environments remain fragmented, measurement approaches will need to rely more heavily on models that can evaluate the full economic impact of media across markets and channels.
Otherwise the industry risks optimizing against smaller and smaller fragments of observable data while assuming that those fragments represent the whole system.
In a retail economy that is increasingly complex and privacy constrained, that assumption is becoming harder to defend.
The real opportunity ahead for our industry is not simply finding new ways to track consumers.
It is building measurement frameworks capable of understanding what actually drives growth, even when the identity of the individual consumer is no longer visible.
The solution: OmniChannel Performance Platform
If measurement frameworks built around identity are becoming structurally incomplete, the industry needs a different approach.
At Pēq, we’ve built the Omnichannel Performance Platform to measure the true incremental impact of marketing across retailers, channels, and markets, even when consumer identifiers are missing.