I came across a job advert the other day from an advertising agency for a ‘Research Scientist’. Just to be clear, I was not looking for a new job having just acquired one! But it was the job title that caught my eye. They were looking for a data scientist who understood market research. Equally, they were looking for a market researcher who understood data science. Or something like that.
Maybe advertising for a market researcher on its own does not have any allure; or that everyone is calling themselves data scientists these days, and Harvard Business Review called it the sexiest job of the 21st Century? Or maybe it is an acknowledgement that just being one or the other is no longer sufficient in our fast-moving, mobile-first, media world? Or maybe they are one and the same nowadays anyway? Whatever it is, one does wonder though where these people are going to come from. Given the relatively recent hype around big data, it would need some fast-tracked Darwinian intervention to provide a ready-made talent pool of ‘research scientists’.
In my experience, a dedicated data science team sitting alongside, or interspersed with, researchers, as can often be found in market research companies today, can yield fantastic benefits – whether it is broadening the skill sets and experience of both sets of individuals, or delivering the optimum project team to clients. Collaboration really is king.
A fine example of this is with MMS in Sweden, the body that oversees the television audience measurement. There are a number of data streams from different sources and different companies providing the audience data. GfK will integrate these datasets into a single Total Video Currency through methodological consultancy and data science expertise. That really is research science in action.
Whilst data fusion is not a new concept in media research (the German industry research body, AG.MA, started developing applications in the 1970s), it demonstrates that integrated analytics of data are now laying the foundation stones for today’s audience measurement across all media. This is important because both cost and participant burden preclude designing a vast single-source study from scratch. In simple terms: optimise the component parts and weld them all together. It’s not rocket science.