A brand new key

James Whitmore, managing director of Route, warns of the danger of combining complex data.

There is a seesaw in the media playground. On one side sits creativity, judgement, knowledge and experience. On the other sits mathematics.

The dynamic has shifted occasionally over the years. One would say that the long-term trend has been for the right brain (creativity) to dominate the left. There have been fleeting moments when the equilibrium has been disturbed but there haven’t been any really wild shifts.

That is, until now.

After a period of pussyfooting, there is, increasingly, evidence that the siren call of data is seducing more than just the visionaries and early bullshitters. It’s more than a feeling. It’s everywhere. Take two random examples from the past week.

Sky IQ’s recent survey suggests that three fifths of agency respondents do not agree that creativity is more important than data in making a successful TV campaign.

ABC’s Interaction 2014 event conducted a poll of the auditorium, which suggested that by two-to-one, “big data” management outscores reaching the right audience as the key concern of our industry.

Exciting times, scary times, changing times.

Let’s start with a caution. It is often easier to see the comedy of error in other areas of life.

It is reported that in personality tests, Paul Flowers the Crystal Methodist, outperformed the other candidates for the non-exec Chair at the Co-operative Bank. It cannot be that the selection board ignored the merits of the competing track records, CVs and interviews in favour of a “neutral” set of statistics. Can it?

In their new book, “Where Historical Figures Really Rank”, Skiena and Ward order the relative importance of historical figures, artists and literary figures by dint of an algorithm. The computation relies on what can be found on the internet, principally from Wikipedia.

Volume of coverage is a key criterion. Would Aristotle move up from number eight if more verbose contributors had written his entry? And what would happen if Wikipedia gave due reference to his tablet use that pre-dates Steve Jobs by more than two thousand years?

In a normal experiment based on the laws of physics, one would expect the heavier character on the seesaw to prevail. But what if one player simply expands by dint of hot air? It grows and grows and has yet to pop. It is not heavier, just bigger. Do you cower in awe before the mighty blob? Or do you stand up for equilibrium whilst others simper in acquiescence?

We’ll go back to the ABC event. The audience witnessed a presentation on the mysteries of “viewability” for online ads and within an hour managed to vote the same medium as the most accountable. It was all a bit of fun and people’s guards were down, so we should not take it too seriously. On the other hand, it perhaps hints at a broader truth that for some, the presence of data is often enough in itself.

But that is to carp. There is a huge amount of creativity and experimentation in data aggregation and agglomeration. The opportunity to identify sources of information and make inspired choices about how and when to employ them is immensely liberating.

The warning for an audience research body, such as the one I represent for the out-of-home medium, is that it cannot just go out and bang data together. Livelihoods depend on a robust, consistent and rigorous currency.

In many ways, Route is already a pioneer in combining complex sets of census data with equally labyrinthine survey results. For example, we know every click of every gate of every tube station for every quarter hour of every day of every week for every month of the year and we combine them with second-by-second GPS statistics for known individuals as a start point (!) to understanding who sees what in an underground station.

On the other hand, seemingly easy wins take an age to even get into first gear. Figures such as those of mobile phone operators would appear to offer a compelling source for location information. They may, but first one must overcome obstacles such as what the data truly represent and how representative they are.

Understand what is to be found, and what is lost, in an aggregation of data points. Learn of the consistency and accuracy of the readings. And then determine the value of the gain to the industry’s understanding of how the population conducts its day-to-day business and finally judge if it is a price worth paying.

I’ll finish with a cautionary tale. One very late night about twenty-five summers ago, I found myself squashed in the back seat of an overcrowded, overheated car, circling a village near Ely in search of a “blinding party”.

We were definitely in the right place but could not hear a single beat or sight a flashing light of any kind. What was not discovered until the following day was that our trusted guide and driver had misspelt the name of the location. The similar sounding village that we ought to have sought was over twenty miles away in another part of Cambridgeshire. In research terms, the “fusion hook” – the ability to spell, was wonky.