Data Mining Business Intelligence

Visual RFM - Segmentation You Can See

Learn how to use Visual RFM to see the "big picture" of your customer base all at once.

by Alan Weber, Advisory Consultant, Management Analytics Group
(Published in DM News magazine)
Ask us!
Questions?
Just ask!
VISUAL RFM is a way to present RFM data to help marketers decide both who to contact and how to vary offers to maximize their effectiveness. It maximizes the effectiveness of both the list and the offers. It can also help decide upon the most cost-effective frequency of contact.

SEGMENTING WITH OFFERS AND MESSAGES

Maximizing list effectiveness is important but so is maximizing the effectiveness of the offer. Recent first time buyers should be followed up quickly. Small-dollar one-time buyers should receive retention offers sooner than larger-dollar multi-buyers who are likely to remain active for a longer time. More loyal buyers should receive more frequent contact than less loyal buyers.

If in fact, customer segments are different, shouldn't different segments receive different offers? For example, if a customer typically spends over $500, does it make sense to send an offer for free shipping on orders over $50? Does it make sense to send an offer for 10% off of the next order over $500 to a customer who typically spends less than $50 per transaction?

The key to maximizing the effectiveness of offers and messages is seeing and understanding why different segments are different. By changing strategy to take advantage of differences in customer buying habits, marketers can improve upon the performance in each segment. With the right offer, segments that otherwise would be unprofitable become profitable -- and the profitability of better segments can be improved.

UNDERSTANDING VISUAL RFM

Look at the sample portion of the Visual RFM sheet below. Customer segments are based upon Recency, Frequency and Monetary. If a cataloger has very different customer groups, they may need more than one sheet. For the sake of simplicity, this example is for one set of buyers based upon a company with an average order of about $50.

The segments each represent a Recency-Frequency-Monetary grouping. For example, the upper left cell is all one-time buyers who spent less than $49 during the last 30 days. Those who made two purchases and spent over $100 are in the upper right cell, and so on. The example chart below shows that only 1-time and 2-time buyers for twelve months. Typically, the full chart covers 1-time, 2-time and 3+-time buyers and recencies from four to six years.

THE NUMBERS IN EACH SEGMENT REPRESENT...

VISUAL RFM - Segmentation That You Can See
    Frequency = 1 Frequency = 2
Recency   $0 - $49 $50 - $99 $100+ $0 - $49 $50 - $99 $100+
0-30 Days   Contacts 1,350 450 360 675 338 321
$/Contact $5.45 $6.10 $7.45 $8.18 $9.16 $11.18
% Response 10.90% 12.20% 14.90% 16.40% 18.30% 22.30%
31-90 Days   Contacts 2,612 871 697 1,306 653 620
$/Contact $4.10 $4.59 $5.60 $6.15 $6.89 $8.40
% Response 8.20% 9.20% 11.20% 12.30% 13.80% 16.80%
4-6 Months   Contacts 3,811 1,270 1,016 1,906 953 905
$/Contact $3.60 $4.03 $4.92 $5.40 $6.05 $7.38
% Response 6.80% 7.60% 9.30% 10.20% 11.40% 13.90%
7-9 Months   Contacts 3,916 1,305 1,044 1,958 979 930
$/Contact $2.95 $3.30 $4.03 $4.43 $4.96 $6.05
% Response 5.90% 6.60% 8.10% 8.90% 9.90% 12.10%
10-12 Months   Contacts 4,002 1,334 1,067 2,001 1,001 950
$/Contact $2.55 $2.86 $3.48 $3.83 $4.29 $5.23
% Response 5.10% 5.70% 7.00% 7.70% 8.60% 10.50%

By looking at the responses to a past promotion, it can be determined which segments should be contacted. Counts can be determined for an upcoming promotion by adding up the number of contacts in different segments. An appropriate offer can be sent to each segment.

For catalogers, offers are usually varied with cover wraps since ink jet messages tend to have relatively little effect. A series of "We miss you" or "We want you back" offers may be sent to less recent buyers. First-time buyers may get a quick follow-up enticing them to make a second order. Offers are made to bump up the average order in lower dollar segments.

BUILDING A STRATEGY BASED UPON RFM

Once a marketer has access to Visual RFM information, immediate improvements in communication strategy usually become obvious. For example, it becomes clear that instead of sending catalogs four times a year to everybody, some groups should be mailed more often. (Sometimes the mailings use exactly the same catalog and sometimes the cover wrap is changed.) It will become clear that other groups should be contacted less frequently. The marginal segments identified by the Visual RFM analysis should probably be contacted only once during peak buying seasons.

Many catalogers do not have a strategy to convert 1-time buyers into multi-buyers quickly. However, most do have a retention strategy for less recent buyers. A Visual RFM analysis often shows how it makes more sense to work hard to turn 1-time into 2-time buyers while they are still recent and more likely to respond than it does to spend large amounts trying to reactivate older marginal buyers.

Of course, a marketer can build a solid communication strategy while using a traditional RFM model. The difference is that the traditional model is designed to emphasize a computer's abilities while the Visual RFM approach is designed to emphasize a decision-makers capabilities.

PUTTING THE PLAN INTO ACTION

Once the chart has been created for one past offer, it can be reproduced for each promotion. Counts are determined by producing the chart with only the number of available contacts and then adding results as the campaign progresses.

\Often it is just as effective to track each segment by key code in the traditional manner. The visual RFM chart is then created only on an occasional basis perhaps quarterly or annually, to determine if the marketer is still "on track."

The important thing is adding the Visual RFM knowledge to the marketer's understanding. Even when the best RFM model is used, there is still room for improvement if offers and messages for different buyer segments are not being used effectively!

Ask us!
Questions? Just ask!     [Back to How-to Resources]      [Top]
About MAG | Privacy Policy | Contact Us | © Copyright 2009 Management Analytics Group LLC. All rights reserved.