Small Catalogs Forum: Gimme an R, Gimme an F, Gimme an M

Jun 01, 2002 9:30 PM  By

“My house file is too small for any segmentation beyond recency” or “I can mail all my buyers and get a great response” are the two reasons small catalogers give most often when explaining why they don’t use recency, frequency, and monetary value (RFM) segmentation.

Indeed, in reviewing numerous circulation plans of small catalog companies, I have found that few use full RFM segmentation. And when they do use RFM, they don’t typically do so effectively. But failure to use RFM segmentation frequently results in overmailing or undermailing specific portions of the buyer file — a waste of valuable resources.

The most powerful determinant of whether a customer will buy again is how recently he made his previous purchase. The recency chart near the upper right corner of this page shows the declining response ratio that results as buyers “age” for a hypothetical catalog company whose buyer file produced an average of $4.01 per book mailed. The most recent segment of customers, those who bought between July and December 2001, produced 1.75 times the revenue per catalog as the average for the total file. But response quickly falls below the buyer file average as the file ages, with those who haven’t bought since 1998 producing only 31% of the average sales per book mailed, or $1.24 per catalog.

When segmenting by recency, be sure to label the cells by actual date ranges (July-December 2001) rather than by regressive spans of time (0-6 months). For one, specific date ranges can encompass buying seasons. Spring buyers often have a different response from fall/holiday buyers. If one is doing a merge in March, a 0- to 6-month buyer bridges two very different groups of customers, the holiday gift buyer and the early spring buyer. If you mix multiseasonal buyers in a single segment, differential response goes undetected and cannot be used to one’s advantage. In selecting calendar periods, chose those that reflect seasonality specific to your business — for example, spring outdoor decor buyers vs. fall indoor decor buyers.

Also, specific buyers do not migrate from dated ranges until they buy again, allowing you to control the number of mailings sent to a buyer. Customer Jack Jones who last purchased on Feb. 17, 2001, will remain a buyer in the January-June 2001 segment until he buys again. If you want Jack and his peers to receive three catalogs during the year, you will mail the January-June 2001 segment three times. But if you use regressive time periods, you won’t know when Jack migrates from segment to segment, and therefore you cannot control the number of contacts to him.

Frequency is the second most powerful indicator of response. For our hypothetical catalog, the frequency chart on page 54 shows the number of buyers who have purchased once, twice, and at least three times. Those who have purchased once produced $2.88 per catalog mailed, or 0.72 times the average buyer response of $4.01. Two-time buyers produced $5.54, or 1.38 times the average. Those who bought at least three times produced $11.75, or 2.93 times the average. In making mailing decisions as your file ages, you can justify mailing multiple buyers much more frequently than one-time buyers.

Finally, buyers who have made large purchases are more likely to continue to spend at higher levels. The monetary value chart on page 54 shows the portion and response of buyers at three average order levels for our catalog. The buyer whose average order is $150 or higher responded at $7.04 per catalog mailed, or 1.76 times the average. In the $75-$150 average order range, buyers performed at $3.74, or 0.93 times the average. Finally, a buyer with an average order of less than $75 performed at $2.26, or 0.56 times the average.

You should select dollar ranges that both encompass sizable portions of the buyer file and demonstrate significant differences in response rate. This will vary depending on your average order, but mailers typically have three segments, each of which encompasses roughly one-third of the house file.

Another tip: When applying segmentation, take care to assign buyers to segments based on average order or mean order. Using “total dollars to date” will duplicate the meaning of frequency.

Combined segmentation

The chart below shows the results of combining frequency and monetary segmentation with a recency segment. Note the tenth segment: “Buyers under $10.” I have found when segmenting files for catalogs that for various reasons, buyers reside in this segment. These could be buyers who returned their purchase, buyers who are really inquirers who purchased their catalog, or file conversion errors where dollars were dropped. You should isolate these buyers and analyze response to this segment separately. If you find you have significant numbers of these buyers, you may also identify a procedural or systemic problem that you should address.

By multiplying the ratios that we established for frequency and monetary segments, we have new ratios of how, on the average, each of these segments should perform in relationship to one another. For instance, the chart below shows that a three-time, $150+ buyer should perform 5.16 times the average buyer response of $4.01 for our hypothetical catalog (multiply the 2.93 frequency ratio by the 1.76 monetary value ratio to get 5.16). Meanwhile, a one-time, less-than-$75 buyer should produce only 0.40 times the average response.

For some catalogs with small buyer files, we could not have established these relationships based on an analysis of the response rates of individual segments in any single time period. The quantities of buyers in any single segment would be too small, given that you’d want at least 100 records in a segment. But when examining the relationship over entire mailings, we’re more likely to achieve statistical significance and repeatable results.

Having established these ratios on frequency and monetary value, we can then combine them with recency ratios. Again, multiply the frequency/monetary ratios by the ratio for each time period. In the example below, the combined ratio for a two-time $75-$150 buyer is 1.28. Multiplying it by 1.75 for the time period of July-December 2001 yields a factor of 2.24. Multiply 2.24 by the average revenue per book of $4.01, and you’ll get an expected response rate of $8.98 per book. After you have established all of the ratios for your catalog, you can produce a spreadsheet with an approximation of the expected response rate for every segment.

Additional segmentation

In addition to RFM, catalogs may have other, equally important factors to take into consideration. For one, the gender of the customer may influence response. For one client, whose catalog featured men’s merchandise, past women buyers responded at only 60% of the rate of men. When we combined this reduced response ratio with RFM ratios, we could not justify ever mailing another catalog to a woman who had spent less than $50 on a purchase. Since little profit was made on any order of less than $50, the client was tempted to put a notice in the catalog that women buyers who were spending less than $50 were not welcome!

In addition to gender, the type of product (say, male-oriented vs. female-oriented, or indoor vs. outdoor), multiproduct vs. single-product orders, full-price vs. off-price, source of name, and geography have proved to be factors worth additional segmentation. But you can effectively manage only a limited number of factors in a segmentation scheme, so pick the three to five that are most influential. When your file grows to several hundred thousand buyers and you have a multitude of factors to assimilate, it is time to move on to house file regression modeling.

Based on general ratios, you can stop mailing small numbers of names from recent periods that are likely to be unproductive and mail segments in older periods that will prove responsive. You can make this a mathematical process or simply have the ratios in your mind as you make mail/do not mail decisions segment by segment coming out of a merge/purge.

Granted, this is not a precise process. But a catalog that had taken an “all or nothing” approach based solely on recency can achieve a 10%-20% lift in response rates for the same mail quantity simply by adopting RFM.


John Lenser is the principal of Lenser, a catalog consulting firm based in San Rafael, CA.

Recency

EXAMPLE FOR SPRING 2002 CATALOG
RATIO $/BOOK
July-December 2001 1.75 $7.01
January-June 2001 1.00 $4.04
July-December 2000 0.87 $3.48
January-June 2000 0.63 $2.53
July-December 1999 0.71 $2.86
January-June 1999 0.59 $2.36
July-December 1998 0.45 $1.81
January-June 1998 0.39 $1.56
1997 and older 0.31 $1.24

Buyer Segmentation

RECENCY MONETARY FREQ. RATIO COMB. (1.75X) RATIO EXPECTED RESPONSE
July-December 2001 $150+ 3x+ 5.16 9.03 $36.21
July-December 2001 $150+ 2x 4.05 7.09 $28.43
July-December 2001 $150+ 1x 1.27 2.22 $ 8.90
July-December 2001 $75-$150 3x+ 2.73 4.78 $19.17
July-December 2001 $75-$150 2x 1.28 2.24 $ 8.98
July-December 2001 $75-$150 1x .67 1.17 $ 4.69
July-December 2001 $10-$75 3x 1.64 2.87 $11.51
July-December 2001 $10-$75 2x .77 1.35 $ 5.41
July-December 2001 $10-$75 1x .40 .70 $ 2.81
July-December 2001 0-$10 1x+ ??