Data that speaks volumes

Twenty years ago, I worked for the UK Government on The British Crime Survey. Each year this study would count and classify the amount of reported and solved crimes across the UK, to allow the government of the day to make decisions on criminal justice policy.

The research not only identified those crimes that were on the increase and the crimes that were not being solved, but also the crimes people were most worried about.

It was good training for identifying the best way to treat data in all its forms and taught me three lessons that are relevant today in the media world.

First, cultural behaviours can be hard to record. Crime statistics need to be representative but, they often weren’t. Some communities, for example, never spoke to the police – so crimes were never recorded. At the same time, some police forces were incentivised to focus on certain crimes, which made their statistics biased. Getting the full story, I realised, meant understanding communities and cultures, and crucially, people.

Second, new technology can skew the results. Crime statistics need to be consistent over time – but you need to account for how new technology changes behaviours. When online banking became popular, for instance, bank robberies went down. Crime wasn’t dropping, it was just easier to clone a credit card. Understanding the impact of tech is just as important in the media world today.

Third, stats can’t tell us anything about emotions. Importantly, the government policy on crime needed to be empathetic. The hard data was important but not the whole story. Our research highlighted that certain groups of people had a heightened fear of certain crimes (like older people and mugging). Such fears needed to be addressed directly and frequently.

from crime to media

In the late 1990s at MediaCom, we spent 98% of our research budget generating data on who was watching and listening to what. This was important for media trading but it never got close to viewer emotions.

In 2002, we added empathy and a greater level of understanding to our research. That meant talking to people quantitatively and qualitatively about why they watched and read things and how it made them feel.

This has even more relevance today. With so much data available, brands need to understand the narrative behind various data points, as well as be assured that all sources are representative.

We recently worked with an entertainment client. The client produced video content centrally and used it globally, with varying degrees of success. One data source said video content in this category was more influential than any other. A second source said our client’s content was under-performing against its competitors.

We used our Content Assessment Tool (CAT) to provide a third source of insight. It combines claimed metrics of content performance with measured behavioural responses. This is done through a standard online quantitative study and facial coding.

It allowed us to compare and determine how well the content scored on a rational level, and in terms of generating more unconscious, emotional reactions.

Such a multi-spoke approach can appear obvious, but unfortunately there are too many examples of brands who try to attract specific audiences without getting to know them.

human data at scale

We now speak to more than 500,000 people in 40 countries, using the same questionnaire, so our facts are consistent. The respondents remain on our panels so we can re-contact them to find out more – whether this is about one of our client’s brands or their adoption of a new technology.

We can also fuse other data sets with what we already know about our panellists, and talk to them individually to ask ‘why’. Our task is not just to quantify behaviour, but qualify it and bring it to life to connect the ‘what’ of classical media data with the ‘why’ of human insight.

Our goal is to replicate the human social experience, because the more you know someone, the easier it is to communicate with them.

Marketing is a very subjective business, full of opinion (and increasingly, algorithms), but what matters is that we don’t lose touch with real people. As we become more immersed in a data-driven digital world, the smartest brands will be those which maintain that human connection.

Mick Mernagh is chief insight officer at MediaCom