Advertising has never measured its performance so extensively. Impressions, viewability, attention, clicks, conversions, brand lift, incremental sales, attribution, MMM: advertisers now have an unprecedented amount of data to evaluate their investments.
Trust itself has moved back to the center of industry discussions. Last March, France’s UDECAM devoted its annual Rencontres to “new levers of trust,” examining in particular the relationships between brands, media, and citizens. But trust doesn’t only concern relationships between market players. It also concerns the instruments their decisions are built on.
This sophistication has deeply improved our understanding of advertising effectiveness. It has also created a singular situation: in some environments, the players who sell ad space also hold the exposure data, the activation technology, and their own measurement tools.
A simple question is therefore worth asking: does a measurement produced by the seller carry exactly the same value as one produced by an independent third party?
The question isn’t about pitting proprietary measurement against independent measurement. Both can be methodologically solid and, above all, complementary. It touches something more fundamental: the trust an entire market can place in the measurement its decisions rest on.
1. Advertising has never measured its performance so extensively
For a long time, measuring advertising effectiveness mainly meant tracking a handful of broad indicators:
- Reach
- Frequency
- Awareness
- Recall
- Purchase intent or sales evolution
Digitalization considerably enriched this toolkit. Every impression can now generate data. Every exposure can be linked to a behavior. Every campaign can produce dozens of indicators tracking different stages of the consumer journey.
Methodologies have grown more sophisticated at the same time.
- Brand Lift studies try to isolate a campaign’s effect on brand perception.
- Attribution attempts to identify the touchpoints that contributed to a conversion.
- Incrementality tests compare exposed and unexposed populations.
- Marketing Mix Modeling seeks to measure the contribution of different marketing levers to business results.
This evolution is global.
In Europe, the UK launched Origin, an advertiser-led cross-media measurement system under the ISBA’s umbrella. It measures deduplicated reach and frequency across different touchpoints and runs on governance bringing together advertisers, agencies, media, and major platforms. Origin notably claims a methodology subject to independent audits.
In the US, the industry is also working toward common frameworks. In 2025, the IAB, the Media Rating Council, and the Coalition for Innovative Media Measurement published a shared attention-measurement framework, built with input from more than 200 experts across brands, agencies, publishers, and measurement companies. The goal is precisely to make approaches more consistent and comparable across media and platforms.
The same shift can be seen in Oceania. In Australia, OzTAM brought Netflix into its measurement ecosystem in 2025. For the first time, a major global streaming platform is now measured independently alongside linear TV and BVoD, with the ambition of bringing these worlds together around shared metrics. OzTAM is also working to integrate new data from millions of connected TVs to evolve its measurement system.
All these initiatives respond, in different forms, to the same paradox: the more capable an industry becomes of measuring, the more it has to figure out how to make those measurements comparable, shareable, and credible.
Because more data doesn’t necessarily mean advertisers have a simpler view of their performance.
In 2025, Nielsen found that only 32% of marketers worldwide reported measuring their media investments holistically across digital and traditional media. In Europe, that share dropped to 23%.
Measurement, then, is no longer just a reporting tool. It has become an instrument for allocating marketing capital.
A campaign deemed effective can receive more budget. A channel judged less effective can lose some. A technology able to demonstrate a stronger ROI can gradually reshape a media plan.
Measurement no longer just describes performance. It helps decide where the next euros get invested.
This shift mechanically gives more weight to the systems that produce that measurement. And it raises a question rarely asked out loud: who measures the seller?
2. When the seller also becomes the measurer
Major advertising platforms have a considerable advantage when it comes to measuring the campaigns they run: they know their own environments with a depth no outside player can fully replicate.
They can observe impressions, exposure frequency, interactions, and, in some cases, link that data to behaviors or transactions. This closeness to the data has enabled the development of extremely sophisticated, fast, and granular measurement tools.
These capabilities have contributed enormously to the success of digital advertising.
But they have also introduced a quieter break in the history of advertising measurement.
For decades, advertisers progressively asked media not to be the sole judge of their own performance. Television, radio, print, and out-of-home structured themselves around methodologies, trading currencies, and measurement bodies that let buyers and sellers rely on shared references.
The separation was never absolute — media obviously have their own data and studies. But when it comes to building a reference used to buy, compare, or arbitrate investment, having a measurement recognized by the different parties became an essential component of how the market functions.
With the rise of major digital platforms, part of that architecture changed.
The same player could now sell the ad inventory, control most of the exposure data, operate the activation technology, and itself produce part of the indicators used to evaluate performance.
And advertisers largely accepted this new arrangement.
The paradox may lie right there: by entering the “walled gardens,” advertisers sometimes accepted, on measurement, exactly what they had spent decades trying to avoid with traditional media.
When sophistication in measurement isn’t enough
Facebook’s history offers a particularly interesting illustration of this tension.
In 2016, Facebook acknowledged an error in how it calculated average video view duration. Views shorter than three seconds were excluded from the average, which led the platform to admit a 60-80% overestimation of that metric.
That same year, Facebook identified several other issues in its metrics.
In Page Insights, organic reach over 7 and 28 days had been calculated by summing daily audiences without properly deduplicating users returning across multiple days. After the correction, Facebook stated that the displayed figures would be 33% lower on average over 7 days and 55% lower over 28 days. The platform specified that paid reach was not affected.
Facebook also found that average time spent on its Instant Articles had been overestimated by 7-8% due to a calculation error. Another metric, related to app referrals, was overestimated by around 6% on average for its heaviest users, as some clicks made inside Facebook had been incorrectly counted as external referrals.
A month later, the platform acknowledged yet another misclassification, this time involving reactions to Facebook Live. The total number of reactions was correct, but some had been attributed to the wrong category: the fix led to an average 500% increase in one category and a 25% decrease in another.
These episodes obviously don’t prove that proprietary measurement is inherently wrong. They demonstrate something more interesting: a measurement can be technologically advanced, extremely granular, and methodologically sophisticated while still being exposed to the limits inherent in a system where most of the data, methodology, and control remain concentrated with a single player.
Facebook itself drew certain conclusions from this.
Following these incidents, the platform announced it wanted to strengthen third-party verification and cited comScore, Moat, Nielsen, and Integral Ad Science among others. It also explained it was responding to partner requests for an independent measure of the time ads were actually visible on screen.
In other words, the response to a crisis of confidence in proprietary measurement was, in part… more outside measurement.
The question, then, isn’t choosing between two systems
Proprietary data remains indispensable; a platform holds knowledge of its users, its environments, and its signals that an outside institute cannot fully replicate. Its tools can bring considerable speed, granularity, and experimentation capacity.
Conversely, an independent third party doesn’t necessarily have the same depth of data.
The issue, then, isn’t pitting the two models against each other. It’s about distinguishing two notions that often get conflated: methodological robustness and independence of measurement are two different qualities.
A proprietary measurement can be excellent. An independent measurement can itself have limits.
But when the same player sells the inventory, holds the data, and measures the performance, the question of independence naturally becomes a question of governance.
Because the problem, in the end, isn’t only whether a number is accurate. It’s determining under what conditions buyers, sellers, agencies, and media can accept that number as a shared reality.
And it’s precisely to answer that question that the advertising industry has historically built trusted third parties.
3. Why advertising has always needed trusted third parties
This issue, in the end, isn’t new.
The advertising industry was historically built around a separation between several functions.
- Media produce audiences.
- Sales houses commercialize them.
- Agencies advise advertisers and arbitrate investment.
- And specialized players measure audiences, environment quality, or campaign effectiveness.
This organization never eliminated everyone’s economic interests. It made it possible to build something essential to how a market functions: shared frameworks robust enough to let players with different interests make decisions from a common reality.
An advertising trading currency is only genuinely useful if buyers and sellers accept its definition.
But this logic now extends well beyond audience measurement alone.
Independent measurement has specialized too
As advertising grew more complex, the third parties tasked with evaluating it developed different areas of expertise.
Institutes like Kantar measure, among other things, the effect of campaigns on brand, sales, or ROI, and try to reconcile performance across platforms and formats. Its LIFT solution, for example, measures exposure and impact across different channels to identify their individual contribution as well as their synergies.
This independence can also operate inside walled gardens themselves. Kantar is a certified Google Measurement Partner, providing advertisers with a third-party measure of reach and Brand Lift. Its integration with Ads Data Hub allows independent evaluation of YouTube campaigns while using the aggregated data Google makes available.
Other players operate on different dimensions.
Integral Ad Science (IAS) measures viewability, invalid traffic, brand safety, and brand suitability, among other things. IAS explicitly defines itself as an independent third-party provider: it can therefore measure ad environment quality on platforms it doesn’t control. This third-party measurement exists today notably on Facebook and Instagram, and was further extended in June 2026 to YouTube Audio Ads campaigns.
New specialists are also emerging around metrics the industry is still trying to standardize.
xpln.ai, for example, is developing an independent measure of media quality and advertising attention. Its approach combines impression characteristics, context, ad position, eye-tracking data, and predictive models to go beyond viewability alone.
In other words, independence no longer necessarily means a single institute producing one number. It’s progressively becoming a verification ecosystem.
- Audience
- Exposure
- Attention
- Brand Lift
- Brand Safety
- Incrementality
- Sales
- ROI
At each step, different methodologies and players can be involved.
Third parties don’t replace proprietary data
This is an essential point.
Google knows better than anyone what happens inside YouTube. Meta holds signals no outside institute can directly access. Amazon has an exceptional depth of transactional data within its own environment.
It would be absurd to give up that information in the name of independence.
But the market’s recent evolution shows that another architecture is possible: the platform keeps the richness of its data, while a third party brings an outside methodology, comparison, or validation.
That’s precisely the principle behind the Google-Kantar integration: data stays protected inside Ads Data Hub, while Kantar can produce an independent measure of advertising effectiveness.
The line between proprietary and independent measurement is therefore becoming more subtle than it used to be.
It’s no longer necessarily about pulling data out of the platforms. It’s about letting third parties access it under sufficiently controlled conditions to confront, complement, or validate the measurement the ecosystem itself produces.
From measurement to the governance of proof
This shift deeply changes what measurement is for.
- An advertiser may hold its own CRM data.
- A platform, its own exposure data.
- A retailer, its own sales.
- An agency, its knowledge of the media plan.
- An attention specialist, its own metrics.
- An institute, a Brand Lift measurement.
- A verification player, data on viewability, fraud, or environment quality.
Taken separately, each holds part of the reality. None necessarily holds the whole of it.
The challenge, then, is less about identifying the perfect measurement than about organizing the confrontation of several sources solid enough to support a decision.
This is precisely why contemporary measurement is moving toward more triangulation.
And that leads to an important distinction: independence doesn’t mean a third party holds the whole truth — it means no interested party should be the only one allowed to define it.
This is probably one of the most important functions trusted third parties serve in the modern advertising economy.
They’re not there just to produce more numbers. They let several players make decisions based on proof that none of them fully controls.
4. What if the next challenge were measuring what happens after attention?
For several years, a large share of advertising innovation focused on one question: did we actually capture attention?
That shift mattered.
A viewable impression doesn’t necessarily mean it was looked at. And an ad viewed for longer has, all else equal, more chances of producing an effect.
But attention itself hits a limit.
Being looked at doesn’t necessarily mean being believed. And being believed still doesn’t necessarily mean triggering a decision.
This is precisely what part of the advertising market is starting to express.
Jae O., Head of Ads, Formats, Placements, Measurement & Audiences at LinkedIn, put it particularly bluntly recently: “Attention is easy to measure, influence is harder, but it’s what matters.”
His reasoning is interesting because it doesn’t stop at influence.
In B2B, he explains, a video doesn’t succeed simply because it was watched. It succeeds when the groups involved in a purchase gain enough confidence to move forward in their decision. LinkedIn then tries to connect that influence to pipeline and revenue.
The measurement value chain starts to shift: Exposure → Attention → Influence / Trust → Decision → Outcome.
Attention is no longer the outcome. It’s becoming a step.
A shift that goes beyond LinkedIn
It would be excessive to claim the entire advertising industry is already turning trust into a new KPI. That’s not the case. But several signals suggest the major media owners are now trying to understand what turns attention into decision, rather than treating attention as an endpoint.
Google, for instance, uses similar language around influence and trust in 2026. In its presentation of the year’s commercial developments, Google explains it wants to turn YouTube creators’ organic influence into real business impact, pointing to the role trust in those creators plays in moving from discovery to purchase.
In the world of European premium media, the path is different, but the question of trust becomes explicit there too.
RTL AdAlliance devoted its TV Key Facts 2025 to the role of media in building brand trust. Its study identifies awareness, cultural relevance, and presence in credible media environments as three components of that trust. Two out of three Europeans surveyed also consider brands appearing in long-form professional content more trustworthy. RTL AdAlliance goes as far as presenting trust as “the ultimate KPI” behind a brand’s success.
These approaches aren’t measuring exactly the same thing.
- LinkedIn focuses on the trust that lets a group of B2B buyers move forward in their decision.
- Google is working on creator influence and its business impact.
- RTL AdAlliance studies the ability of the media environment to strengthen trust in a brand.
But they share one thing in common: all of them are trying to understand what happens between ad exposure and its final result.
The real still-poorly-measured territory may sit right there. The market is relatively good at measuring both ends of the chain.
On one side: impressions, reach, frequency, viewability, attention. On the other: clicks, conversions, sales, acquisition, incrementality, ROI.
Between the two lies a zone that’s much harder to observe:
- What convinced them?
- What reassured them?
- What reduced their uncertainty?
- What gave them enough confidence to act?
This is where concepts like influence, credibility, proof, or trust start to matter in a particular way.
And measuring them poses a problem far more complex than measuring an impression or a click. Trust isn’t a binary event.
It can come from the brand itself, the media context, a recommendation, an expert, a creator, past experience, a customer review, or several of these at once.
It can also vary depending on the product, the perceived level of risk, and the stage of the purchase journey.
Measuring trust, then, is less about creating a new counter than about understanding how it actually contributes to a decision.
From measuring attention to measuring conviction
This may be where the next frontier of advertising measurement is taking shape.
The industry has learned to measure exposure, and is now improving how it measures attention. It’s progressing quickly on measuring outcomes but still understands imperfectly the mechanism linking all three.
The signals observable in the market do, however, call for an important nuance: it would be premature to conclude there’s a global convergence toward “Trust is the new attention.” A different shift, however, appears much more clearly: “Attention is no longer enough.”
This nuance matters.
The point isn’t to artificially invent a new universal KPI called “Trust,” but to recognize that between an ad that was seen and an ad that produces a result lies a decisive, still imperfectly understood step: conviction.
And if that conviction genuinely influences the decision, the question of how to measure it will inevitably come up. Having learned to measure whether an ad was seen, then watched, the industry may eventually need to learn to measure whether it was believed.
Conclusion — Measuring more may no longer be enough
For years, the advertising industry sought to produce more data.
Then it sought to produce better measurement.
The next step could look different. It may no longer be just about whether a measurement is technically strong, but about determining under what conditions it can become credible enough to serve as a shared reference.
This doesn’t mean replacing proprietary measurement with independent measurement.
- The former often has a depth of data and speed that’s hard to match.
- The latter can bring comparison, distance, and outside legitimacy.
Maturity may lie precisely in their ability to coexist, challenge each other, and complement one another.
Because when a measurement starts to determine the allocation of millions — sometimes billions — of euros in advertising investment, the question is no longer just: “Does it work?”
Another one appears: “Who can demonstrate it credibly enough for everyone to accept it?”
And that may be where one of the next challenges of advertising measurement now sits. The value of a measurement doesn’t only depend on what it demonstrates — it also depends on the trust the market is willing to place in it.
