Nobody clicks a billboard, which is why it can be measured honestly

Out-of-home advertising has a measurement problem that everybody knows about and an advantage that almost nobody uses. The problem is obvious: nobody clicks a billboard, so the tools built for digital advertising have nothing to count. The advantage is less obvious. Because a billboard sits in one place, it has a geography, and for a business that closes sales by phone or in person, geography is something you can measure against your own records.

The industry's answer to the problem has been to measure attention, and increasingly to measure movement using location data from people's phones. Both have their uses. Neither tells a law firm, a roofer or a dental group how much revenue a board on the interstate produced.

This post covers what an out-of-home figure actually counts, where the newer measurements get their data, and a method that works from the one record every business already trusts: the list of sales it closed.

A billboard's audience is an estimate of attention. Its revenue is a question your own sales records can answer, if you set the board up to be measured before it goes up.

What an out-of-home impression actually counts

An audited US out-of-home impression is an estimate of how many passers-by are likely to notice a display over a set period. It measures the chance of attention, not visits, calls or sales.

The figure on a media plan is often read as "people who saw the ad". That is closer than it used to be, and still not what it says. It is a modelled count of people likely to notice the display, derived from traffic and visibility research, over a defined number of hours per day.

This matters because an impression count invites a calculation it cannot support. Divide the cost of a board by its impressions and you get a cost per thousand that looks comparable to digital advertising. It is not comparable, because digital impressions are served to a screen and out-of-home impressions are estimated from a road, and neither figure says anything about revenue.

• Use impressions to choose between boards: which location, which face, how long.
• Do not use them to judge whether the campaign paid for itself.
• Do not add them to digital impressions in a report as if they were the same unit.

Footfall attribution, and where its data comes from

Footfall attribution estimates whether people exposed to a display later visited a location, using location data from mobile devices. Its results are only as trustworthy as the provenance and consent behind that data, which buyers rarely examine.

The appeal is easy to see. If a vendor can say that devices seen near your board were later seen at your showroom, the medium finally has a conversion. It is increasingly offered alongside out-of-home campaigns, usually as a report of exposed visits against a baseline.

What those reports rarely foreground is where the location data comes from. Much of it originates in mobile apps, collected through software development kits embedded in apps whose users may not have had advertising measurement in mind when they allowed location access. Two cases from the same month show why the question is worth asking.

Neither case is about billboards, and neither means that every footfall study is built on data like this. The point is narrower and practical: a measurement that depends on tracking individual people's movements inherits every question about how that tracking was done. If a vendor cannot explain in plain terms where its location data comes from and how people agreed to it, that is information about the report you are about to rely on.

There is also a methodological problem that has nothing to do with privacy. The people who pass your board are mostly people who live or work nearby, and people who live or work nearby are already more likely to visit you. A footfall report has to separate the effect of the board from the effect of proximity, and the separation is only as good as its baseline.

Before buying a footfall report, ask where the location data came from. A vendor who cannot answer in a sentence has told you something.

A number on the board is the cheapest measurement you can buy

A phone number printed on one board and used nowhere else turns the calls that board produces into a lead source. Matched against closed sales, it gives a conservative floor for the board's revenue that no model can argue with.

For businesses that close by phone, this is the single highest-value thing to do with an out-of-home budget, and it costs the price of a tracking number. A personal injury firm, an HVAC company or a dental group running three boards should run three numbers, one per board or per route, and never reuse them on the website or in other advertising.

1. Issue a dedicated number for each board or each distinct route, configured so that it never appears on your website or in other channels.
2. Make the number memorable if you can — a vanity number is worth more on a board read at speed than on any other medium.
3. Export the call log with the tracked number, the caller's number and the date.
4. Match those calls against your closed sales on the caller's phone number, over a window long enough for your sales cycle.
5. Report the result as the board's floor: revenue that can be linked to it directly.

Two cautions. The floor is a floor: many people who see a board remember the name rather than the number and search for it later, and their sales will land elsewhere. And the floor is only honest if the number stays exclusive. The moment the billboard number is printed on a flyer or added to the website, the calls stop meaning anything about the board.

A vanity web address does the same job for businesses whose customers are more likely to type than to phone, with the same condition: a short address used only on the board, redirected to a page that records where the visitor came from.

The rest of the effect shows up in other channels

Most of a billboard's influence arrives through other doors: branded search, calls to the main number, walk-ins, and word of mouth. Those sales are real, and they appear in your records credited to something else or to nothing.

This is the part of out-of-home measurement that trips up the most careful buyers. The board works, the calls to its number are modest, and the budget is cut — while branded search volume, direct calls and referrals quietly fall over the following quarter. The board's effect was real; it was simply being recorded under other names.

None of this can be traced to an individual without tracking individuals, and it should not be. It can, however, be measured in aggregate, because every one of those doors leaves the same trace in your records: a sale, with an address, on a date. And a billboard, unlike an advert on a phone, has an address too.

Measuring by geography instead of by person

A geographic comparison measures a billboard's total effect by comparing revenue in the areas where it could be seen with comparable areas where it could not, before and during the campaign. It needs sales with an address, not data about anybody's movements.

The idea is old and robust: if a campaign runs in some places and not others, and those places behaved alike before, the difference that opens up afterwards is the campaign's effect. For out-of-home it has a particular advantage — the medium is geographic by nature, so the areas are defined by the placement rather than chosen after the fact.

For a local business the practical version is simpler than the statistics suggest.

1. Define the exposed area around each board — the postcodes or ZIP codes whose residents plausibly pass it — before the campaign starts, and write it down.
2. Choose comparison areas that behaved similarly over the previous year: similar revenue, similar seasonality, no board.
3. Take closed sales with the customer's address or postcode, and total them by area and by week for the year before and the months during.
4. Compare the change in the exposed areas with the change in the comparison areas. The gap is the board's estimated effect, and it includes the sales that came through every other door.

The discipline that makes this defensible is the same one that makes an attribution window defensible: decide the areas and the comparison before you see the results. An exposed area drawn afterwards around the postcodes that happened to grow is not a measurement.

The number on the board tells you the least the billboard did. The geographic comparison tells you roughly what it did in total. Report both, and the argument about out-of-home becomes a conversation about evidence.

What the out-of-home line should look like in a report

A defensible out-of-home line reports the directly linked revenue as a floor, the geographic estimate beside it with its areas and comparison stated, and impressions only as a planning figure.

The mistake to avoid is collapsing these into one number. A board's matched revenue and its geographic effect are answers to different questions, and presenting only the larger one invites a finance director to ask how it was produced. Presenting only the smaller one cuts a channel that may be doing most of its work through other doors.

• Direct: revenue matched from the board's own number or address to closed sales.
• Geographic: the estimated change in revenue in the exposed areas against the comparison, with both defined in advance and stated.
• Planning: impressions and cost per thousand, labelled as such and kept out of the revenue columns.
• Unmatched: sales with no traceable source, reported as such rather than distributed to the board or to anything else.

CloseRev matches your closed sales against your lead sources on the phone number or email address in both files, so a billboard's tracking-number log is a lead file like any other and its calls are credited only when they match a sale. Anything that cannot be linked to a source is reported as Direct / Unknown rather than credited to the board, so the floor stays a floor and the geographic comparison has an honest remainder to work from.

A short checklist

Most of the value comes from decisions made before the board goes up.

1. Give every board or route its own number, and keep it off every other channel.
2. Decide the exposed areas and the comparison areas before the campaign starts, and write them down.
3. Make sure your sales records carry an address or postcode, so the geographic comparison is possible at all.
4. Ask any footfall vendor where its location data comes from and how people consented to it.
5. Keep impressions in the planning section of the report and out of the revenue columns.

Out-of-home advertising deserves better than to be either taken on faith or cut for lack of clicks. For a business that closes its sales in person or on the phone, it is one of the few media whose effect can be measured without tracking anybody — which makes it, oddly, one of the most honest channels to measure.