Route measures advertising in the public spaces. To sell outdoor ads, we need to understand where people go, how they do so and what the look at whilst doing it. Route has been measuring this since 2013.
2016 saw Route and Ipsos Connect agree a new contract for the provision of out of home audience measurement through until 2023. As part of this evaluation process we decided that it was time to upgrade our methods and modernise how we collect our data. In doing this we considered and rejected various approaches before settling on our course of choice. This required a considerable increase in investment from our underwriting stakeholders and led to the development of a world-leading approach in people tracking.
These days, there are a growing number of ways to measure people’s movements. This may be through GPS tracking, using Wi-Fi and Bluetooth data or making use of other sensors. Increasingly, many of these means are packed into modern-day smartphones. The mobile operators are able to harvest data from these sensors and are willing to make this available to third parties, at a price. With smartphone penetration reaching mass numbers this certainly is seductive. It is a potential dataset that could be informative for our purpose of measuring people’s movements when they are out and about.
Coupling a mobile and big data solution which can be used to underpin our method is certainly appealing. And, on paper at least, the positives associated with this dataset begin to stack up quite quickly. The plus points include:
- It can offer huge volumes
- It offers a quick and elegant solution, as the readings are already in existence
- The information is available in “real-time”
- It’s a passive measure so is not affected by recall bias
- The figures are collected continuously so will show where the devices move through time
- It’s mobile, which automatically means that it’s a pretty “sexy” data set
Despite these seductive attributes, we rejected folding the data into our methods. How could this be?
Quite simply, Route is a Joint Industry Committee. This means we are data-centric and ultimately, data purists. We represent the industry as a whole. We need to be confident that the quality of our data is as good as it possibly can be. We can’t be swayed by trendiness or convenience. Ultimately, we decided to collect the data ourselves.
As part of our contract evaluation process we road-tested the mobile operator data and found four key reasons which led us down an alternative route. Before outlining these, it’s perhaps worth taking a second to consider the provenance of the mobile data.
To do that, we should first think of the meat-farming industry (trust me, just go with it). The meat farmer has a prize cow which is beautifully reared, fattened and ready for the slaughterhouse. To maximise the value of her cow, the farmer knows that she wants to turn said cow into as much high-value steak as possible. However, they are also fiercely aware that in the butchering process, not everything can become steak and that some meat may fall by the wayside. What does the butcher do with that meat? They can either leave it to waste and bin it, or else they can sell it onwards as mince. The mobile operator data is, in effect, that mince. It’s a by-product of some other core process. For the mobile operators, the high value steaks are the 24-month iPhone X contracts at £65 a month, whereas the data showing where those devices ping up and connect to the network is very much a secondary consideration.
As with much “big data”, the location data was not designed specifically for the purpose that we want to use it for – for measuring people’s movement to help understand who sees out of home advertising. That’s not to say that the data has not got great value for marketing purposes, for understanding smartphone use and for geo-fenced ad effectiveness purposes etc. – merely that it’s not right for us in what we’re doing.
Upon looking at all the evidence before us, there were four core reasons why we opted against using the mobile data…
Firstly, and fundamentally, the data which is available from the mobile operators is not precise enough for our purposes. We need a solution that will allow us to determine whether people are able to see OOH advertising. This needs to work equally across the spectrum from those seeing a large ‘jumbotron’ digital screen to those seeing a small paper poster in a pedestrian precinct somewhere. As the mobile operators use cell-tower triangulation, they are able to locate the device to a range of around 150m. This is, quite simply, not precise enough for our purposes. This will undoubtedly improve in time as the take-up of 4G and particularly 5G increases (it’s estimated that 5G will lead to an accuracy closer to 5 metres). However, this will still lead to variable levels of accuracy until these antennae are nationwide. With a need for a level playing field across the industry, this again is not something that we could readily accept.
A wise man once said “Never set out on a journey using someone else’s donkey” and this very much rings true with the potential adoption of third party data into a currency. Collecting the data ourselves means we become masters of our own destiny. It ensures that we can remain independent. This means we’re not as reliant on continued availability of third party data to entirely underpin our currency (though we do still make use of third party data as inputs for our Traffic Intensity Model). Collecting the data ourselves, also ensures that we have complete control over the numbers which we produce. It means that we can collect exactly the data which we need. We are able to design it specifically to fit our purpose rather than retro-fitting and bending the currency to fit data that is already in existence. This gives us a degree of flexibility and capacity to change things if and when we require without being beholden to the data owners and their existing data formats.
Ultimately, we are in the business of selling advertising. Advertising, lest we forget, is about selling more stuff to people. As obsessed as we all are with the latest gadgets and gizmos, devices don’t buy things, people buy things. So, for us, it’s vitally important to have people at the heart of what we do. It is crucial that we know exactly who we are speaking with. Having real people involved is an absolute must. We also want to be able to understand their lives so we really need to be able to ask them questions too. In the current climate where GDRP is a concern for everyone, it’s also important that the people whose data we are using know that we are using it and what we are using it for. Being based on devices and lacking the facility to append more explicit attitudinal or demographic information to our cause, the mobile data also falls a little short.
Finally, the mobile data includes a significant amount of, somewhat opaque, data modelling to append demographic data to the devices being tracked. There are also models to extrapolate from devices to the total population. The fact that there is data modelling happening is not necessarily the issue. Rather, it’s the lack of transparency as to what processes are being applied, which we had concerns about. Without sight of these and an understanding of what is being applied and why, we find ourselves in a position whereby assessing the quality or indeed the validity of the data becomes difficult or even impossible. As a JIC and an arbiter of data quality this is not something that we could accept.
So, what are we doing instead?
We are using a new generation of meters which implement Multi-Sensor-Tracking that take advantage of Blade-runner-esque technology. They may look like a pager from 1992, but are packed full of cutting-edge sensor technology. Each sensor is gathering information on a second by second basis. They tell us where people are, when they pass Bluetooth beacons or Wi-Fi access points. We know the direction people are facing and the speed they are travelling. We can identify when they twist and turn. We know the air temperature so can tell when people are inside or out. Knowing the air pressure allows us to determine their height and whether they are going up stairs or down escalators. These measurements all enable us to locate participants no matter where they are, even when they are out of GPS range or cellular. We can pinpoint participants to an accuracy of 2 metres (75 times more accurate than the current mobile operator data and 2.5 times more accurate than the best-case when 5G is in place in case you were counting).
This all means that we can track people whether they are inside, outside, underground, overground or wombling free.
With more readings to hand, we have more scope to improve the accuracy of our current models. After a year in the field, we are beginning the process of incorporating new information from the new MST devices. The first environment to benefit will be rail. Looking at the initial MST findings for Paddington station we now know:
- That the average time spent in Paddington station is 19 minutes 14 seconds
- That 8% of this time is spent walking purposefully from A to B, 29% is spent waiting around and the remaining 63% wandering about or shopping (wending).
- That 71% of visits to the station involve switching levels (via stairs or escalator), so this may be going down to the tube or up to the shopping areas.
- That on average people walk for 206 metres per visit to Paddington
This new data will first take effect in Release 28 of Route which will be published in September 2018.
So, to summarise, Route rejected the siren call of mobile operator data, despite its offering huge volumes, quick turnarounds and being readily available. Instead, we’re collecting our own bespoke data. We’ve upgraded our technology. We can track people above and below ground, inside and out to an accuracy far greater than any other means available. We’ve prioritised data quality and gone for rigour over something de rigueur. We’re favouring transparency in our method over black box solutions to ensure data validity
In short, we’re placing our core JIC values at the heart of everything we’re doing. This is to ensure that we continue providing accountable, transparent and objective evidence of who sees OOH ads to the industry.
This piece has been adapted from a presentation originally given by on 5th December by Euan Mackay at the annual Media Research Group (MRG) Conference.
Below you will find a link to download a pdf version of the presentation, complete with the full presentation transcript.