The latest Route data (R24) now incorporates views from on-train audiences who see platform ads through carriage windows in its audience measurement calculations.
Our audience measures now take account of people travelling on trains through railway stations who have a realistic likelihood to see advertising through their windows. This applies both on the approach to stations and while stopped on the platform. This creates an audience additional to the usual counts of travellers entering and leaving each station by foot.
How this was determined?
The new calculation is underpinned by two strands of independent research. Ethnographic work conducted by ACB Research and eye tracking research carried out by the University of London, Birkbeck College.
The ethnographic research
The ethnographic research used a sample of CCTV recordings of both inter-city and commuter trains departing from and arriving into Liverpool Street station. Recordings of all trains were made available. All recordings were treated in confidence and no people were personally identifiable.
With privacy in mind, rather than focusing on the people on board, the research identified a random selection of seats per carriage and observed the behaviours of the person on that seat. If the seat was empty, then no behaviours were recorded.
It was agreed that a view should only be valid through the nearest window to the passenger. As an ad can only be on one side of the train, it is assumed then that a maximum of 50% of all passengers are potentially able to view it through the window of the train.
The next step was to determine how many people look out of the window on trains, for how long and when do they do it. To understand this, we coded up the amount of time that a person in a nominated seat looked out of the window closest to them. We also noted when this occurred. The analysis focused on two key journey moments:
- When the train was in motion
- When it was stationary at a platform
Analysis determined that during peak times, of the 50% who could potentially view an ad (i.e. those on the correct side of the train), 17% spent time looking out while stopped in a station. This factor is taken forward and built into our calculation.
The eye-tracking research
Working with experts from Birkbeck College we undertook a series of lab tests monitoring eye-tracking behaviours where people were exposed to photos of views from train windows while at a station.
The objective of this research was to determine the probability of seeing different OOH ads and the distance at which they were viewable. A total of 85 photos were tested.
• 40 small, portrait images (an equal mix of 20 x 4 sheets and 20 x 6 sheets)
• 20 large, landscape images (which included 48 sheets)
• 25 decoy images.
The relative probabilities of seeing the different ad formats are also fed into our calculations.
How the calculation comes together
Using count data from the Office of Rail Regulation (ORR) we first estimate the number of trains stopping at each station per day and then the number of passengers per train, after making appropriate allowances for those entering or leaving at the station concerned. This in essence gives an opportunity to see (OTS) measure.
From there we establish the proportion of people on the train who are actually looking out of the windows. This is our realistic opportunity to see measure (ROTS). This is calculated by assuming that 50% of the train passengers will be on the relevant side of the train and that 13% (of the 50%) will be looking out of the windows.
Beyond this, we apply a probability of the frame being in the cone of vision from the window. This is taken to be 12%, based on estimates of platform length. The visibility of the ad itself is then adjusted according to the frame size and distance from the train in line with the results of the visibility research. Allowances are also made for the possibility that the view of the platform may be blocked by another train. This gives us our measure of likelihood to see (LTS).
What is the impact on the audience data?
Overall, this has led to a relatively small increase in rail advertising audiences. 48-sheets on platforms benefit most from this change, though increases are also seen for other formats too. Overall, a campaign of all 799 48-sheets on rail platforms have increased impacts by 3.8% and cover by 8.7% relative to R.23.
Below is a table outlining the effect on audience by frame size. This relates to all rail platform ads only.
Analysis at a station level shows more variety in the audience numbers. Busy interchange stations and smaller rail stations with relatively low numbers of passengers boarding or alighting see proportionally the greatest increase in audience.
It’s worth stressing that the development only affects stations where trains ‘pass through’ and so busy terminus stations, such as London Waterloo, Victoria or Southend-On-Sea remain unaffected in their audience numbers.
Where are the biggest changes in audience seen?
The top 10 stations which see the largest proportional increases in impacts in a single-week period compared to R23 are:
In absolute terms the stations seeing the greatest increases in audience impacts are:
Will the development also be applied to other environments?
While the ethnographic research was also conducted on buses to better understand in-bus behaviours, there is no reliable information available that provides us with the number of people travelling on buses at any given time. Nor are there results of visibility research to inform the probability of seeing different ads from bus windows. Without these, it is not currently possible to apply similar measures to other environments.