In 2015 one of the topics we would like to discuss more often on this blog is data mining.
You have all heard about Big Data, data mining, business intelligence ; but what those terms really cover remains mysterious for most of us. I guess there is also a desire to keep these terms opaque and fuzzy so that they actually seem complicated and reserved to specialists. But is it really the case?
Everyone can do data mining
What I do believe is that everyone is able to perform data mining in his/her own company. I don’t believe that so-called experts (especially if they are external) can do a great job at data mining alone. They may have some specialized technical skills to extract the data, but the crucial part remain the analysis. The analytical part of the data requires business intelligence, in other words in requires that someone with knowledge of the business have a look at the data to interpret it. If you are a regular reader of this blog you may also remember what I wrote about how crucial this analytical part is. Too often (especially in big corporations) I’ve seen people with little or no practical business knowledge trying to give sense to the data. The consequence was that correlations were made between events that were purely accidental, conclusions were drawn that were completely erroneous.
What can I do in terms of data mining?
All companies have tons of data at their disposal : it can be customer data, sales data, … you name it. Most probably this data is stored in different files (despite all the CRM, ERP on earth, experience shows that firms are still in love with Excel by the way). Either one file is rich enough to come to a conclusion or, most probably, you’ll have to merge data from different files to come to interesting conclusions.
One concrete example
Retailers often ask us a simple question : who are my customers? There are very different ways to answer to that question and you’ll have to indeed provide many different responses to get a precise idea of who they are. Let’s start with something simple. Have you ever wondered how big your captation area is where your customers are living. The size of the captation area will give you insights on how attractive your store is. Obviously, the more attractive your store, the longer the distance covered by customers to visit it will be. If your store is a “destination store”, it may be interesting to represent on the map where your customers live. There is a very good online tool for that which is called geobatch.
What do I do with a map of my customers ?
Once you have this map the next question arises : what do I do with it. Here comes the crucial part. You need to understand the business (it’s called business intelligence) and you also need to understand customers’ behaviors (but I guess business experience will bring you anyway this customer intimacy). What you could see for instance is that most customers live within 1 km of your store (it’s a pretty classical setting). But filtering this data by levels of spending would perhaps let you see that your most profitable customers come from a tiny area 20 km from your store. If you have records of all purchases by customer, you may even see that they spend their money only on Saturday’s. Wouldn’t it be great to understand why your shop is so attractive to people living in the same area and what you could do to make their Saturday’s shopping trip even more enjoyable?Tags: consumer behavior, sharing economy