Everywhere you look these days it would appear there is someone banging the drum and sounding the rallying cry for big data. Our ability to analyse large volumes of customer information, in order to make smarter decisions, is being portrayed as the silver bullet of the marketing world.
But before you all rush out to sign up to this new wizardry, here are a few questions we think you should consider.
Can you trust your data?
Let me give you an example, imagine a customer who is looking for a new jacket. They type in the url of one of their favourite retailers into a device of their choice. They navigate straight to jackets and spend 15 minutes looking at various products, after which they click off the website leaving a nice digital trail. Later that week they go into store, locate the jackets they were interested in and after trying them on, they select one and make their way to the check out where they pay with their card again, leaving a transactional and product data trail.
This trend, known as webrooming, is very much on the rise amongst consumers who don’t want to wait or pay for delivery and who may want try a product prior to purchase. But this presents the retailer with a problem. On the one hand an online customer journey that ended without a transaction and a store visit that resulted in a sale. Both of these packets of data could be interpreted in a multitude of different ways but without the ability to uncover the simple truth, that they are part of the same customer journey, the retailer is in danger of being mis-informed and taking actions that are invalid.
Data does not equal an answer.
Brands of all shapes and sizes right across the globe are trying to find the answer to the never ending question of how can we identify, anticipate and profitably supply our customers needs and desires? Clearly understanding the customers existing preferences would be a distinct advantage in delivering this commercial goal and big data can certainly help with that endeavour. But the tricky part comes when we try to use historical data to predict what the customer will do next. This is because at some point in the process an inference has to take place in order for an insight to be created and action ultimately taken. Someone, at some stage, has to nail their colours to the mast and say the data means this, and that is inherently flawed because we will all see the data through the prism of our own experiences and preferences.
Duncan Watts, whilst Principle Research Scientist at Yahoo, said that every time they ran bucket tests to try and develop new products and services and used their intuition, experience and knowledge to try and anticipate how their customers would behave they got it wrong. This occurred despite the vast amounts of data and expertise the team at Yahoo had at its disposal.
What versus why.
Understanding how your customers behaved historically is an important part of the marketing picture, but it is only a part. Understanding why your customers do what they do is the real money question. It’s a classic case of cause and effect with customer behaviour, and by definition big data, being the effect. So if we are serious about influencing our customers we need to understand what drives their actions. As we stand currently we are still struggling to come to terms with causation because it is inherently difficult to capture, analyse and understand all of the factors that shape how people think and feel. Without causation brands will be left with regression analysis and propensity modelling methodologies as mechanisms to determine their future activity.
In conclusion big data definitely has a role to play in the future of effective marketing communications. However as we stand, it can only present you with a partial picture of what some of your customers did historically.