Customer Data Mining

You Can't Afford Not to Be Mining Your Customer Data

May 27, 2004

Data mining is commonly defined as the discovery or the extraction of patterns or models from sets of data. In customer data mining, the data from which patterns or models are discovered or extracted represent the business that you do with your customers, as well as information about them and the relationships that they have with you. Customer data mining is an analytic approach that solves business-critical problems; delivers bottom-line benefits; is more powerful, more effective, and more consistent than SQL and/or OLAP; and leverages your investment in data warehousing. As such, we recommend that data mining become a foundation technology or technology standard and a core skill within your organization.

Answers to Business-Critical Questions

Wouldn’t you like to know, and know with reasonable certainty, the answers to the following critical questions about your customer relationships:

* Acquisition. Which offer is each of your prospects most likely to accept, thereby turning that prospect into a customer?

* Retention. Over the next twelve months, which of your customers is likely to desert you and join up with your competitors?

* Value. You can easily calculate a customer’s historical value. But what is the future value of a customer, and what will determine how to maximize that value?

* Costs-to-Serve. What are the factors that will move your customers toward the Web for self service and away from your call center for assisted (and expensive) customer service?

* Satisfaction. What specific issues dissatisfy your customers? What are the things that matter the most for each group of customers?

Customer data mining can answer these questions. Data mining methods and technologies offer some of the most powerful analytics you can use to understand your customers and to improve their experience of doing business with you. If you’re not already doing so, we believe that it’s time for you to harness the power of data mining. Your adoption and application of data mining methods and technologies can help strengthen your customer relationships and can make yours a more profitable, more customer-centric organization.

Consumer Industries Use Customer Data Mining…

Today, customer data mining is most commonly used by companies in consumer-oriented industry segments: retail banking, brokerage, insurance, telecommunications, retailing, and catalog marketing. These companies have large numbers of customers, large target markets, and/or lots of products. Some of these companies have been using customer data mining for many years. Its methods and technologies have given them the mechanisms to understand their customer relationships (which may number in the tens of millions) quickly and consistently with a high level of automation. These companies don’t have the time, the staff, or the bandwidth to use more conventional and more personnel-intensive mechanisms such as reports, query tools, or OLAP.

…But Most Companies Can Benefit

But customer data mining isn’t just for the largest consumer companies. Smaller consumer companies and many B2B companies can also benefit from customer data mining. Simple and manual approaches to customer segmentation, cross-selling, and up-selling have high costs and low returns. Retention initiatives conducted after customers have left are really nothing more than targeted acquisition programs. Customer data mining can address all of these issues efficiently and effectively.

In addition, the move toward cross-channel marketing, sales, and service has generated huge volumes of customer behavior and transaction data that have to be analyzed in order to understand your customer relationships. As your data volumes increase, your conventional analytic approaches may not scale. Discovering patterns in very large data sets is the bread and butter of customer data mining.

Big-Time, Bottom-Line Benefits

We’ve researched the benefits that can be achieved using data mining. You’ll recognize the companies involved in the following examples:

* Sumitomo Trust and Banking Company increased its rate of customer acquisition via direct mail marketing campaigns from 1.2 percent to 6 percent by using data mining to improve targeting.

* Jubii, Denmark’s most popular Internet portal and a subsidiary of Lycos Europe, has 2.3 million visitors in a typical month. The firm used data mining for customer profiling in order to improve the effectiveness of its Web advertising. As a result, click-through rates rose 30 to 50 percent, and advertising revenue increased by fifteen percent. The data mining project paid for itself in twelve months.

* AXA Financial used data mining to recognize that 80 percent of its cross-sells came from 30 percent of its consumer and financial professional customers, so it was able to reduce costs and increase profitability through improved targeting.

* used data mining to predict fraudulent transactions. In the first six months of its implementation, this fraud detection system reduced fraud rates by 50 percent.

* BT Retail, the UK’s largest communications service provider, uses data mining to help identify and understand the drivers for satisfaction with the customer service provided by the firm.

* Marks & Spencer used data mining techniques including clustering to identify eleven core customer segments. The firm uses its segments along with additional profile analysis of its customers to ensure that the products in its particular stores are the ones customers want. As a result, conversion rates and basket value have been improved.

* Victoria’s Secret used data mining to improve the targeting in a cross-channel test marketing campaign. The campaign results demonstrated that customers who buy Victoria’s Secret products through all three of the firm’s channels--retail stores, the Web, and catalogs--spend three to five times more than customers who buy through only one channel.


Perceptions of Cost and Complexity

We’ve seen data mining in action for understanding customer relationships in large organizations that sell their products and services to consumers--airlines, banks, wireless telecommunications companies, and retailers, for example. But despite its huge potential benefits, data mining is not widely used. Just the opposite, in fact. Some of the reasons we’ve heard for not using it are:

* It’s too complex.
* It’s too expensive.
* It takes too long.
* I need a PhD in statistics to use it effectively.
* It doesn’t work.
* It’s magic.
* We’re not that sophisticated.

Some of these reasons have aspects of validity, but, really, they’re just symptoms. The root cause of resistance to data mining is that it is different. Using data mining will take change. You’re concerned that effective use of data mining will take training, software acquisition, and experience. And you’re right. But non-users might be like the folks who didn’t get ...

Sign in to download the full article


Be the first one to comment.

You must be a member to comment. Sign in or create a free account.