Every day, we create 2.5 quintillion bytes of data — so much that 90% of the data in the world today has been created in the last two years alone. This statement comes from IBM and they must know a thing or two about data. But what can we do with so much stuff? Is it humanly (and commercially) possible to digest such a gargantuan level of information?
Take for example something close to the heart of many PR practitioner: web analytics. Only a few years ago we were just happy with the number of hits a page received. Then we realised that this was totally meaningless and we moved into unique visitors, then even more data became available and our company now regularly produces reports with complex correlations between traffic sources and other indices freely available from tools such as Google Analytics. Matters become even more convoluted when advertising campaigns are involved, while the advent of social media has increased the available data exponentially.
When I studied advanced statistical analysis, in between rowdy university parties and lectures, I remember that the very first thing that was hammered into us was the need to become aware of spurious correlations. With data of any kind you can easily fall into a trap of putting two and two together and literally making five. To an extent poor data analysis mirrors what may happen with poor market research. Everyone knows you can easily create an opinion poll that is structured to tell you exactly what you expect to know. If you have access to vast quantities of data you could be equally selective and simply push for the correlations that may appear to be relevant but which are in fact totally spurious. Scientists take great care to avoid this pitfall, but how many companies out there take such an equally professional approach?
I was recently reading an article from another PR colleague who berated Pizza Express for continuously sending offers out by email, but apparently not making much use of the data they have gathered about their customers. I suspect the poor boffins at Pizza Express are simply inundated with data and are probably trying way too hard to make sense of why Capricorns prefer a Veneziana (do they still make them?), as opposed to Pisceans going for a Four Season instead.
So what can we do with all this stuff? Well, the first thing to do is to select only the data we really need. The second is to start with the knowledge that ultimately data only has meaning if it gets used to transform and change something, preferably for the better. So perhaps when we are asked to produce vast quantities of data for our clients we should ask ourselves the same question: what’s it all about? Is this data really going to be used to transform and improve, or is it merely a cosmetic exercise?