Retro Motor Company: potential problems with data mining

Company Overview:

Retro's product line consists of a mid-sized sedan (Impression), a
minivan (Express), a small sport utility vehicle (Altitude), and—the
most recent addition—a sporty coupe (Vortex) available with a hard or
soft top. The company's first vehicle, the Impression, was extremely
successful. Sales were high, and Retro developed valuable brand equity.

Over the past few years, however, a variety of trends and issues have
arisen that Retro's executive board is particularly concerned about.
Sales for Retro's Impression model have leveled off, and although sales
for all Retro vehicles have increased over the past years, the sales
numbers have not met the company's projections.

Project Scenario

Retro Motor Company blends the classic look of the past with the
technology of today.

Established in 1991, the St. Louis–based manufacturer is well known in
the automotive industry for the quality and ingenuity of its designs.

In 1993, Retro unveiled its Impression vehicle, and three other popular
models soon followed: Express, Altitude, and Vortex.

You were recently hired as manager of analytical services for Retro.

Working with an outside data mining consulting firm, you have analyzed
processes and problems in various departments. Your goal has been to
ensure that Retro operates more efficiently, maintains quality, and
increases sales.

Retro's executive board is pleased with your results and wants you to
apply data mining techniques to three other challenges facing the
company.

First, many customers are unhappy with Retro's platinum warranty
service.

Second, Retro's advertising campaign has not reached the customer market
and has not increased sales as expected.

Third, Retro wants to create predefined options packages for its car
buyers.

You will evaluate how data mining could help Retro address the first
problem. Then you'll determine how to manage this data mining project,
including its staffing needs and ethical, privacy, and legal issues.

Using the findings of Apex consulting, you will also interpret the
results and potential solutions of data mining techniques as applied to
Retro's three problems: the warranty service problem, the customer
market problem, and the options package problem.

In addition, you will assess the infrastructure changes needed to
support these data mining projects.

Finally, you will present and justify your findings and a data mining
implementation plan to Retro's executive board.

As manager of analytical services, you will demonstrate how data mining
can help Retro stay competitive and do business more intelligently.

My role

In one of your recent data mining projects for Retro's marketing
department, I worked with an outside data mining consulting firm (Apex
Decision Support, Inc.) to help determine the most profitable group for
a direct mailing campaign. The costs of the campaign, including the
consulting fees, were lower than previous campaigns, but the response
rate was higher. Pleased with the results, the marketing department and
the company's executive board want me to use data mining to help analyze
the other business problems across the company's departments.

Memo from Jack Holsey, a member of the executive board and director of
business operations at Retro (and the person I directly report to),
sends me an e-mail message about these problems and what he expects from
you and your data mining capabilities.

To: Manager of analytical services

From: Jack Holsey

Subject: Your New Assignments



Retro faces a variety of challenges if it is going to compete
successfully in the new millennium. We need to learn more about our
customers and how to better satisfy them. After being in business for
more than a decade, we have a great deal of customer and operations
data, but we need to improve our efforts at turning it into useful
information that can help us make better decisions. I know that you and
others in your department are exploring the use of data mining to
enhance analysis of our data.

Members of the executive board have identified problems that data mining
might help solve but are largely unfamiliar with data mining techniques.
We'd like you to

evaluate how data mining could help Retro address these problems

assess the staffing, management, and infrastructure needs of potential
data mining projects

help us interpret the results of various data mining techniques so we
can find optimal solutions to our problems

make recommendations to our executive board so that Retro improves its
decision-making

Specifically, I want you to help with three key problems facing Retro
Motors:

The service division needs help in understanding why many customers are
unhappy with our platinum warranty service program.

The marketing department needs assistance in figuring out why customers
in the market are not responding to a nationwide advertising piece.

Retro wants to determine the best predefined options packages for
customers purchasing new vehicles.

Problem: Look at the problem of dissatisfied platinum warranty
customers.

Brian Nichols, Retro's service division manager, is very concerned about
this. Customer response card and e-mail response data, as well as
service attendant anecdotes, indicate that many people who have
purchased the platinum warranty for their cars are unhappy with this
service option.

You have customer data (including age, address, date of purchase, and
model purchased) for the 540,000 purchasers of the platinum warranty. In
addition, you have warranty satisfaction levels for the 295,000 people
who sent back the customer response cards and responded to service
e-mails. You also have Retro's automotive service data (type, cost,
length of time for service or repair). You have lots of data about these
platinum customers, but what really separates the satisfied ones from
the unsatisfied ones?

Further discussion with Jack Holsey, Retro's director of business
operations, reveals that Brian Nichols is not familiar with data mining.
You decide that a well targeted memo to Brian will be a good starting
point for this data mining project