Nine out of home advertising measurement mistakes to fix

What to take away

• The most common mistake is treating reach and impressions as proof a campaign changed anything; they show delivery, not effect.
• Agreeing the metric after sites are booked removes the baseline, so fix the metric and baseline before the order is signed.
• Roadside, rail and retail formats do not share one audience definition, so each needs its own measure.
• Any campaign using personal data for targeting or political messaging needs a lawful basis and an audit trail.

This list covers UK out-of-home plans, judged on whether the error stops a buyer evidencing effect. Keep England, Scotland, Wales and Northern Ireland data separate.

Mistake one: counting impressions as outcomes

An impression records that a panel was available to be seen, not that anyone noticed it or bought anything. Impressions alone cannot tell a strong site from a weak one. Pair delivery with a business metric such as enquiries or footfall. Our guide to out of home advertising measurement and reporting explains how to match a method to the decision.

Mistake two: metric set after booking

Teams often book sites first and decide how to measure them later, by which point the baseline week has passed. Agree the primary metric, the baseline period and the data owner before the order is signed. If a before-and-after window is impossible, say so rather than presenting a post-campaign figure as evidence.

Mistake three: one audience definition

A roadside panel, a station concourse screen and a checkout display reach different people in different moods. One definition flatters some formats and penalises others. Build a definition per format, state the geography it covers, and note whether the estimate comes from a traffic count, a panel or a model.

Mistake four: ignoring data protection rules

Audience targeting that touches personal data, especially around political messaging, sits under the Information Commissioner's Office rules for personal data in political campaigning. Treating a location list as anonymous because it carries no names is a common error. Check the lawful basis and the retention period before data reaches a media owner.

Mistake five: no creative compliance check

Measurement collapses if the ad is pulled. Claims, pricing and promotional wording are covered by the CAP Code, and the ASA can refer persistent non-compliance to Trading Standards through its trading standards referrals process. Build a compliance step into the schedule.

Mistake six: content-led formats as posters

Native and content-led placements are judged on engagement and dwell, not gross rating points. The IAB UK native distribution principles are a sensible reference for labelling when editorial-style creative runs in a paid placement. Without that labelling, response looks better than it is.

Mistake seven: reporting one blended number

A single campaign figure hides which sites, weeks or creatives moved the metric. Split delivery by site, week and creative, then show the business outcome beside it. A structured out of home advertising reporting dashboard makes that split repeatable, so the next buy starts from evidence rather than a rebuilt spreadsheet.

Mistake eight: modelled versus observed data

Modelled audiences are estimates built from counts and assumptions. Observed data comes from a device, a panel or a transaction, and carries a different confidence. Label each figure with its source type, and never present a modelled uplift as measured.

Mistake nine: glossary and audit trail

If two people read the same report and define reach differently, it has failed. Keep a short glossary with the campaign and version it.

• Reach: estimated unique people or devices exposed in a period.
• Impression: one opportunity for a panel to be seen.
• GRP: impressions as a percentage of a defined audience.
• Dwell: time spent in front of a placement.
• Uplift: change in a business metric against a stated baseline.

Common questions

Which measurement mistake costs the most?

Booking sites before agreeing the metric. It removes the baseline and leaves the buyer unable to prove effect, even when the campaign worked well.

Can one metric cover every format?

No. Roadside, transport and retail placements differ in audience, dwell and context, so each needs its own definition.

How often should a report be reviewed?

At least weekly during a live burst, so a weak site or a pulled creative can be corrected while spend remains.