The measurement debate TV really needs to have
Opinion
Audience measurement matters. But it is not where the budget battle is being lost, says the founder of media research and econometrics agency, Colourtext.
Justin Sampson’s recent call for the industry to defend shared, independently governed audience measurement standards is right and important. A fragmented landscape in which platforms self-declare their own performance and restrict scrutiny would be bad for advertisers, bad for accountability, and bad for the market.
But I want to raise a different question. Even if every issue Sampson identifies were resolved tomorrow, would it move the needle on advertising budget share for television, radio and outdoor?
I do not think it would. Winning that argument is necessary, but not sufficient. The budget battle is no longer being fought on the terrain of audience measurement. It is being fought on brand outcomes. And that is a game in which offline media is systematically disadvantaged by the way outcome measurement works.
The shift to outcomes-based evaluation
The advertising industry has moved towards outcomes-based evaluation. That sounds sensible. Brands want to know what their advertising actually achieved, not just how many people might have seen it.
But the measurement systems driving those evaluations were largely designed for performance channels, for clicks and conversions that happen immediately and can be tracked to a specific touchpoint. Attribution was built for a world where a consumer sees an ad, clicks it, and converts within a measurable time window. It works beautifully when the effect is instantaneous and online.
Offline media does not work that way. Television, radio and outdoor build their effects over time. They persist in memory. They influence behaviour weeks or months after exposure, often through pathways that never touch a click trail.
Attribution struggles to see this. It systematically favours performance channels because those channels are structurally easier to measure in the short term.
When offline media is evaluated by these systems, it looks weak. Not because it stopped working, but because the measurement technology stopped seeing it.
The offline attribution trap
I have seen this problem up close. Colourtext was asked to examine radio response measurement for Radiocentre in London. The standard attribution testing window was basically “did anything happen within a few minutes of the ad playing out?” Five minutes. Perhaps twenty, if you are feeling generous.
Under that framework, radio was getting absolutely hammered. It looked weak. It looked disposable. It looked like the sort of budget line item you cut to fund more search ads.
But when we modelled it properly, allowing for delayed response, we found something very different.
Those short attribution windows were capturing only about 8% of radio’s actual ad response. The remaining 92% occurred outside the measurement window and were credited to the channel that was ready to collect them. Usually paid search.
Under those circumstances, the post-campaign meeting writes itself. “Search is driving results. Radio isn’t.” It sounds empirical. It sounds like data. But it is the wrong data, produced by measurement systems that were never designed to see how broadcast channels work.
This is what I call the offline attribution trap. It is not that broadcast stops working. It is that attribution stops seeing it. And in a world where budget allocation increasingly follows model outputs, a channel that becomes invisible to the measurement system becomes “ineffective” in budget meetings remarkably quickly.
The problem with waiting for better audience numbers
Joint-industry currencies like Barb and RAJAR are immensely valuable because they create a shared, auditable basis for trading. But they measure audiences, not outcomes. The industry conversation has moved downstream.
When a CFO asks what the media spend delivered, nobody answers in TVRs. They answer in conversions, in cost-per-acquisition, in attributed revenue. The question being asked in budget meetings is not “how many people saw this?” but “what did it do?”
That is the terrain where offline media is losing ground. Better audience measurement, necessary as it may be for other reasons, does not address the root of the problem.
What would help?
What offline media needs is the ability to show up credibly in the systems that drive budget allocation. That means econometric modelling and attribution, not just audience currencies. This requires four things.
First, media owners need their own econometric capability. Most marketing mix modelling is commissioned by brands and agencies. It answers their question: how should I allocate my budget?
That is legitimate, but it leaves media owners in a reactive position, waiting to see how their channel performs in someone else’s model, built on someone else’s assumptions.
Media owners need sufficient knowledge to recognise when a model has been poorly specified, and sufficient capability to challenge results that quietly cut their channel out of the story.
If offline media does not engage with econometric modelling on its own terms, it will continue to be judged by methodologies designed for activation, not brand-building
Second, the data has to be in the room.
The uncomfortable truth is that offline media is often underrepresented in models simply because the data is unavailable. It is late, incomplete, or in a format that cannot be joined to anything else.
Digital has the advantage of being born model-ready. It arrives with tags, timestamps and a click trail. So when the modelling starts, digital is sat at the table with its name badge on, and offline is still outside looking for the door.
This is a solvable problem, but it requires media owners to invest in and focus on their own post-campaign data systems. Global seems to be leading the way on this with its Global: IQ project.
Third, the industry needs to understand what econometric evidence can reveal about how offline channels actually work.
Properly specified models can capture effects that attribution misses entirely: the persistence of television and radio’s influence on brand outcomes over weeks and months, not days; and the interaction effects between channels, where broadcast amplifies the effectiveness of performance media rather than simply operating alongside it.
These findings only become visible when you model delayed responses and channel interactions over long time horizons. They change the commercial story significantly. But if media owners are not commissioning this work, they will not have the evidence to make the case.
Fourth, offline media cannot simply cede the attribution layer to performance channels.
Project Lantern, sponsored by the major UK television networks, represents one response to this challenge. It aims to provide attribution measurement for television based on panel data, giving TV a presence on terrain that has historically belonged to digital.
Econometric modelling and attribution measurement serve different purposes and answer different questions. Offline media needs to be credible in both conversations, not just one.
Two debates, not one
I am not suggesting we should ignore audience measurement. Justin Sampson is right that a shared currency of truth matters, and that allowing platforms to self-declare performance while restricting scrutiny is a bad outcome for everyone.
But audience measurement and outcome measurement are two different debates.
The outcome debate is where the budget decisions are increasingly being made. If offline media wins the first debate and loses the second, its share of advertising spend will still continue to decline.
The channels that cannot be seen by the systems that drive budget allocation will become the channels that get cut. Not in one dramatic moment, but through a series of sensible-looking decisions based on incomplete evidence.
A small cut here because the model did not show much. A budget shift there because search looked more accountable. A planning rule that quietly makes offline channels the thing you invest in when there is money left over from the activation budget.
That is the measurement problem television, radio and outdoor really need to solve. It is not the same problem as the one Barb is fighting.
Jason Brownlee is the founder of Colourtext, a media research and econometrics agency working with media owners and trade bodies, including Radiocentre, Global Media & Entertainment and Television Audience Monitoring Ireland.