I wrote this for the IPA's website in 2014 - supplemental post here - but recently discovered when trying to refer to it that it apparently got lost in a website redesign (or at least, the link broke and I can't find it anywhere), so I'm re-posting from a Word version I found on my hard drive for future reference.

In the media industry, “Digital” used to have a very clear meaning. In fact, around 2010 or so, the problem was that it was too clearly defined – so much so that in agencies we were trying to avoid it being siloed off and isolated, not properly integrated with other media, and not well enough understood by non-digital specialists.

Today, it is clearly a different story. Everything is going digital.

Working in media research, this has never been clearer. BARB is in the process of the significant transition from measuring viewing on TV screens to measuring TV viewing on any device, through any delivery method. The NRS survey – the foundation of our social grades system – is being shaken to its core for not reflecting digital behaviours well enough. Even OOH advertising can be bought by day part, and updated in real time. We can buy ads to run alongside newspaper or magazine content that might be more up to date than the content itself.

Arguably, the very term “Digital” in media has become so broad as to have lost any clear definition. Meanwhile, we are watching the explosion of a new world of “data”.

As with any new buzzword, we seem to be keen to apply “data” to anything we can. Where we used to look for “research”, we now pull the data from TGI, we get dashboards full of web data, we plug the latest sales data into our econometric and attribution models. And so on and so forth.

Of course, there isn’t anything wrong with this. Its perfectly fitting with the definition of the word.

But somewhere along the way, the word “digital” came to represent something different in a media context. The important thing about “digital” in media was never in its binary nature – it was the qualities that “digital” brought specifically to media, that was unlike anything the industry had dealt with before and turned it on its head. We all know that we aren’t talking about the kind of “digital” that led to the boom of CDs in the 1980s, or multichannel TV. When we say “digital”, we really mean something different: “online.”

But when we talk about “data”, a lot of the time we are just talking about “information”. I wonder whether there is a danger that in the hype, what the experts mean when they talk about media “data” is actually something a bit different.

This kind of media data I mean has some unique qualities;

Human data – relating directly to actual people – not a defined sample, set to quotas and weighted to represent a total population (where a single respondent can represent tens of thousands of members of a population.)
Living data – representing what those people are doing now and constantly updated – not what they were doing three months ago when the fieldwork was carried out.
Trackable data – that we can use to see whether people are taking the kinds of actions that we would like them to take. (Is this the first time they have visited our website? What did they look at last time? Have they clicked on any of our ads? Did they put anything in their basket last time they visited? How far along the purchase process did they get?)

But probably most importantly of all, Targetable data, that we can use to feed into our media campaigns.

Data that we can buy impressions against – or exclude from campaigns – so we don’t waste money on people who we know are unlikely to buy our products (or telling them about products we know they have already bought.)

Data that accumulates, making our spending more efficient, our targeting more accurate, and constantly improves our understanding of that infamous half of our media budgets which we know is being wasted.

I think that potentially, agencies could have as many specialists in targetable data in 5-10 years time as we have digital specialists today. But for this to happen, we need to really understand when we are talking about information - qualitative insights, quantitative consumer research, sales and market statistics and so on - and when we are talking about real 'media data'.

Because if we can’t be clear about what sort of data we are talking about, can we really be really clear about we can do with it?

[2021 addendum - I followed up with a post on "Data-driven insights" that is along a similar theme about the overly-broad use of the word "data".]

Some Random Nerd