Data Mining

Data mining is the process of analyzing large amounts of collected data in order to summarize it into useful information. This is achieved by determining correlation and patterns in the data, which helps us to gain a better understanding of how it can be used to improve performance or outcomes. Data mining is used in a number of capacities, and it can be especially useful for companies that have a strong customer focus. Large retail chains, grocery stores, banks, and credit card companies, all use data mining to better understand their customers and tailor their offerings to suit them based on information obtained.  What information might we glean from analyzing data collected from customers and how is it collected? Information can be collected at any interaction between the customer and a business, for instance, Point-of-sale, customer surveys, customer loyalty programs, coupon usage, website usage, etc. Additionally, data can also be purchased by a company. We are able to discover an abundance of information about a consumer through these resources, for instance, where they live, their ethnicity, purchasing habits, family situation, estimated income, product preferences and so on.

In a fascinating article written for The New York Times, by Charles Duhigg entitled “How Companies Learn Your Secrets”, we are given some inside knowledge about how Target has used data collected about customers to develop a system for determining whether or not a woman is expecting a baby. This information is greatly useful for Target because, as the article explains, humans are creatures of habit who tend to follow routines in their purchasing behavior. Opportunities to gain new customers arise when these customers are going through life changes, such as pregnancy or marriage. This occurs because during these transitions customers are likely to begin new patterns of purchasing, and by being able to predict these life changes early, a company like Target is able to focus specific advertising efforts at these consumers and thereby retain them in the long run.

Various types of data mining exist; a common one in business, one that companies like Target employ frequently, is known as Market Basket Analysis. Basically, it is the awareness that when a consumer purchases one item, he or she is more likely to purchase another related or complementary item. In the case of Target and their ability to predict pregnancies; they were able to infer that if a woman began to purchase items such as vitamins and unscented lotions, there was a higher likelihood that she was pregnant. They were then able to direct coupons and advertisements for baby related merchandise at these expectant mothers. This allowed them to capture a new group of shoppers during a transitional period in their lives and then retain them as future customers as they developed new shopping habits based around their changing needs.

There are other types of data mining that can be useful in business as well, for instance regression analysis is used to identify how changes in one variable will affect another variable. In the case of marketing this can be a highly useful tool, as a business can use simple regression analysis to analyze factors such as how sales have been affected by a certain promotion or product. The reason businesses find data mining so useful, is that it can identify trends in data that could be otherwise overlooked and subsequently help the business use this knowledge to increase sales. Oracle is well known software for data collection; however, several tools also exist for smaller businesses to take advantage of this tool.

For more reading on this topic, I have attached a link to an article here from Entrepreneur.com showing several ways a smaller company can take advantage of sites like Facebook and Google Analytics to perform their own data analysis.