Facebook studied its ads with out-of-home. Together: 13% more efficient than expected.
Facebook commissioned Kantar and Accenture to test its ads alongside out-of-home, and running both beat every other mix. Below is the report we build to find your number, from your own data. No device IDs, no location panels.
Facebook's advice to advertisers: run its ads with out-of-home
Facebook sells against every other medium. It still published research, run with Kantar and Accenture, saying its ads do better next to out-of-home.
- Broader and younger. The pairing reached Gen Z and Millennials far better than TV or print did.
- It moves purchase decisions. Younger Europeans were significantly more likely than the general population to say Facebook and out-of-home help them decide what food and drink to buy.
- We measure this in your account. Our paid social read compares completed views on the ads you already run, DOOH markets against holdout markets. If the screens make your Facebook ads work harder, it shows up there.
Sources: Facebook for Business, OAAA summary, Social Media Today. Published 2019. The 13% comes from one campaign in France; the survey covered the UK, France and Germany.
What the published research says out-of-home does
The headline numbers and this section are published results. The sample study that follows is simulated. These studies are why search and site visits are the first two reads in our design. DOOH is also where the money is moving: the OAAA reported DOOH revenue up 18.5% year over year in Q2 2026, 38.4% of a $3.16 billion quarter (OAAA).
| Study | What it looked at | What it found | Source |
|---|---|---|---|
| Facebook with Kantar and Accenture, 2019 | Consumers in the UK, France and Germany, plus a Danone campaign on TV, out-of-home and Facebook | Running Facebook ads and out-of-home together worked best, 13% more efficient than expected. The pairing reached a broader, younger audience than TV or print. | |
| OAAA and Kochava, May 2026 | Device-level data from hundreds of campaigns in seven verticals | Median lift of 20% on in-person outcomes and 14% on digital outcomes (site visits, app downloads) for out-of-home, against about 10% for CTV and for broadcast TV. | OAAA |
| OAAA and Kochava, May 2026 | Large quick-service restaurant studies inside the same report | Out-of-home exposure came before the search 96% of the time and before the social action 94% of the time. Conversion rate to a digital action rose 7.1 times between 1 and 10 exposures. | Billboard Insider |
| OAAA and The Harris Poll, August 2026 | Survey of 2,015 US adults | 86% say they would act on a retail out-of-home ad that interests them: 45% would visit the retailer's site or app, 42% would search for the retailer and 33% would visit a store. | Billboard Insider |
| OAAA and Comscore, 2022 | Survey of 1,580 US consumers aged 16 to 64 | 41% had used a search engine after seeing an out-of-home ad in the past six months. Out-of-home drew 22% of search actions on 4.1% of ad spend, 5.3 times its share. | OAAA |
| Keen Decision Systems and Accretive, October 2026 | Marketing mix data from the two companies' client brands | 89% of brands buying out-of-home saw a profitable return. Median return was $1.76 per dollar for DOOH and $1.78 for out-of-home overall. The bottom quarter of DOOH buyers got back less than a dollar. | PPC Land |
| BlueAlpha for beehiiv (published 2026) | Geo holdout on a $300,000, four-week New York subway campaign | About 100,000 incremental new site visitors at roughly $3 each, at 95% confidence. Only about 100 incremental signups, so it was cheap for traffic and expensive for direct conversion. | BlueAlpha |
What this means for a test
- Search comes first. In surveys and in device data, searching for the brand is the most common thing people do after seeing an out-of-home ad. It is the first read here.
- Weight matters. Response climbed steeply from 1 to 10 exposures. That is why each test market gets a planned weight per adult, and why we check lift against weight.
- Last-click reports hide it. If the screen comes before the search, the search ad gets the credit. A holdout market is how the screens get theirs.
- The sample numbers below are cautious. The published median for digital outcomes is 14% among people exposed. We set a true lift of +10% on search and +6% on site visits across whole markets, where many people never pass a screen.
Read these with care
- Most of this research was paid for by the out-of-home industry or by companies that sell measurement.
- It covers all out-of-home, not only digital screens.
- Results vary a lot. A quarter of DOOH buyers in the Keen data lost money, and the beehiiv campaign bought traffic far more cheaply than signups.
- In Azerion's 2026 survey of UK planners and buyers, 58% named uncertainty about measurement and attribution as the biggest obstacle to spending more (ExchangeWire).
- None of these studies is about your brand. A test with a holdout is how you get your own number.
Forty markets, twenty pairs, one coin flip each
From here down, every number is simulated. The published medians above come from people who were exposed to the ads. This design reads whole markets, where many people never pass a screen, so its lifts are smaller by construction and it comes with a range on every read.
What this sample study read
We rank the markets by size and pair each one with its neighbour on the list, so both markets in a pair have similar search and site traffic before the flight. A coin flip in each pair picks the market that gets DOOH media. The other market is the holdout and gets no screens from this buy. Everything else the brand runs (search, social, TV) carries on as normal in both, so it cancels out.
- Test unit
- Markets. In the US, DMAs. In the UK, towns and cities, each built from the postcode districts that make it up, so Google Ads, Google Analytics and the screen list all describe the same place.
- Media
- DOOH media in the test markets only: billboards, transit, retail, fuel, gyms, offices and other venue types, with every screen picked by AMAP audience data. Each test market gets a planned weight, counted as DOOH impressions per adult per week.
- Proof of play
- Log-verified plays by screen and market. Before any lift is read, the logs confirm each test market got its weight and holdout markets got zero plays from this buy.
- Read 1
- Brand search and non-paid site visits by market, from the brand's own Google Ads and Google Analytics, plus the response to the paid social the brand already runs. No pixels and no device matching.
- Read 2
- Optional. A panel partner surveys adults in the same test and holdout markets after the flight.
- Why markets
- A screen is seen by many people at once, and none of them carries a cookie past it. Most DOOH measurement gets around that by matching mobile device IDs to screen locations. This design doesn't. The comparison is market against market, so nobody's device or ID is matched or handed over, and it works under GDPR and US state privacy laws without changing anything.
Brand search: DOOH markets vs holdout markets
Weekly searches for the brand's name in each group of markets. The average of the eight weeks before the flight is set to 100. The flight is shaded red and the two weeks after it are shaded grey. Both lines dip in the weeks around the launch, because the whole category did. The read is the gap between them.
Fake launches before the flight, no screens running
- Fake launch 5 weeks before the flight: −0.7% (−5.2% to +4.0%)
- Fake launch 4 weeks before the flight: +2.3% (−2.3% to +7.2%)
- Fake launch 3 weeks before the flight: +3.4% (−1.0% to +8.0%)
- Every range includes zero, so the DOOH markets weren't already pulling ahead before the flight.
20,000 random coin flips
Swapping test and holdout inside pairs at random shows what pure chance produces. 174 of 20,000 random flips came out as far from zero as the real assignment (p 0.009). Site visits: p 0.030. Paid social completions: p 0.007.
Non-paid site visits: the visits no ad paid for
A screen has no click. Every visit DOOH media causes arrives as direct, organic search, organic social or referral traffic, which is what Google Analytics counts as non-paid. Paid clicks are left out: the brand bought those, and its ad reports already show them.
Fake launches before the flight, no screens running
- Fake launch 5 weeks before the flight: −0.5% (−3.5% to +2.7%)
- Fake launch 4 weeks before the flight: −1.3% (−5.3% to +2.8%)
- Fake launch 3 weeks before the flight: −0.8% (−4.7% to +3.4%)
- Every range includes zero, so the DOOH markets weren't already pulling ahead before the flight.
Why before-and-after isn't enough
- Brand search, DOOH markets, flight vs their own baseline: +2.4%
- Brand search, holdout markets, same weeks: −4.1%
- Non-paid visits, DOOH markets vs their own baseline: −1.1%
- Non-paid visits, holdout markets, same weeks: −8.5%
- The whole category dipped during the flight. A before-and-after read of the DOOH markets alone would have called this campaign flat. The holdout shows what would have happened without the screens.
The ads the brand already runs worked harder
The brand's always-on paid social video ran in all 40 markets, with the same budget, bids and creative before and during the flight. We count completed views by market from the ad platform's own reporting. Nothing about those ads changed, so more completed views where the screens ran means people there were more ready to watch.
Fake launches before the flight, no screens running
- Fake launch 5 weeks before the flight: +0.8% (−4.3% to +6.2%)
- Fake launch 4 weeks before the flight: +1.8% (−2.9% to +6.7%)
- Fake launch 3 weeks before the flight: +0.6% (−3.9% to +5.2%)
- Every range includes zero, so the DOOH markets weren't already pulling ahead before the flight.
This is a supporting read. It measures attention to the brand's other ads, not sales, and the industry has no published benchmark for it yet. It needs paid social that runs evenly across the test and holdout markets, with a change freeze for the length of the test.
Does more DOOH weight mean more lift?
Not every test market gets the same weight. In this study the heavier ten test markets carried about 1.5 times the impressions per adult of the lighter ten. Splitting the pairs by weight is a check that the lift comes from the screens. If it does, heavier markets should move more. Each half is only ten pairs, so the ranges are wide.
Search +6.8% in heavier pairs vs +7.6% in lighter pairs; non-paid visits +10.6% vs +3.9%. Non-paid visits lined up with the weight; brand search did not separate. With ten pairs in each half the ranges overlap, so this run's weight check is suggestive, not conclusive. We report it either way.
Brand lift survey in the same markets
After the flight a panel partner asks adults in the test and holdout markets the standard brand questions. The holdout markets give the baseline, including the people who say they remember an ad they could never have seen. Everyone in a market counts, whether they passed a screen or not, so these lifts are smaller than an exposed-vs-unexposed survey would show. In exchange, no device has to be matched to a screen.
1,500 completes in each group, 75 per market. Recall seeing an ad for the brand came back clearly above zero (+5.9 pts). Aided brand awareness read +2.1 pts (−2.9 to +7.2); Would consider the brand read +2.1 pts (−0.4 to +4.6); Likely to buy in the next 3 months read +2.2 pts (−0.9 to +5.3): all positive, none able to rule out zero. The true lifts we set for those were +4, +2.5, +1.5 points, so these are misses on sample size, not on the media: see Sizing.
Every range contains the true lift
Because this world is simulated, we know the right answers. A good method lands near them, and its 95% range contains them about 19 times in 20.
| Read | True lift we set | What the study read | 95% range | Contains the truth |
|---|---|---|---|---|
| Brand search | +10.0% | +7.2% | +2.0% to +12.7% | Yes |
| Non-paid site visits | +6.0% | +7.2% | +0.7% to +14.0% | Yes |
| Paid social video completions | +6.0% | +7.8% | +2.6% to +13.3% | Yes |
| Recall seeing an ad for the brand | +7.0 pts | +5.9 pts | +2.4 to +9.3 | Yes |
| Aided brand awareness | +4.0 pts | +2.1 pts | −2.9 to +7.2 | Yes Can’t rule out zero |
| Would consider the brand | +2.5 pts | +2.1 pts | −0.4 to +4.6 | Yes Can’t rule out zero |
| Likely to buy in the next 3 months | +1.5 pts | +2.2 pts | −0.9 to +5.3 | Yes Can’t rule out zero |
7 of 7 ranges contain the true lift. Some reads land under the truth (brand search +7.2% against +10.0% set) and some over (non-paid visits +7.2% against +6.0%). That spread is the noise in a 40-market test, and it's why every read comes with its range.
How big the test needs to be
We re-ran this study 300 times at each true lift, with fresh noise every time, and counted how often the read came back clearly above zero. A test should find the lift you hope for at least 80% of the time. Below that, a real effect can easily read as nothing.
Search and site visits, 40 markets, 6 weeks
Survey awareness, by sample size
| Markets in the test | Smallest search lift found 4 times in 5 | Smallest site-visit lift found 4 times in 5 |
|---|---|---|
| 20 markets (10 pairs) | +11% | +12% |
| 40 markets (20 pairs) this study | +7% | +8% |
| 60 markets (30 pairs) | +6% | +6% |
At the volumes in this sample. A true +4 point awareness lift was found 53% of the time with 1,500 completes per group and 71% with 3,000, which is why three of the four survey questions here can't rule out zero. Before a real test we size it on the brand's own search and visit history, market by market, and tell them what lift the test can see before they commit.
From kickoff to the read
- Read-only access. Google Ads and Google Analytics, with at least 12 weeks of history, ideally a year. For the paid social read, reporting access to that ad account. We read search, visits and views by market; we never change an account.
- Pick and pair the markets. We size each market from that history, drop the ones too small to read, pair the rest and flip the coins. The brand sees the list before anything runs.
- Build the plan in the test markets. AMAP audience data picks the screens and the plan sets a weight for each market. We plan for at least half an impression per adult per week, because thin weight spread across a whole market is a common reason a DOOH test reads nothing. The plan activates in the DSP the brand already uses, or we manage the buy.
- Keep the holdout clean. No screens from this buy in holdout markets. Other media can run as normal, as long as it runs the same way in both groups. Tell us about any regional pushes and any out-of-home the brand already has up.
- Run the flight. Six weeks of DOOH media in the test markets, with a weekly proof-of-play check that every market is getting its planned weight.
- Field the survey (optional). The panel partner surveys both groups in the last week of the flight or just after it.
- Read it two weeks after the flight. Search, site visits, paid social and the survey, with fake launches, random reassignments and the weight check, in a report like this one.
What this design can't do
- Anything that happens in only some markets during the flight (a local news story, a competitor's regional push) can look like lift. Pairing many markets dilutes it, and the fake launches catch most of it.
- Screens don't move, but people do. Commuters and travellers carry a little exposure into holdout markets, which makes the read smaller than the truth, never bigger.
- It needs many markets. A campaign in one or two markets can't be read this way. For those we rotate weight between groups of screens inside the market, or stagger the start dates, and read the same signals.
- It reads search, site and survey response. Store visits and sales need their own read, set up before the flight.
- A small brand with little search or site traffic per market can't be read this way at a sensible budget. We say so at the sizing step rather than after the flight.
Running CTV as well? See the CTV lift sample study, built on the same design.