DMA Database Seminar
Chicago, Ill.

This was one of the best and most intense seminars I have attended. Database marketing involves a lot of number crunching, and often when people talk about it, it’s from an IT perspective using a lot of tech jargon.

Rick Courtheoux, (B.S. Yale, M.B.A. University of Chicago, M.S. in Computer Science from the Weizmann Institute of Science) who taught this course, presented it from a marketer’s perspective, breaking down thorny technical issues in a way that made it understandable and actually enjoyable for us non-technical folks.

Database marketing has been around for decades, but there are a lot of new and exciting things on the horizon. In order to be ready for these direct marketing innovations, I believe it’s important to have a working knowledge of the current state of database marketing. This seminar brought me up-to-date on how it’s done, who’s doing it and what is and what isn’t possible.

Of course I can’t relay everything I learned in the course in a few pages, but I want to share my notes with you in the hope that it will impart a fair understanding of the current state of database marketing.

What is database marketing?

A way to understand the customer and the right mix of products and marketing

• It’s information intensive
• Based on facts and information
• Long-term and strategic
• Builds up strategic capabilities over time

Benefits

• More and better information increases efficiency by reducing waste.The more we know about our customers the better able we are to reach them in an efficient way.
• Extend reach to additional names.Database marketing helps us discover who our best customers are and gives us the ability to find more like them.
• Understand the market better.Database marketing allows us to track and analyze customer behavior over time and to integrate customer and product perspectives. Database marketing helps us see how we got to where we are and where to go from here. It helps us to understand customer dynamics: how their behavior changes over time. It allows us to integrate customer and product perspectives and to track and analyze performance over time.
• Strengthen customer relationships.We’re better able to stay top-of-mind by expanding direct communications.
• Manage multiple channels.Database marketing helps us to speak to the customer the same way across all channels. Customer management system integration provides the means to recognize a customer’s behavior and treat that customer in a way befitting his or her status.
• Grow into specialized markets.Database marketing gives us the tools to understand customers’ special needs and wants and to fulfill them.

Develop a competitive advantage. Database marketing favors market-leading companies because they can afford to develop the infrastructure.

Costs

• System investments are similar to operational systems. New relationships between marketing and operational systems must be established.
• Long term planning is required.Successful database marketing programs require a considerable investment in infrastructure. Reactive companies are often not a good fit for database marketing.
• Organizational changes.Companies that are organized around products don’t use database marketing as well as those that are organized around the customer. Stability and low-employee turnover are a must in order for database marketing to succeed.

Key success factors

1. Long term relationship with the customer. In order for database marketing to be useful, you must have repeat buying opportunities. You will have to have a buying history with your customers in order to understand their behavior.
2. Closed loop of information.Database marketing requires that a dialogue happen between the customer and the company. You need to know what you said to him or her and how he or she reacted. You need to know the result and if the communication was successful.
3. Significant variable margin.Costs for database marketing are significant. You will need a high margin in order to make it pay off.
4. Specialty niche markets.Database marketing works best when you can be specific.
5. Reduced distribution stages.It’s important that you be able to sell directly to the customer and achieve fundamental economies. Database marketing in this situation can shorten the sales cycle and allow you to know when to speak to the customer.

Industry success levels based on key success factors

High

• Merchandise
• Travel
• Financial
• Hi-tech

Medium

• Retail
• Publishing
• Not-for-profit

Low

• Package goods
• Dealers
• Distributors

While database marketing is able to give us a wealth of information, there are some things that fall outside our ability to know or predict. For instance, it is difficult to determine what caused a person to buy. We can predict what channel a person will choose to use to buy, but it is not possible to know which channel influenced their purchase. An attendee at the seminar said that in the new car business, 70% of new car buyers arrived in the showroom with a direct mail piece in their hands.

The two most telling pieces of information about a person are the magazines he or she reads and the car he or she drives.

Customer relationship management

Customer relationship management is an extension of database marketing. It addresses the whole customer lifecycle and all communication channels. Good CRM requires the integration of many applications, including: call center, sales force automation, customer service interactions and e-commerce.

CRM is very complicated and few companies have been able to realize its full potential.

• CRM has to supply all information needed to handle every interaction with every customer
• Very mixed picture of success: 2/3 to 3/4 fail rate
• CRM only works for very big companies
• Citi Corp and Amex are the companies that have come closest to making CRM work

Database marketing trends

Positive trends

• Innovations driven by tech systems
• Demographic heterogeneity
• More channels
• Globalization

Negative trends

• Privacy legislation
• Restrictions on telemarketing
• Postal service economic stress
• Environmental concerns (waste of paper)

Building database marketing capabilities

• Identifying and integrating all useful data
• Management issues: costs, budgeting, ROI, good business practices
• Verifying correctness of data
• Technical issues

Types of data

• Transactional
• Demographic
• Psychographic
• Geodemographic

Of all the types of data, transactional is the strongest. That’s why it’s so important to understand how each of your customers act. When you know what they buy, how and when they buy, you can begin to predict their future purchase behaviors.

Database marketing depends on detailed information

• Event-based data
• Exact products or services purchased
• When people bought
• How often people bought
• Track inquiries about products or services
• Types of behavior: service calls, complaints, returns, payments
• Contact history: how does you customer prefer to talk to you?
• Determine the best way to communicate: right level, right medium, right mix
• Get data out of operational systems and into marketing systems
• Use telemarketing records
• Record online and Web site activity
• Acquire application form data
• Market research
• Credit and payment information

It’s important to assign each customer a unique identification number so that you can keep track of that individual. A unique ID number also helps to reference that customer back to the individual systems involved in the customer data integration loop.

Predictive modeling

Predictive modeling is very much in vogue these days, but few of us have a good grasp of what it is and how it can be used. Predictive modeling (and database marketing in general) works by looking at the past, then using that data to predict the future.

Think of predictive modeling as driving your car by looking out the rear window. If the road is straight, then you have a good chance of staying on the road. However if the road has a lot of curves, then steering your car based on where you’ve been won’t work very well. So, if you’re in an industry where things change a lot, predictive modeling might not be for you.

Most popular predictive modeling techniques

• Multiple regression
  First used by Reader’s Digest in the 60s
  Most widely used
  Used to build models of lots of types of behaviors
• CHAID
  Chi-Square Automatic Interaction Detector
  Developed by University of Michigan in the 70s
  Creates segments by repeated splitting of cases
• Neural Networks
  Mimics living systems
  Takes a lot of computation
  Very powerful, but hard to integrate into existing systems

Others

• Genetic algorithms
• Pattern recognition
• Discriminant analysis

Garbage in, garbage out

• 80% of modeling success depends on data quality and preparation

Model on one group, test on a different group

• Separate data samples

Predictive modeling allows you to score your customers and find the best ones, then use external data to find more customers like your best customers. As marketers, we have to develop ways to make yes/no decisions on who to market to and how much.

close