How Tube advertising is measured

Underground advertising is measured on opportunity and frequency, not on clicks. Understanding what the numbers do and do not say is the difference between a useful evaluation and an argument after the fact.

Measurement is a separate thing from network scale. How heavily the network is used is published by TfL and is a matter of record; how many of those people had an opportunity to see a particular campaign is modelled, and the two are routinely confused.

Coverage and frequency

Coverage is the share of your target audience with at least one opportunity to see the campaign. Frequency is how many opportunities the average person in that group had. The out of home audience currency models both from passenger data, site locations and visibility characteristics.

Station entry and exit counts

TfL publishes annual entry and exit counts by station. They are useful for sizing a station and comparing two of them, and they are not a campaign metric: a count of journeys is not a count of people, and a commuter passing twice a day appears many times over.

Reading the numbers on a plan

A plan arrives with coverage and frequency against a defined audience, and the two only mean anything together. High coverage with a frequency of one is a campaign most people see once and nobody remembers. Lower coverage with a frequency of four or five, concentrated where your audience actually travels, is usually the better buy — which is why cutting the station list often improves a plan rather than weakening it.

Check what the audience definition is before comparing two plans. Coverage against all adults and coverage against a narrow target are different numbers, and a plan can be made to look stronger simply by widening the group it is measured against. If two plans are being compared, they have to be measured against the same audience or the comparison is meaningless.

Deciding what to measure before it runs

The evaluation is far easier to agree before the campaign is booked than after it has run. Decide what would count as success, and whether the mechanism to observe it already exists — a brand tracker with a pre-campaign wave, store traffic data, a landing page that only this campaign points at. Retrofitting an answer to a question nobody asked in advance is where most out of home evaluation goes wrong.

Format choice affects what is observable, too. Digital sites return play-out reporting confirming what ran, where and when; paper returns proof of posting. Neither is an audience measure, but both answer the more basic question of whether the campaign went up as booked.

What can be measured beyond that

• —Brand lift studies — pre and post surveys against an exposed and unexposed group. Worth it on larger campaigns, not on a two-week poster burst.
• —Footfall studies — for retail, matching campaign windows against store traffic.
• —Play-out reporting — for digital, confirming what ran, where and when.
• —Proof of posting — for paper, confirming the site went up as booked.

We do not publish performance figures from past campaigns anywhere on this site unless the client has approved the number and its source. A case study without a citation is a claim, not evidence — which is why our campaign write-ups describe what ran and where rather than what it delivered.