Jamie Out of Home

James Whitmore, managing director at Postar, bakes a fiendishly complicated algorithm cake in his attempt at measuring the audience of out-of-home media. Make sure you have a pen and paper ready…

Everything is on the move. The adverts and the audience. People don’t have a ready mental framework to describe what they have seen. Traditional approaches, à la BARB, RAJAR et al, will not work.

The solution is fiendishly complicated. I thought that I would try and give a brief overview of how to cook up a new currency using a few handy ingredients and some master chef alchemy. If you are sitting at work missing daytime TV, have no fear; it’s a bit like this.

First, take a map of the country. Turn each road, pavement and passageway into a series of connected links. They must all be to scale. Convert it to a digital file. Then take a map of each of the indoor spaces that you wish to record – shopping centres, tube stations, rail stations, airports and so on. Make these the same scale and integrate them to your digital map.

You should now have something that allows you to seamlessly follow any route by any mode of transport anywhere in the country.

Next, sprinkle on some media. Position some 360,000 advertising sites on your map. Take care to ensure that they are all correctly orientated to north. Be certain that they are accurately placed to the nearest metre. Remember to record the flavour of each sign, its size, illumination, whether it is static or digital.

For buses, you will need to create a digital copy of their routes and note the frequency of movement by time of day. Meld this to your map too.

Place these to one side for now; we’ll come back to them later.

Now, you want to catch some fresh respondents. Go out into to wild and snare 23,000 people. Give them each a state-of-the-art GPS device and ask them to carry it for nine days. That’s two weekends and one working week.

Make sure that you hunt everywhere in the country and be sure to bag a representative cross-section. You might have to look a little harder to get young bucks or rare creatures that are constantly leaping and bounding everywhere.

This will produce millions of spaghetti strands that reflect the routes these people have taken. Colour-code each one so that you do not forget who produced it and carefully lay them on to your digital map.

Now that you are about half way through, you should be able to see thick ropes of coloured spaghetti passing by the advertising sites that you had prepared earlier. Yum.

We are now going to turn our map into a model. (This will include using “algorithms” that I shan’t begin to explain. Let’s pretend they come ready-made in a tin.)

Our dish needs a bit of volume. If you look hard enough, you can find counts, that is, the number of people or cars or passengers that pass through some of the links on your map. Some of these data will be by the quarter hour by day for the entire year.

Some will be for the year in total. And many points in between. To make sure that all the counts are comparable, you will need to reduce down some of them. Others might require a bit of whisking.

Next, extrude these counts so that they populate all the links on your map. You will require some adept culinary science here. What we seasoned chefs call a “Traffic Intensity Model”. You must add a scoop or two of tinned algorithm.

Once you have created your model, you can meld your respondent spaghetti with the counts.
Take a step back and admire your work so far. You should be able to see exactly which type of person is passing along each link, when and how often.

Add further spice by recording the speed of these movements. You can get this from the GPS records. This is important as it helps to give a firm base for the final part, the topping.

Every good meal needs a scrummy topping and for our dish we have chosen cream of visibility. Eye-tracking research can determine the probability that someone sees something based on their distance from an object, the size of the object, its offset and orientation.

You will wish to update this to include the effects of movement (it is an increasingly digital age) and illumination. So that the icing does not collapse under its new weight, you will need to make sure that you are absolutely certain of all the existing visibility ingredients. Check them thoroughly and make adjustments if needs be.

To properly fix the topping to the main dish you will want to match the speed and direction of each flow of audience to the visibility cream. A few more scoops of tinned algorithm will help here.

Now put this on a baking tray together with whatever is left in the algorithm tin and whack it in the oven.

What comes out is a dish with a riot of flavours. It will reference all types of audience and place them carefully against the full spectrum of zests and tangs that the outdoor world has to offer. Delish.

You will find the ingredients at a range of stores; Ipsos MediaCT, MGE Data, University of London, Telmar and Cuende Infometrics being the main ones.

Multi-tasking to the max.