The marketing agenda has become dominated by big data. And for good reason, as it is hard to ignore the heady predictions about the effect that data can have on business performance. McKinsey, for example, has estimated that a retailer using big data can potentially increase its margin by more than 60%.
No wonder that marketers are becoming increasingly data savvy, seizing the opportunity to move their profession to a world of tangible measurement, one where they can have greater certainty of the ROI of their activities. It’s a brave CMO who has not embraced the possibilities afforded by big data.
This ‘big data gold rush’ has led to a huge investment in the technology infrastructure to capture and store data but also panic that the data scientists needed to analyse big data have become members of a sexy profession and are in dangerously short supply.
The tools to access data within organisations are rapidly becoming a reality for many businesses, technology costs are falling and data sources are increasing. While this is good news it perhaps brings with it another challenge – are brands potentially in danger of reaching a plateau as the playing field levels out? And linked to this is the role of the data scientist. This highly-qualified and well-paid cadre of professionals are much sought-after but, coming as they do from disciplines such as statistics, computer science, applied mathematics, and economics, do we need to draw on a wider pool of skills to ensure that an understanding of consumers is properly reflected in the data analytics?
My view, as set out in the book, ‘Humanizing Big Data’ is that it is time for brands to start rethinking how they approach big data by recognising the role of the human at its heart. This is not only the consumer behind the data point but also the individual, with all their human brilliance and failings, who is analysing the data. Perhaps if brands are to fully recognise the value of their infrastructure investments, both these areas need a greater focus. Let’s discuss each in turn below.
The Human Side Of The Data Scientist
I was recently discussing these issues with the head of analytics for a major media organisation. She was talking about how they employ some of the most brilliant minds they can find to leverage value from their data; computer scientists and econometricians create elegant and complex models to determine how to target consumers and predict behaviours. The challenge, however, is that often the models will characterise human behaviours and predict outcomes that have no intuitive base in reality. As she said, “I know that consumers simply do not act like that”. So they are told to go away and rethink. Of course notwithstanding that we need to be prepared for counter-intuitive findings this is clearly a key challenge for marketers. How can we ensure we have the right skill sets engaged in making sense of the data?
There is also a growing body of evidence that our ability to use big data to identify patterns of behaviour and predict outcomes is not as simple as we may like to think. So, for example, we all experience apophenia, thinking we can see patterns in random data, we cannot rely on statistical significance as it errs on the side of positive given enough data points and prediction of complex systems such as human behaviour can be very tough to get right. Of course we need our analysts to have the technical and numeracy skills to explore data sets but, importantly, we also need them to have an understanding of human behaviour. This allows us to navigate a way through the data, distinguishing between, to paraphrase Nate Silver, the signal and the noise. While it is tempting to think that technology will allow us to do this and that the ‘data will speak for itself’ the reality is that this falls a long way short of what marketers need from big data.
We are starting to see a call for a greater understanding of the human side of big data analytics. Donald Marchand and Joe Peppard suggested that ‘IT fumbles analytics’ because of this very issue. They believe that alongside the IT professionals, there should be specialists in the cognitive and behavioural sciences who understand how people perceive problems, use information and analyse data in developing solutions, ideas and knowledge. They make the point very clearly that while analytical techniques and controlled experiments are tools for thinking, it is people who do the actual thinking and learning.
The Human Behind The Data Point
So how can we frame the thinking in a way that allows us to better understand the human behind the data? I believe that this requires two elements. The first, as we touched on above, is a better conceptual understanding of how humans operate. That’s not to say that data scientists do not have this but my view is that this is often a limited understanding based on implicit beliefs rather than having any scientific underpinning.
So, for example, there is a widespread belief in people as homo economicus that runs throughout much of the business world. This is the idea that we are rational agents capable of identifying and evaluating our needs and desires which they broadly translate into action. We often hold this belief implicitly rather than it being something that is overt and stated, so its validity is not considered and tested. But this belief often sets the agenda for the type of analysis which is conducted. We typically fail to then take into account other ways of seeing the world that might explore, for example, the way in which social effects shape our preferences or the way in which fast and frugal decision-making shapes outcomes.
Limiting our thinking in this way means that a key source of potential differentiation for brands is then missed, which is a huge opportunity slipping through brands’ fingers. While social scientists are getting very excited about new ways of seeing the world, marketers are in danger of missing this opportunity to engage with consumers in ways that have potential to add real value to the brand but also have the added advantage that the barriers to entry are high and therefore hard to replicate. Data scientist may be the most ‘sexy’ job but perhaps social scientist is going to be one of the most important roles that brands can employ to leverage their data sets.
The second element available to brands already exists within most large organisations – the consumer insights team. While social science frameworks are helpful to guide data analytics, there needs to be a level of ‘practical intelligence’ applied to the process and outputs. Academic social science can be a little ‘rarified’, not least because it is often developed from experiments with participants that are ‘WEIRD’ (an acronym for western, educated, industrialised, rich and democratic) and where the conditions are, by necessity, highly controlled to examine the effects of one variable at a time. This does not always reflect the messy, multi-variate, non-linear nature of the ‘real world’. Insights professionals by definition have a very good understanding of consumer behaviour and as such they are an invaluable addition to the data analytics process. Their role will not only be to help frame the questions but also to apply a sanity check to the outcomes, asking whether they ring true, based on their extensive knowledge of the way consumers act in their category.
Many have pronounced that we are now in the ‘trough of disillusionment’ around big data, the hype around it having surpassed the reality of what it can deliver. To make it out of this trough, I believe we need to start engaging with a new agenda, which is a much more human-centred approach. If brands can ‘humanize big data’ then they have the opportunity to grow in a way that offers long-term, sustainable differentiation rather than the potentially shorter-term, replicable benefits created by access to data and technology alone.
Colin Strong’s new book Humanizing Big Data is available in the US from Amazon.com and BN.com. For a special 20% discount everywhere else use discount code HBDCMO at www.koganpage.com