WHY DATA MINING?
by David Danziger, Trajecta

A major long distance company (lets call them Company A) recently contacted me. This company offered me $100 to switch from another carrier (which we'll call Company B). I had previously been a customer of Company A, and only switched to Company B because they offered me 5,000 bonus miles on my favorite airline's frequent flyer program. Why the offers? Do these companies know something about me? Absolutely, and they know it as a result of data mining.

As Charles Morgan, the Chairman of Acxiom, stated, "The more someone knows about me, the less likely they are to send me an offer about a sewing craft kit." (1) Data mining enables large and small companies alike the opportunity to know more about me. Not in a way that's threatening, but in a way that enables them to make educated decisions regarding the marketing levers to pull when presenting me with an offer. By knowing my sensitivity to certain types of appeals, these companies can tailor their offers specifically to me.

Marketers who utilize data mining tools are also able to accurately predict my likelihood of responding to a particular type of offer. Marketers can now segment their customer base in an intelligent way, run simulated marketing campaigns, and can predict results before going into the field.

The two long distance companies mentioned above may know a bit about my demographic profile, my lifestyle, and my calling patterns. Using this information along with sophisticated data mining techniques, Company A and Company B can craft marketing strategies for me that they know will entice me to switch companies (again!). The critical piece of the puzzle for marketers of all stripes, whether direct or retail, is understanding what marketing mix will yield the best possible results in the customer universe. In the long distance phone example above, several marketing factors are involved:

1. Via what medium should the offer be made to ensure maximum response?
2. What is the appropriate price of the offering?
3. What discounts or promotional tie-ins could maximize response?
4. What product attributes should be emphasized in the offering?

In the past, answering these questions has been accomplished by looking back at historical data and applying intuition. Data mining changes that. With data mining and data modeling tools such as neural networks, companies can rapidly use powerful analytical tools to anticipate the results of various actions. For example, consider the following:

• Given a database of 100,000 names, which persons are the least likely to default on their credit cards?
• Which types of transactions are likely to be fraudulent given the demographics and transactional history of a particular customer?
• If I raise the price of my product by $.02, what is the effect on my ROI?
• If I offer only 2,500 airline miles as an incentive to purchase rather than 5,000, how many lost responses will result?
• If I emphasize ease-of-use of the product as opposed to its technical capabilities, what will be the net effect on my revenues?
• Which of my customers are likely to be the most loyal?

In fact, data mining can help solve a wide range of marketing problems for companies which collect or can acquire quality data on their current or potential customers. Today, data mining is being applied in many different markets including: telecommunications, payment systems, banking, financial services, insurance, health care, pharmaceuticals, publishing, transportation, hospitality, utilities, and non-profit.

The ability to tailor marketing efforts based on the predicted responses to these questions empowers companies to achieve their goals with remarkable precision and substantially lower costs. Who gains as a result of this efficiency? Buyers and sellers both do to a certain degree. In any free market transaction, a buyer voluntarily exchanges money for the goods and/or services of the seller. The buyer values the product or service more than the money, while the seller values the money more. The use of data mining by its very nature squeezes costs out of the process which ultimately leads to

1. lower costs to the buyer and/or
2. higher profits to the seller.

(1) Information Week, June 10, 1996, Data Dilemma

For more information, see http://www.trajecta.com/