Data Mining and Data Analytics for SMBs - why should you care?

[date-stamp]When thinking about Data Mining, most people think about big companies and how they analyze billions of records of customer data to find trends and patterns. True, more big companies than small companies use data mining, but there's no rational reason for this. The benefits are there for all companies regardless of size, so being small should not be an obstacle. The advantages include (among others) increased customer loyalty and lower customer turnover.

 

Data Mining and Data Analytics - what is the difference?

Data Mining is not the same as data analytics. Data Mining is "looking for gold", while data analytics interpret data. Confused? No reason to be, imagine Data Mining as Uncle Scrooge with his pickaxe digging for big lumps of gold in the Yukon gold rush, while data analytics is someone with a map of where gold has been found and then trying to predict where more gold will be found. Such predictions can be very valuable, since they will direct future efforts in locating the best places to dig for more gold.

 

What would help you to know?

Given your current customer base and how your business is operated, what would be beneficial to know in order to have a better foundation for the decisions taken in your business? In general, everything that affects the income, expenses and efficiency of your company are of interest. And that's most of what a company does :-) One obvious place to start though is your customers. What to look for in customer data can be categorized in three main categories:

1. Sales forecasting

Look at what customers have bought from you, and try to predict what they will buy from you next time. If your business is a web shop and you have sold a capsule coffee machine to a customer, chances are that the customer will buy some coffee capsules next time (since that's a natural follow-up buy.) However, the next buy does not have to be a follow-up buy like that; this depends wholly on the nature of your business and what you provide to customers. For instance, consultancies that sell services to their customers will have to do forecasting based on their knowledge about each customer.

2. Database marketing

By digging deep in your customer data and finding patterns and relationships between your customers based on buying patterns and demographics, it is possible to create and market products that are specifically targeted towards groups and segments of your customers.

3. Market segmentation

Divide the market into subsets and try to target each subset with different products and strategies. This can be very useful if the market is big and non-homogenous.

 

What works and what doesn't

For other things than customers and buying patterns, data mining and data analytics can help you identify trends and facts that can help explain why your company is doing well or why it’s not. To know why things are going well may not seem so important (things are going well, so why change anything?); but even if things are going well, they could be going stellar. So if things are going well it would be good to know why things are going well. Chances are you have found a product, a way of selling, or a way of running your business that fits perfectly for a resourceful segment of customers. Figuring out which characteristics of your current process that attracts customers will help you to develop these characteristics to something better and even more widely adoptable, which will then reach a wider selection of customers. So instead of relying on what feels right or which strategy worked five years ago, using facts and data to develop the optimal way to interact with your customers is the way to go. Performing a survey among your customers may be the way to go, and then analyze the results to find common answers and trends among your customers, which might help explain how things are working as of now.

One example of finding an optimal way that works is the way some companies charge their customers when ordering; and at this point, the products being sold have not been produced yet. So there have been no expenses towards manufacturing or delivery at the point when money is being received from the customer. Using a build to order model like this means you do not have to have products sitting in storage, thus avoiding capital costs on these products and eliminating the period where you have paid for manufacturing and the products but have been bought. With a build to order model, you have the money to manufacture the products from the customer, so there's no need for capital in this phase of the manufacturing. And that's valuable and pretty profitable too.  Trying to find a sweet spot like this is one of the goals of data mining and data analysis.

 

How to do it?

There are consultancies that can help, there are off-the-shelf products for data mining and data analysis, and there are do-it-yourself solutions. So there are many ways to Rome here, and it’s hard for anyone without some knowledge about your company to say which path is the best one for you. As a person in the cloud-computing world, I can recommend cloud applications for those that have their data stored out there in the cloud. There are some good software-as-a-service solutions for data mining and data analysis out there, and they don't cost that much either :-)