OOH impressions vs online impressions: We use one word for two different things, and we price the world on it.
An impression online is a receipt. Something happened, a system recorded it, you can go and look at the record.
An impression out-of-home is a forecast. Nothing was recorded. A model estimated how many people were likely to have seen your billboard, and that estimate became the number on the media plan.
Same word. A completely different kind of object. And almost every pricing conversation in our industry rests on the assumption that they are interchangeable.
I know the first kind well. I spent over three years on Google’s AdWords team covering the UK and Ireland, watching impressions accumulate in dashboards in near real time. Then I moved into out-of-home and spent the last six years buying billboards, bus sides, DOOH and transit panels, where I found the same word doing an entirely different job. Nobody has ever handed me a log file for a poster site. They hand me a forecast, and we both call it a measurement.
What the OOH campaign forecast is made of
The UK, where I buy, has one of the more rigorous systems anywhere. Route is jointly owned by advertisers, agencies and media owners, with fieldwork run by Ipsos. A large-scale GPS travel survey maps how people in Great Britain move. Independent traffic and footfall counts calibrate it.
Then a visibility adjustment, informed by eye tracking research, estimates not merely who passed a frame but who was realistically likely to notice it. Route calls that realistic opportunity to see, and the likelihood curves vary by transport mode and format before being applied to the traffic passing each site.
Change the market and the names change, not the architecture. Geopath performs this role in the United States and is currently rebuilding its methodology, with Ipsos running a next generation pilot. Germany has iDOOH, France has Mobimetrie, Australia has MOVE, New Zealand has Calibre, South Africa has the OHMC. Different surveys, different curves, identical underlying moves: nothing can be counted, so something must be estimated.
Then there is everywhere else, which is a much bigger share of the world than our agendas suggest. Much of Asia still has no unified measurement framework at all, which is why the push there is being led by media owners, platforms and technology vendors rather than by a joint industry body. Large parts of the Middle East and Africa are in a similar position, with currencies under construction rather than established.
Follow that to its conclusion and it gets uncomfortable. If a market has no independent currency, the audience estimate attached to a site is produced by the company selling you the site. Not necessarily dishonestly.
There may be no alternative source of data, and a media owner with good footfall research is doing something more useful than a media owner with none. But the buyer is being handed a forecast made by the seller, presented in the same vocabulary a London or New York planner would use for an independently governed one. Same word again, third meaning, and this time nobody is even pretending there is a referee.
That is the case for measurement bodies, not against them. If a jointly owned system with Ipsos fieldwork and eye tracking research still produces a forecast rather than a count, a market without one is running the same uncertainty with none of the safeguards.
The work building these currencies deserves respect. It is a serious answer to a genuinely impossible question, and I would not want to buy in a market without it. There is no ad server bolted to the side of a building. No render event fires when a bus goes past you. If you want a number at all, somebody has to build one.
But built numbers behave differently from counted ones. Two competent teams can model the same site and land in different places, both defensibly. Every input carries error bars, and none of them travel with the figure. What reaches the media plan is a clean integer under a heading that says impressions, borrowing every ounce of authority the word earned in a medium where it meant something else entirely.
Before anyone points at digital
This is the part where I am meant to say online measurement is spotless and the out-of-home industry should be embarrassed. It isn’t, and we shouldn’t be.
Fraudlogix examined 105.7 billion ad impressions across 2025 and found a global invalid traffic rate of 20.64 percent. Roughly one in five of those immaculately logged impressions carried the signature of bots or other non-human activity. Juniper Research expects digital ad fraud losses to reach 100 billion dollars this year. The ANA’s transparency work puts wasted programmatic spend at about a quarter of budgets.
So, the receipt has its own defect. It records with total precision, bots included. The forecast has the opposite defect. Everyone inside it is real, because the model is built from how actual humans travel, but the total is a probability rather than a tally. That contrast is the part I wish more people sat with. Digital offers a precise count of an audience that is partly fictional. Out-of-home offers an approximate count of an audience that is entirely real. Neither is the hard currency it presents itself as. They are different wagers and treating them as the same unit is precisely where buyers get hurt.
The comparison that breaks everything
It breaks the moment a OOH CPM lands next to a digital CPM in a spreadsheet, which happens in pitch decks daily, in every market on earth.
A CPM is price divided by impressions. In out-of-home advertising, the price is real and the impressions are forecast, so the CPM is only ever as solid as a model the buyer will never see. Then comes the genuinely corrosive bit. Put two OOH quotes side by side on CPM and the comparison quietly rewards whoever forecast most generously. The vendor with the sunniest footfall assumption wins. Nobody set out to build a market that selects for confident modelling over accurate modelling, but that is the market the metric produces.
I have watched quotes for broadly comparable sites diverge wildly, and the honest response to “how can both of these CPMs be correct” is that neither is correct or incorrect. They are two forecasts wearing the costume of a measurement.
None of which makes our rates unjustifiable. The UK alone posted record revenue of 1.44 billion pounds in 2025, close to two billion dollars, up 2.6 percent against a grim economy, according to Outsmart and PwC. Advertisers are not idiots. They return because the medium works. My argument is narrower than it might sound: the rates are earned by what out-of-home does, not by the impression arithmetic we drape across it afterwards.
What I use instead of impressions
If impressions are soft currency, what is hard? I do not have a complete answer, and I distrust anyone who claims one. But six years of buying has left me with an order of preference.
Price per site per period comes first, because it is the only figure in the entire transaction that nobody modelled. A roadside billboard in Leeds for a fortnight costs what it costs. So does one in Ohio or Osaka. It is comparable across vendors, auditable, and impossible to inflate with an assumption.
It is also the only number a smaller advertiser can realistically interrogate. A global brand has an analytics team to pull an audience model apart. The cafe owner buying their first billboard has a quote and a leap of faith. SMEs deserve a medium that is straight with them, and a visible price is the cheapest, fastest place this industry could start.
After that, share of voice in physical space. A billboard does not divide your audience among six other advertisers in the same second the way a feed does. Hold both frames at a junction and you own the entirety of the advertising attention at that junction. That is a structural property of the medium rather than an estimate of it, and we undersell it constantly.
Then context you can physically occupy. Dwell time at a bus stop requires no probability curve. You can go and stand there. No digital planner has ever visited their placement.
And finally, outcomes.
When we ran two adjacent boards on York Street in Leeds earlier this summer, I judged the campaign on what moved: new website traffic up 188 percent, press pickup, bookings that followed. Search lift, direct traffic, QR scans, footfall studies. Every one of them is imperfect. All of them measure effect rather than exposure, which is the thing anyone is really buying.
Where I am not sure
I should admit the hole in my own position, because it is a real one.
Price per site is honest, but it is not portable. It cannot tell you whether a billboard in Leeds is better value than one in Manchester, or whether roadside beats rail for a given brand. Audience models exist because planners need to compare unlike things, and my preferred metric is useless for that. Scrap the forecasts entirely and you do not get a purer market; you get one where nobody can plan across formats and out-of-home loses its seat at the cross-media table altogether.
So I am not arguing for abolition. I am arguing for demotion. Publish the price as the headline and the audience estimate as supporting evidence, with its assumptions attached, rather than the reverse. Say forecast when we mean forecast. The number survives. Its false precision does not.
What happens next
Route reaches 98 percent of the UK population weekly on its own figures, and measurement bodies worldwide are investing seriously in attribution, footfall uplift and playout verification.
That work is genuine and improving. My concern was never the people building the models. It is the trading habit that receives their careful forecasts and files them as counts.
Out-of-home will lose a precision contest against digital, and it should stop entering one. Our honest pitch beats our borrowed one: real people, in the real world, who cannot skip you, block you or scroll past you, at a price you are allowed to see. The industry that learns to sell what it really is, rather than what it once borrowed a word to imply, is the one that finally earns a bigger share of the plan.
I left the company that industrialized the modern impression for a medium that cannot count a single one. I have never been more convinced by the second. I would just like us to describe it truthfully.
Jamie Roberts spent three years on Google’s AdWords team for the UK and Ireland before moving into out-of-home. He is the founder of Loud! OOH, an independent UK agency.
