Big Data is fast becoming an invaluable tool that is changing the way we approach marketing. As a way of developing insights into consumers' habits, it is the lifeblood of our trade. But it also has its drawbacks, and no one can or should try to understand their consumers through Big Data alone.
Big Data allows us to pool the plethora of data we get from consumer's complex lives, and draw valuable long and short-term insights into our consumers across multiple media. It enables us to draw together a picture of what consumers are doing, where they are, their likes and habits, and -- many would argue -- what they might do next.
Having this data in real time means we have much more immediate data than ever before, down to a last action, click or purchase. And for media agencies like us, this is vital in terms of interrogating the habits of target consumers.
However, marketers should draw conclusions from Big Data with extreme caution, and avoid falling into the obvious traps. Categorizing consumers according to specific habits, preferences, or demographics makes us run the risk of pigeon holing them, and leaping to the wrong conclusion.
By defining people according to generic types, and making assumptions or predictions, we are forgetting one of the most fundamental facts about human nature: we are unpredictable and not always rational. People don't necessarily behave according to types in real life, or act according to their age, gender or lifestage. So why do we expect our consumers to fall relatively neatly into these categories?
This is where Big Data falls short. None of us likes being put in a box and defined, and our 'consumers' should not be treated any differently. As we discussed in the 'MediaTel Media Playground' AdWeek debate, we need to hone down on individuals rather than groups.
Take a very simplified example from the media industry: you run a banner ad for an online sofa company on The Telegraph home page every Monday for four weeks. The first two weeks you make 40 sales, followed by 20 sales on the second two weeks. If we had predicted the outcome after two weeks, we would of course have come to the wrong conclusion.
While this example is clearly unrealistic, it illustrates the point that you cannot necessarily predict people's behaviour according to how they have behaved in the past, which is what Big Data asks us to do. People are not always measurable. And we must recognise the random quirkiness of our audiences, their creativity and unpredictability of thinking, and then make sense of it all with the data analysis now available to us.
By fusing an understanding and appreciation of human nature with the logic of Big Data, we can gain much more valuable insights upon which to base conclusions.
No one can survive on Big Data alone -- the real value of big data will only ever be unlocked if you have intelligent people to help comprehend it in more real terms.