DATABASE SCREAMS!
The DMSI e-NEWSLETTER... Rants and raves from the database pros at DMSI.

In This Issue:

Dear Friends,

Reality - "where the rubber meets the road". A colleague of mine often used this expression when lamenting the difficulty of executing sound practices - that only exist on paper. Such is the case this month with our discussion of two database marketing concepts: RFM analysis, and Lifetime Value analysis. Read on for real world issues in implementing these popular techniques.

We also highlight new Advance Householding features in our increasingly popular WiseGuys Marketing Software program - for both consumer and B-B application. As always, please let us know how we can fit into your organization's Database Marketing plans!

Bruce Gregoire,
President and founder,
Desktop Marketing Solutions, Inc (DMSI)

Adjunct Professor, Marketing Information Systems
Johns Hopkins University Graduate School

Why not RFM? 4 Most Common Objections to using RFM Analysis
Do these objections to RFM Analysis sound familiar in your organization?

1. "We mail the whole file - we can't afford not to". Generally the fear here is that RFM might screen out important responders.
2. "More than half of our buyers are one time customers - RFM can't handle that". The implication here is: why score one time customers if they all end up with the same Frequency score?
3. "We have too many exceptions in our selection process to use a generic RFM approach". This seems to be a variation of the time-honored maxim "we are unique".
4. This from a veteran cataloger: "Most of us are really people with merchandising passions - we like finding products with huge margins. Performing RFM is just not a high priority".

1. Go ahead and "mail the whole file" (at least some of the time). But use RFM in three important ways:

   A.   Embed each customer's RFM score as part of their source code, to determine later (with back end response analysis) which of your RFM cells are the highest responders.
   
   B.   Use RFM to version the message in your mailing. For instance, lapsed customers with a low RFM score should receive a different message than your high scoring customers (perhaps including a discount to win them back)
   
   C.   Between your "mail the whole file" campaigns, you can add another more targeted mail campaign. This would be sent to selective segments based on their RFM scores. For instance, you may want to move your near-Gold customers up to a Gold status.

2. One time customers can still be scored using RFM. First, perform Recency scoring. Then, if 60% of your customers ordered just once, and you are scoring using quintiles, break the 60% up into 3 equal segments of 20% each. But give a "tie-breaking" higher frequency score to the customers in segment who are more recent.
   
   
3. Deal with the exceptions in your customer data head-on. Use query filters to screen out "noise level" in your data, before performing RFM scoring.
   
   
4. To address the lack of priority for RFM in your organization, demonstrate the savings from better targeting. Most execs understand that savings in postage and printing go right to the bottom line. You might also explain the 80/20 rule: the importance of targeting the 20% of your customers that yield 80% of your sales.

Learn More: To see how these and other sticky RFM issues are addressed by our WiseGuys Marketing Software - Click here
Return to top5

MOM and Ecometry users: WiseGuys training planned for October in St. Paul.
DMSI is planning to host a 1 day WiseGuys Marketing Training program in St. Paul, MN in late October. This hands-on training will cover both the theory and practice of RFM and Lifetime Value analysis, using WiseGuys on a workstation. Only $295 per student - great for users of Mail Order Manager and Ecometry who need extra database marketing capabilities.

Learn More: Check the DMSI News section of our web site as details develop.
Return to top5

Risk Factor influence in LTV formulation

Customer Lifetime Value Analysis (LTV) is one of those marketing topics that nearly all marketers talk about but only a few measure. Instead, mass marketers spend extraordinary advertising dollars on acquisition (e.g. Super Bowl ads) rather than retaining existing customers. The question always arises: Why? Isn't LTV the best way to truly measure customer loyalty? Let me take a shot at explaining why LTV measurement is scarce, discuss one particularly difficult obstacle, and provide a tool for you to measure your own LTV.

First a definition: Arthur Hughes (author of "Strategic Database Marketing" and a frequent DMAW presenter) defines LTV as "the net present value of the profit that you will realize on the average new customer during a given number of years."

1. Average customer spending rate: if it goes up, LTV goes up.
2. Customer retention rate: if it goes up, LTV goes up.
3. Variable cost (cost to service a customer): if it goes up, LTV goes down.
4. Acquisition cost (to attract 1 new customer): if it goes up, LTV goes down.
5. Discount rate: if it goes up, LTV goes down (only slightly).

By far the most controllable variable for a marketer is your retention rate. It turns out it is also the most influential on LTV. Then why hasn't this become a mainstream tool? Perhaps marketers don't want the headache of measuring the variables for their own organization. Perhaps marketers are generally creative types and not quantitative by nature. More likely, marketers may not have a database that can yield the data for the 4 variables that they need. Or, getting to the point of this discussion, they may intuitively sense that other real-world factors, such as Risk Factor, can muddy the LTV waters.

Risk Factor: A modifying factor that is often overlooked in the LTV formulation is Risk Factor. As Hughes points out, the customer LTV of an enterprise is subject to serious risks, including interest rate fluctuation, product obsolescence, and new competitive products. But how can risk be measured?

A study that appeared in the May 2003 edition of Harvard Business Review looked at this issue from a different perspective. It introduced the concept of Risk Adjusted Lifetime Value (RALTV). RALTV deals with the problem of unpredictable customer behavior by following the practices of sophisticated investors in stocks, whose prices fluctuate in unforeseen ways.

Since customers can be considered a risky asset, this same type of analysis can be applied to a "portfolio" of customers as well. The model is similar to that of investment. Before a firm decides whether to acquire a customer (at a certain acquisition cost), it must first decide whether the addition of the customer will have the desired effect on the riskiness of the portfolio.

Conclusion: If you are familiar with the "beta" formula for the valuation of stock, you may find the Risk Adjusted Lifetime Val