Mining Your Customer Files

Many catalogers, especially smaller and medium-sized ones, are seriously challenged when it comes to developing a stronger revenue and profit stream from their customer lists. The following challenges are endemic to all catalogers in working their customer lists:

Knowing which are the best customers (i.e., the ones with the highest lifetime value (LTV) and those most likely to respond to the next mailing);

Knowing how to build customer loyalty without having to buy it with discounts, premiums or extensive (and expensive) point programs;

Knowing when and how to reactivate those once-loyal customers who you haven’t heard from in some time;

Knowing how to reduce customer attrition and turnover; and

Knowing how to use technology to build personalized communications aimed at targeted segments of the customer list.

Whether you prefer to call it loyalty marketing, relationship marketing, one-to-one marketing or personalized marketing, understanding your housefile customer list and maximizing sales and profits depends on your ability to segment and manage the list, as well as to build a customer contact strategy.

ALL CUSTOMERS ARE NOT CREATED EQUAL

As a generality, the customer list will outperform a list of non-buyers or rented names by a factor of 2x, 4x or as much as 10x. The chart on the opposite page helps explain the tremendous leverage that the buyer list exerts on catalog economics. Most catalogers understand that prospecting costs money, while the buyer list produces profits to pay shareholders, finance the growth of the business, and pay overhead and general and administrative costs.

Buyers have varying degrees of loyalty and some are more apt to be repeat purchasers from the catalog. In general, you can expect higher response rates and higher average order values (AOVs) from this group, especially if they are recent and repeat buyers.

Naturally, there are a number of gradations within this chart, but a hierarchy is always present. The chart displays the simple concept of RFM or recency of purchase, frequency of purchase and lifetime monetary value of purchases.

Let me demonstrate this concept with an example. Several years ago, when working on the circulation plans for a medium-sized catalog client, we were able to go into the catalog’s database and quantify the customer list RFM (recency, frequency and monetary value) segments and use them for test mailings. A simple segmentation (not unlike the one in the chart) was mailed with the following results:

• The top customer segment pulled 32 percent, with a $69 AOV or $22.08 sales per catalog mailed.

• The second best customer pulled 17 percent, with a $63 AOV or $10.71 sales per catalog mailed.

• The third customer RFM segment pulled 7 percent, with a $57 AOV or $3.99 sales per catalog mailed.

• The lowest customer segment pulled just slightly better than outside rental names with a 2.5 percent response and a $53 AOV or $1.33 sales per catalog mailed.

• The best outside rental lists, the continuations, pulled just over 2 percent, with a $52 AOV, while new test rental lists were below these numbers.

We see this “stair step phenomenon” occurring over and over again with customer responses. Understanding this principle is critical to working the buyer list harder.

THE DATABASE DRIVES CUSTOMER MARKETING

As catalogers track customer orders and maintain all-important purchase history, they can differentiate between poor, average, good and great customers. An important starting place in segmenting customers is to build a simple RFM (recency, frequency and monetary value) chart for all buyers. This chart will be different for every mailer as each catalog has different price points affecting the monetary range in the chart. Frequency of purchase will also vary: An office or computer supply catalog might get six or more purchases a year from good customers. A gift catalog might get one order per year.

The chart on page 64 (Visual RFM Segmentation) shows ways to segment a customer list. For each customer cell in this hypothetical catalog, we see the number of names or pieces mailed, percent response and sales per catalog ($/Book). The best customers are those in the upper right hand corner cell—where the customers have purchased in the last 30 days (recency), have made the most lifetime purchases (frequency) and spent the most lifetime dollars (monetary) with the catalog. The worst customers are those in the lower left hand corner cell—where they haven’t purchased for over 60 months (recency), have only made one purchase (frequency) and have spend less than $50 lifetime (monetary) with the catalog. With this information, we can build a ranking or response expectation of all of the catalog’s customers. We have built a data mining tool to help improve campaign results.

HOW TO USE VISUAL RFM SEGMENTATION

There are some wonderful, practical uses of this chart in a data mining or relationship marketing program.

Determine how deep one can mail in the next program. For example, by knowing the breakeven that will produce a 18 percent contribution to overhead and profit, we can mail those cells that meet our criteria and not mail those that don’t. If $3.00 sales per catalog is our breakeven goal to produces a 10 percent pre-tax profit, for example, we can draw the following “line in the chart.” You may elect to mail some of the marginal cells that are close to the $3.00 sales per catalog goal.

Identify problem cells that need specialized or personalized communication. Two problems come immediately to mind in looking at the chart:

• Inactive buyers who have not made a purchase for 12 to 18 months. These customers need a special reactivation message; and

• One-time buyers whose purchase was below $50.

A further study of the database may give you further hints at the distribution of the monetary level of buyers with only one purchase. A customized message is critical to these people to improve their monetary level on their next purchase.

Identify cells that should receive additional mailings. My experience has been that most catalogers under-mail their customers. It makes a great deal of sense to re-mail those cells that have the highest RFM values. A good industry rule of thumb is that re-mailing the same catalog or one with minor cover changes will pull 50 percent to 60 percent of the original mailing. If your breakeven is $3.00 revenue per catalog, it is possible to find a great number of cells that are expected to generate more than $6.00 sales per book.

Thank your good customers. When is the last time you sent your best customers a note thanking them for their business and loyalty? This RFM segmentation can identify them immediately.

Consider a loyalty program for the customers who give you the best long-term lifetime value. The more unique your catalog’s products are, the less a loyalty program is needed. In fact you may be incentivizing customers (or spending dollars on customers) who would continue to purchase from your catalog without any special loyalty program. The more that your catalog’s products are a “commodity,” the more a loyalty program is needed to keep solid customers from buying elsewhere. The database can help you identify your best customers to whom a loyalty program should be mailed. Remember it is difficult to test a loyalty program. You either have a loyalty program or you don’t. Think about your “exit strategy” up front, not after you’re in the mail.

Today’s computer technology has made building a catalog database more cost-effective and user-friendly. Further, applications in personalized printing, in-line message affixing, etc., allow catalogers to personalize the message and take customer communication to a new level. One word of caution—if your tracking of source codes and buyer information is not at the 90-percent to 95-percent level, forget data mining and straighten that out first.

Jack Schmid is president of J. Schmid & Associates, a catalog consulting firm in Shawnee Mission, KS. He can be reached at (913) 385-0220.