WHY DATAMINING? A BASIC CASE STUDY 10.14.97 by Alan Parker, CEO, APower Solutions D S *
The following case study is a true story:
I received my credit-card bill a couple of months ago and there were interest charges. Due to an oversight, I had not paid my bill in full. I pulled my records and realized that I had paid the amount on the first page of the bill and had failed to notice a small charge on the second page. I was $20 short on a $5000 bill. The sub-total on the first page was printed in a bold font. I saw a number, wrote the check, and mailed it to the credit card company. The credit card in question was a loyalty card, which generates frequent flyer points for purchases. This card allows you, in fact rewards you, to keep an outstanding balance.
I called the credit card company to remedy the situation. I was prepared to pay the interest charges but somehow I felt it wasn't my fault since I always pay my bills on time. The customer service person I spoke with was very cordial and agreed to remove the interest charges. He pointed out that I would lose those extra frequent flyer points that had already been awarded for carrying an outstanding balance. For a brief moment, I was unsure whether to keep those miles and pay the charges, but when confronted I elected to drop the interest charges. I was very happy at this point. I had called customer service and my problem had been remedied.
What happened next was fascinating. The customer service representative said "I see that you have been travelling quite a bit... have you ever considered travel insurance to cover your luggage and portable computer?" The customer service representative had turned into a salesperson. Supported by the data from my transaction history, he was able to recognize my needs and present a product offering. The credit-card company was about to close a sale that could generate $1000 in revenue over a five-year period and I had made the call!
This story illustrates several excellent business decisions by the company.
They were professional, cordial, and provided me with the excellent service I have come to expect. Additionally, they broke down the traditional functional silos that exist in most companies, which have sales and customer service as completely separate units. They realize that any customer contact is a sales opportunity and they don't have to initiate it.
The company, however, did not make the sale. The discussion followed:
Company: We have a plan that will insure your luggage and computer for up to $1000 per trip for a $10 charge per trip.
Customer: My laptop costs significantly more than that. Can you give me $5000 in insurance?
Company: Let me see. We have an executive plan for $2000 in insurance?
Customer: I am not interested.
Company: Well thanks for using....
If they offered a plan for $5000, and it was semi-reasonable, I would have signed up. The fact is that the company knew the value of my computer because I purchased it on their credit card. They had used their data to know that I was travelling but they had not looked deeper.
Data mining could have helped this company. It could have identified my particular needs. There are four key approaches to keep in mind where data mining can help in any sales process:
Use data mining to determine what the average consumer wants. Companies develop a static product set and broadcast information about the products.
2) Direct Marketing -- given the product, find the consumer.
Use data mining to determine which consumers want a company's specific product. Companies develop a static product set and only contact consumers with particular profiles, thus reducing the cost of marketing.
3) Relationship Marketing -- given the consumer, find the product.
Companies use consumer profiles to determine which products from a static set should be presented based on the needs of an individual consumer. The credit card case study above is an example of relationship marketing.
4) Dynamic Product Marketing -- given the consumer, make the product.
Use data mining to determine which products you can dynamically create to satisfy the needs of specific consumers.
Direct and mass marketing are widely used. Relationship marketing is increasing in use. Dynamic product marketing is in its infancy. In order to support dynamic product marketing, companies must have fluidity in their product offerings. A number of firms (e.g. Gateway 2000) have been extremely successful at dynamic product marketing. They allow customers to configure a computer system at their web site. They manufacture the system to the needs of the customer. The product you order may be the only one of its kind in the world. Gateway places the burden on the customer to identify the desired product. In the previous case study, the credit card company did not have this luxury. The credit card company had one minute of my time to present a product that I would be interested in. It is unlikely that I would have guided them through a lengthy sales process that I did not initiate. It is in the interest of the company to analyze _all_ the data they have about me to understand the appropriate product to present.
Dynamic Product Marketing is the wave of the future. When companies are fortunate enough to have customer contact, they must be prepared to offer the customer a product based on his/her needs and this particular product may exist only for this customer. Data mining can help companies transform their data into actionable information that will identify those customer needs.
With dynamic product marketing the previous case study would have a different ending:
Company: Is the laptop you purchased in May the one you travel with?
Customer: Yes.
Company: We have an insurance plan that will cover your laptop, a full-replacement value of $4789.34, plus $1000 for your luggage and personal belongings. This will cost you $28.45 per trip when you charge your airline ticket on the credit card. Would you like it?
Customer: Yes.
This case study provides an excellent demonstration of the value of dynamic product marketing. Future articles will look into data mining techniques for mass marketing, direct marketing, relationship marketing and dynamic product marketing.