Measurement
Part of Out of home advertising measurement works when the method matches the decision
What each out of home advertising key metric means for your next buy
A practical guide to out of home advertising key metrics in England, covering reach, impressions, attribution and the reporting habits that keep a buy accountable.
What to take away
- Impressions, reach and frequency are modelled estimates from audience panels and route data. They carry confidence ranges, not proof of attention.
- Effect metrics such as response, visit lift and footfall need a control area or a matched period before the change can be called a campaign result.
- Route is the UK out of home audience currency, published in waves. Note the wave and the audience definition before any figure enters a plan.
- Google retired first-click, linear, time-decay and position-based models in Google Ads in 2023. Last-click and data-driven attribution are what remain there, and neither suits outdoor by default.
Impressions, reach and frequency: what the UK numbers count
An impression counts a possible view of a poster or digital screen. Reach counts the different people who saw it at least once. Frequency counts the average exposures each of those people received.
In England these figures come from audience panels and route modelling rather than a census. They are estimates, and the definition behind a number matters more than the headline itself.
Route is the industry body that commissions the UK out of home audience currency. It is funded through the sector rather than by one media owner, and it publishes estimates in waves. Ask which wave you are being shown.
Formats behave differently. A roadside panel on a commuter route can post high reach and short dwell time. A rail concourse panel can post lower reach with longer exposure. Out-of-home advertising sets out the format families involved.
Roadside vs rail concourse panels
Roadside panel
- Reach
- High
- Dwell time
- Short
- Best use
- Broad awareness
Rail concourse panel
- Reach
- Lower
- Dwell time
- Longer
- Best use
- Longer exposure
Large billboard sites in England generally need advertisement consent. The Town and Country Planning Act 1990 is the statute behind that planning control.
To fix an audience definition you can defend, read how out of home advertising measurement and reporting should match the method to the decision in front of you.
Example: turning loop data into an impression count
The arithmetic below is illustrative. It is not a benchmark and not a rate card.
- Start with the site's modelled weekly audience for a full loop: 360,000 impressions against an adult audience (illustrative).
- Count the slots in the loop. Six ten-second spots in a 60-second loop give you one sixth of screen time.
- Divide360,000 / 6 = 60,000 impressions for your spot.
- Confirm the audience definitionadults 16+, or a narrower group such as commuters.
- Work out cost per thousand. At £600 for that panel week (illustrative), £600 / 60,000 x 1,000 = £10 CPM.
Real rates vary widely with format, location, season and demand. Ask the media owner or your agency for a current rate card, then check what the quoted audience definition covers.
Response, visit lift and footfall: effect metrics that need a control
Effect metrics answer a different question. Did behaviour change? Typical measures are branded search volume, sessions from a campaign area, footfall lift and promo code redemption.
Decide attribution only after the decision is set. Out of home advertising attribution methods explain why the model should follow the question you are answering, not the other way round.
Each effect measure needs a comparison group. Split test areas, hold one back and compare. Where a full test is impossible, use a matched historical period and label it as such.
Attribution models decide which touchpoint receives credit. Google retired first-click, linear, time-decay and position-based models in Google Ads in 2023. As of 2026, last-click and data-driven attribution are the remaining choices there.
Its guidance on attribution models explains how data-driven attribution spreads credit across touchpoints. Outdoor rarely converts directly. It usually sits early in the path, so last-click rules under-credit it.
Choosing an attribution model
Does outdoor exposure precede an online conversion?
use data-driven attribution
last-click may suffice
What each metric measures and where it fails
| Metric | What it counts | Usual source | Main weakness |
|---|---|---|---|
| Impressions | Possible views of a panel | Panel and route model | An estimate, not a count of people |
| Reach | Different people seeing the ad once | Panel and route model | Depends on the audience definition |
| Frequency | Average exposures per person | Panel and route model | Averages hide heavy viewers |
| Response lift | Change in search or site visits | Attribution platform | Needs a control group |
| Footfall lift | Change in area visits | Mobility data | Weather and seasonality |
| Cost per thousand | Cost per 1,000 impressions | Rate card and delivery report | Sensitive to rotation assumptions |
Delivery metrics describe the buy. Effect metrics describe behaviour. Read them in that order and never swap one for the other.
Keeping delivery and effect apart in one report
Keep delivery and effect in separate columns of the same view. Delivery shows whether the buy ran as planned. Effect shows whether it did anything. Blending them lets a strong reach number disguise a flat response.
Refresh live campaigns monthly. Post-campaign, report delivery against plan and effect against control within four weeks, while the data is still clean.
A working reporting dashboard for out of home advertising should show planned against delivered impressions, reach and frequency, then response measures with their control comparisons.
Every number needs a named source, a date and a defined audience. Without those three, it cannot go into a plan. Use the table above as a checking device when a plan crosses your desk.
Common questions
Which metric matters most?
None on its own. Reach and frequency describe delivery; response and footfall lift describe effect. A defensible buy needs at least one of each.
Can I compare out of home impressions with online impressions?
Only with care. Online impressions count served ads; outdoor impressions estimate possible views. Treat cross-channel comparisons as directional at best.
How often should reporting be refreshed?
Monthly works for live campaigns. Add a final post-campaign read within four weeks of the last play date.
Do I need a control group for every campaign?
No, but you need one wherever you intend to claim a behavioural result. Without it, describe the outcome as an observed change, not a campaign effect.
