What is Route?
The Route system is designed to tell advertisers how many people see an outdoor ad, and how often. It contains a wide variety of information about who those people are, allowing advertisers and agencies working on their behalf to plan ad campaigns effectively.
One of the most advanced research projects of its kind, the Route system creates data for use in campaign analysis and planning. Below, we answer some of the most common questions about the data, and the research behind it.
How can I plan with Route?
Route offers a multitude of planning options so that you can target your audience precisely. It is possible to select specific audience demographics because the Route survey questionnaires ask respondents many questions about their preferences. Download the Route respondent questionnaires here:
The questions below address the most commonly asked questions that we are asked about planning functionality.
How is data produced?
Route runs a quarterly cycle of data production. Every three months, we take a 'snapshot' of Outsmart's Inventory Mapping System. The ad sites that are in the system at that time are included in the next Route data release.
Download our latest timetable here.
This section tells you all about how the research was conducted.
For current users of the data
Understanding reliability of the data
The currency encompasses all outdoor environments: airports, buses, cinemas, London Underground, motorway services, National Rail, light rail and tram links, pedestrian shopping precincts, shopping malls, stations, supermarket exteriors, all roadside frames, and taxis.
Audience figures for campaigns are measured by reach, frequency, and GRPs.
The data offer a wealth of lifestyle, media consumption and shopping information about audiences.
The research has been designed around hyper-local geography - you can plan by town (1,600 to choose from), or create your own bespoke geography- as well as choose from 24 conurbations or the 14 BARB areas.
The latest release (R23) included 391,605 in the system.
Inventory data is managed by Outsmart. Please contact Outsmart to find out more, join, and sign up for database training.
In principle, there are several steps to go through that vary depending on your inventory type:
A) Roadside - two steps are required:
1. You will need to supply an XML file with the following information about each frame: width, height, surface area, frame type, number of faces, illumination, address, postcode, latitude and longitude. A frame ID and dimension code for each site should also be included.
2. Photographs, produced in accordance with a specification, should be submitted for each frame.
B) Retail, airports, stations, leisure - two steps are required:
1. You will need to supply an XML file with the following information about each frame: width, height, surface area, frame type, number of faces, illumination, address, postcode, (indoor) latitude and (indoor) longitude. A frame ID, dimension code, map ID number and frame position ID number should also be included.
2. Scale maps. For each location, a scale map(s), produced in accordance with a specification, should be submitted for each frame. Following the submission of data, some administrative work using the Outsmart IMS system may be required. Please contact Outsmart for more information.
Every quarter, to a regular pattern, Route takes a snapshot of the frame data held by Outsmart. This data set is processed through the Route models and thoroughly checked, a sequence that takes three months. Once data has been modelled, audience files are ready for insertion into software platforms.
However, any new environments being included must first be rigorously tested and scrutinized. This work - known as validation - takes place at Route. A technical assessment is made, and the data is approved (or not as the case may be) for publication.
Published data is made available to a regular quarterly pattern.
There will always be a certain amount of fluctuation in the Route data. This is owing to its complexity, and the need for it to be updated with respect to behaviour trends year to year. The data changes because the real world changes.
In addition, Route has taken steps to minimise any discrepancies seen in comparative analyses of similar sets of frames.
The rate card for 2017 is available to download here.
For current users of the data
Yes, it is possible to plan and trade by day-part, as the new study has been designed accordingly.
The current day-parts available are:
Weekday 0600 – 0959am (DP1)
Weekday 1000 – 1559pm (DP2)
Weekday 1600 – 1859pm (DP3)
Weekday 1900 – 0559am (DP4)
Weekend 0600 – 0959am (DP5)
Weekend 1000 – 1559pm (DP6)
Weekend 1600 – 1859pm (DP7)
Weekend 1900 – 0559am (DP8)
With the activation of the new contract, developments are being put in place to further increase the granularity with which Route data can be analysed. The goal is to be able to report at a 15 minute spot level by Q4 2018.
Yes, the data has been designed so that you can plan campaigns with single days. It is also possible to segment the data by 15 minute intervals. Please speak to your software provider for more advice.
The data is built around these time frames, yes. The ability to analyse the data in this particular way would need to be enabled by your chosen software application. Please note that the data allows this to be done within a seven day period only.
No. Not at present. Audiences are averaged over the year. The data is not designed to deal with seasonal changes in travel patterns. The date is only used to account for different daylight hours. Again seasonality of data is something which is currently being investigated with the most recent contract and will be introduced to the data in the near future.
It is possible to incorporate the effect of illumination into a campaign. Frames that are lit from above or behind, or that are digital screens, are more likely to be seen in the hours of darkness. Because the length of these hours changes across the year, we have incorporated an adjustment. So you can opt to include the 'illumination factor' by specifying the date of your campaign. However, the function does not alter audience estimates according to the season or time of year.
Dynamic frames receive a share of the total audience for the site, depending on the share of the total time in view.
There are twelve pre-set loop lengths included within the data:
|Proportion||Loop length %|
|1 in 1||100|
|2 in 3||67|
|1 in 2||50|
|2 in 5||40|
|1 in 3||33|
|1 in 4||25|
|1 in 5||20|
|1 in 6||17|
|1 in 8||12|
|1 in 10||10|
|1 in 20||5|
|1 in 50||2|
No, it is only possible to calculate audiences with twelve loop options. These have been selected as the most used time slots for digital or scrolling adverts. Some software systems may allow you to input any percentage share for rotating adverts. However, the calculation will not be based on this percent; for any percentage loop length that does not match the twelve common percents, the nearest percentage option is selected as a match.
Different types of illumination have more or less ability to increase the likelihood that a frame is seen in the hours of darkness. The factors relate to the learning from the research.
Understanding reliability of the data
What is the minimum number of panels that should be used in a campaign? What is the reason for this?
The minimum number of frames that should be used in any Route analysis is 100. (The frames should also have 100 or more respondents associated with them). Below these two thresholds, Route reach and frequency estimates are not reliable, owing to statistical constraints, and figures should be treated with caution. Statistical analysis to determine the confidence limits for Route campaign analysis is on going. Users will be informed of any alterations to the guidelines.
The minimum sample size for any campaign analysis is 100 respondents.
No, there is no variation in the recommended sample size- it is the same for all types of frames.
Your software application will show the sample size in its outputs.
The fluctuations seen in comparing analyses are a natural by-product of the complex probability models. It is an expected (and correct) statistical outcome. To minimize this effect, we have fixed audiences at frame level.
Data is pre-processed to fix audience calculations earlier in the process. It does not mean that single frame impacts are affected; these are still unreliable, and the 100+ frames and 100+ respondents rule still applies.
As part of Route’s commitment to high quality, the software bureaus that deliver Route data have an obligation to prove they are doing it right. The software applications on which Route data are currently available are built differently. However it is important to note that dataFORWARDS, Kinetic and Telmar, the providers of the systems, must demonstrate to Route that they have correctly applied the algorithm and the data, to produce the same results as our official ‘check figures’. This is done with each data release and algorithm update.
Twelve companies have access to the data for the purposes of creating software: Ayuda, Beacon Dodsworth, Bitposter, Caxis, dataFORWARDS, Holla Me, JCDecaux:NEC, Key Systems, MediaTel, Spafax Aureus, Telmar Communications and Vistar.
How is the data produced?
The typical lead time from initial frame data submission to publication as part of the Route data set is four months. There are two main reasons: the Route data is published on a three-month cycle; and there is a set time for data modelling, which is a complex process. In addition, all data must be rigorously checked and assessed before it may be published.
To find out when the frames you want will appear, please contact the relevant media owner. The management of data upload into the system is the responsibility of media owners and Outsmart.
About the research
30,000 respondents have been included in our sample.
How many questions do they answer on lifestyle, media consumption and shopping?
Our respondents answer over 250 questions, across 19 categories. These answers can be used in the planning process.
The target audience criteria are defined by the questions that Route asks its respondents. These are set by Route’s Methodology Action Group. Each year, in July, we review the two questionnaires. If you would like to make a recommendation, please inform the Committee member at your company (or for non-underwriters, your Outsmart or IPAO representative). However, there is no guarantee that it can be accommodated, owing to necessary constraints.
Ipsos Connect undertake the research fieldwork and probability modelling, and oversee the development of the overall study. Their partners, MGE Data, a Prague-based company, undertake all GIS work as well as creating complex traffic models to produce accurate population flows.
Birkbeck College, University of London, conduct all the eye-tracking studies. Additional, ad-hoc work has been done by ACB Research and TNS.
How are the figures produced?
The research study, undertaken by Ipsos Connect and MGE Data, combines several elements:
- A fieldwork sample of 30,000 people, who each carry a MST meter for nine days. This tells us who, where and when – our travel survey
- A mapped network of all the pathways in the country (including tube stations, retail mall precincts, and all types of roads) to which we affix traffic flows and audience numbers – our Traffic Intensity Model (TIM)
- Also from the TIM, we calculate the realistic opportunity to see (or ROTS)
- Eye-tracking data is used to factor the ROTS numbers to produce likelihood to see (LTS) figures. The data are derived from visibility experiments that include work on dynamic and illuminated frames
- The size, format and precise location of advertising frames are supplied by Outsmart via their Inventory Mapping System (IMS)
Figures from all of these elements are combined in a sophisticated mathematical model to produce audience probabilities for every frame in the country - the final audience data.
In addition to the five elements above, the study has incorporated: ad-hoc count studies; shopping mall flow studies; railway station flow studies; a 'views from windows of public transport' study; an illumination study; a 'bus speed at junctions' study; a study to produce upweights of commuters; a study to produce upweights of airport users; a GPS taxi study and three rounds of dynamic image eye-tracking research.
Route has published three of its eye-tracking studies: ‘Visibility: Estimating the visibility of poster panels for pedestrians’, ‘Visibility: Poster panel visibility for drivers and passengers, a first look’ and 'Visibility: Object visibility in dynamic scenes'. These papers are available to download from this site, see here. Documents concerning the methodology of the Route research are not made publicly available.
- ‘Visibility: Estimating the visibility of poster panels for pedestrians’
- ‘Visibility: Poster panel visibility for drivers and passengers, a first look’
- 'Visibility: Object visibility in dynamic scenes'
Route does not use the term ‘dwell’. It is beyond the scope of Route’s remit, which is to produce audience figures that are agnostic to the creative device that is used. A more appropriate term would be ‘exposure time’, i.e. the duration for which a person passing a site is within its visibility area. The duration is derived from the GPS speed data, and visibility distances.
Route’s visibility work adds a further layer to the calculation, reducing audience estimates from an OTS to an LTS- a ‘likelihood to see’. This information is input into Route models at an early stage. The published research paper ‘Visibility: Poster panel visibility for drivers and passengers, a first look’ may offer some helpful background information.
When measuring the likelihood of a frame to be seen, we include in our calculations a maximum distance, beyond which the frame cannot be seen. The distance varies, based on the size of the frame. The metric was derived from eye-tracking work that Route has undertaken since 1995. Maximum visibility distances are not currently made available to users.
In February 2017, an eighth year of respondents was added to the data set. This took the total sample size to 27,016. In yearly intervals, a new batch of respondents will replace the earliest year’s respondents, keeping the sample fresh and at a level size.
A respondent can only answer the question that they have been asked themselves, so new questions can’t be transposed onto older respondents. The sample number will therefore relate to the number of people that have answered it.
There are two Route questionnaires used to capture information about respondents. Both are available for you to download on the website.
The present data is based on fieldwork conducted between September 2011 and August 2016.
Visibility adjustment is the factor which, unique to the properties of each poster site, defines how likely you are to see a poster.
Generally speaking, a site which is: larger; illuminated; dynamic; next to slow-moving audience flows, placed at a perpendicular angle to an audience flow rather than at an angle- will have an increased likelihood-to-see factor.
What is the multiplier used for vehicle audiences, to represent average passenger numbers- because cars can carry more than one person?
The multiplier used is 1.4, but this is a national average; the figure changes across the regions as appropriate.
The impacts for individual frames are reported but they are not statistically robust.