Mike Weston, CEO of data science consultancy Profusion, discusses how retailers can use data science techniques to understand their consumers better
I hate the time of year when I have to buy a new winter coat. From wool to quilted, waxed, waterproof and lightweight, I am faced with a myriad choices which can give me a mild headache. This situation is not helped by the fact that half the time I don’t even know what I’m looking for. I couldn’t begin to tell a retailer what I want in a winter coat, other than maybe ensuring that it’s warm. If I don’t really know what I want, how can retailers market to me effectively?
When confronted with questions over their preferences, most consumers do not immediately know the answer. This is the market research version of ‘White Coat Syndrome’. When put on the spot, most consumers’ “declared data” is found to be flawed or completely false. Thankfully, data science is set to alter this picture.
Every day consumers interact with organisations across their websites, in store and through social media. These interactions can be used by retailers to understand their customers in ways that were not possible before.
Amongst all this information collected in stores, through market research and online, is what the consumer states that they want – their declared preferences. An individual could, for example, tell a supermarket that they’ve never been on a diet and that they would never consider going on one.
However, by analysing their social media activity it might become apparent that the customer is very interested in people who follow a ‘caveman’ diet – often liking photos of the food and recipes. It might become apparent that they’ve been asking friends on Facebook for advice on different diets and have been browsing recipe sites and blogs. By looking at this behaviour, a retailer can discover the ‘implied preferences’ of this individual.
All these separate threads of information can then be woven together using data science, building a complete view of a consumer. This data be used to inform marketing, pricing, stock and customer service.
In some cases, a retailer would be able to understand more about a consumer than they do themselves. I might not know what kind of coat I’m likely to purchase, but a retailer studying my behaviour probably could make some well-informed inferences.
This kind of knowledge is extremely attractive to retailers. It does, however, require quite a lot of initial legwork.
Before you begin to analyse your data, you have to make sure that you’ve collected the right information, in a format that is easy to access and use. How that data is stored must also be considered, along with who has access to the information and the insights it provides.
It may be rather obvious, but if data on customers is available, the staff working in the company should understand broadly what it says if they are to provide the best service. Yet surprisingly, a large number of organisations still continue to ‘silo’ data. Opening up data to all departments goes a long way in improving how a business functions.
Once you’ve collected your consumer data, the next step would be to interrogate it. The best people to do this are data scientists.
They have the expertise to combine all your consumer information – including social media interactions, online and offline retail data, demographics and declared preferences – and study it using a mix of computer science and statistics.
Retailers using a data science team, or hiring a consultancy to do the work, will find they are worth their weight in gold. Their techniques, which include clustering (creating groups based on seemingly disparate data sets) and hidden decision trees (a technique to quickly mine a huge amount of data – for example, transactional data to detect fraud), can give retailers a profound understanding of their consumers. Data literate retailers will find they have an edge over their competitors.
However, once a data science team has given you the insights, the next, and possibly hardest step is figuring out what to do with them. It can be alarming when data science reveals that long held assumptions about a customer base are incorrect and consequently, the company should head in a completely different direction.
Data science can inform every area of a business and can provide some surprising guidance about how operations could run more efficiently. It takes courage to act on some of the results, particularly when it requires a drastic re-think in direction and strategy. However, data science can tell retailers things about their consumers which they never thought possible. Businesses who employ data science now will have a significant advantage over their rivals who are stuck using traditional methods.
By Mike Weston