Unravelling the Data Mining Mystery – The Key to Dramatically Higher Profits

June 5, 2009

Data mining is the art of extracting nuggets of gold from a set of seeminngly meaningless and random data. For the web, this data can be in the form of your server hit log, a database of visitors to your website or customers that have actually purchased from your web site at one time or another.
Today, we will look at how examining customer purchases can give you big clues to revising/improving your product selection, offering style and packaging of products for much greater profits from both your existing customers and an increased visitor to customer ratio.
To get a feel for this, lets take a look at John, a seller of vitamins and nutritional products on the internet. He has been online for two years and has made a fairly good living at selling vitamins and such online but knows he can do better but isn't sure how.

John was smart enough to keep all customer sales data in a database which was a good idea because it is now available for analysis. The first step is for John to run several reports from his database.

In this instance, these reports include: repeat customers, repeat customer frequency, most popular items, least popular items, item groups, item popularity by season, item popularity by geographic region and repeat orders for the same products. Lets take a brief look at each report and how it could guide John to greater profits.

* Repeat Customers – If I know who my repeat customers are, I can make special offers to them via email or offer them incentive coupons (if automated) surprise discounts at the checkout stand for being such a good customer.

* Repeat Customer Frequency – By knowing how often your customer buys from you, you can start tailoring automatic ship programs for that customer where every so many weeks, you will automatically ship the products the customer needs without the hassle of reordering. It shows the customer that you really value his time and appreciate his business.

* Repeat Orders – By knowing what a customer repeatedly buys and by knowing about your other products, you can make suggestions for additional complimentaty products for the customer to add to the order. You could even throw in free samples for the customer to try. And of course, you should try to get the customer on an auto-ship program.

* Most Popular Items – By knowing what items are purchased the most, you will know what items to highlight in your web site and what items would best be used as a loss-leader in a sale or packaged with other less popular items. If a popular product costs $20 and it is bundled with another $20 product and sold for $35, people will buy the bundle for the savings provided they perceive a need of some sort for the other product.

* Least Popular Items – This fact is useful for inventory control and for bundling (described above.) It is also useful for possible special sales to liquidate unpopular merchandise.

* Item Groups – Understanding item groups is very important in a retail environment. By understanding how customer's typically buy groups of products, you can redesign your display and packaging of items for sale to take advantage of this trend. For instance, if lots of people buy both Vitamin A and Vitamin C, it might make sense to bundle the two together at a small discount to move more product or at least put a hint on their respective web pages that they go great together.

* Item Popularity by season – Some items sell better in certain seasons than others. For instance, Vitamin C may sell better in winter than summer. By knowing the seasonability of the products, you will gain insight into what should be featured on your website and when.

* Item Popularity by Geographic Region – If you can find regional buying patterns in your customer base, you have a great opportunity for personalized, targeted mailings of specific products and product groups to each geographic region. Any time you can be more specific in your offering, your close percentage increases.

As you can see, each of these elements gives very valuable information that can help shape the future of this business and how it conducts itself on the web. It will dictate what new tools are needed, how data should be presented, whether or not a personal experience is justified (i.e. one that remembers you and presents itself based on your past interactions), how and when special sales should be run, what are good loss leaders, etc.

Although it can be quite a bit of work, data mining is a truly powerful way to dramatically increase your profit without incurring the cost of capturing new customers. The cost of being more responsive to an existing customer, making that customer feel welcome and selling that customer more product more often is far less costly than the cost of constantly getting new customers in a haphazard fashion.

Even applying the basic principles shared in this article, you will see a dramatic increase in your profits this coming year. And if you don't have good records, perhaps this is the time to start a system to track all this information. After all, you really don't want to be throwing all that extra money away, do you?

About The Author

Steven Chabotte is president of Big-Web Development Corp, specializing in the development of email productivity and marketing tools for the web. Steven can be reached at or you can visit our websites at or