A presentation from the 2014 National Postal Forum by Gary Seitz, Executive VP of C.TRAC, on Recency, Frequency, and Monetary (RFM) analysis as a simple tool to help mailers.
3. What is RFM?
R = Recency
Last purchase or meaningful transaction (donation)
F = Frequency
How often there is a dollar-earning transaction
M= Monetary
The sum of all revenue generated
4. Why RFM?
Past behavior is best predictor of future
– Segmentation is quantifiable way to distinguish
between best & worst customers
– RFM is quick, descriptive way to segment a
marketing database based on purchase
behavior
– It is not a statistical or predictive model
5. Recency
Most powerful predictor of who is likely to buy
Time-frame since last transaction
“High-end” date on the database, not “today’s date”
Recency Customers Response Revenue Resp Rate Avg Sales
0-3months 38,457 1,597 $42,571 4.2% $26.66
4-6months 29,472 961 $19,274 3.3% $20.06
7-9months 21,947 422 $6,231 1.9% $14.77
10-12months 16,473 247 $2,547 1.5% $10.31
13-18months 11,648 114 $954 1.0% $8.37
19-24months 15,982 88 $341 0.6% $3.88
25+ 26,482 91 $188 0.3% $2.07
TOTAL 160,461 3,520 $72,106 2.2% $20.48
6. Frequency
Second most powerful predictor of response
How often a buyer has a dollar earning transaction
– Purchase or donation
– Exclude credits
Timeframe or Lifetime transactions
Is not number of units (i.e. 3 shipments)
8. Monetary
Least powerful predictor
Useful to avoid mailing low value customers
Sum of all revenue earned
Exclusive of tax and shipping
May be actual dollars, or an average
9. Frequency + Monetary
Impact of adding Monetary
– Who to market
– How to market
Direct Mail offers:
Loyalty programs
“Comps”
Personalized offers
Eliminating frequent “bargain shoppers”
1x
2x
3x
4x
5x+
A$ B$ C$ D$ E$
10. Getting Started with RFM
Recency
Assign 6 to 8 letter codes for time segments
appropriate for your business
Use quarters for first year: 0-3 mo, 4-6 mo
Group years (months) 13-24 mo, 25-36 mo
A = 0-3 months E = 13-24 months
B = 4-6 months F = 25-36 months
C = 7-9 months G = 37-48 months
D = 10-12 months H = 49+ months
11. Getting Started with RFM
Frequency
A simple 1, 2, 3 (3x+)
Triers ~ buyers ~ highers
1x buyers are new customers!
Give them special treatment
Convert them to 2x ~ but not too long
12. Getting Started with RFM
Monetary
Use 3 to 8 letter codes in reverse value
Accumulated or Average value
A = $10,000+ E = $500 - $999
B = $5,000 - $9,999 F = $100 - $499
C = $2,500 - $4,999 G = $1 = $99
D = $1,000 - $2,499 Z = $0 (returned)
13. RFM score
Kevin O’Leary
RFM = B2D
B = 4–6 months since last transaction
2 = 2 transactions
= $1,000 - $2,499
21. RFM score
Can be a 3-character quintile = 543
– Sort file on R descending – breaking into 20% increments
– Sort file on F descending – breaking into 20% increments
– Sort file on M descending – breaking into 20% increments
– Assign top 20% score of 5, next 20% score of 4, and so on
The segments, therefore do not reflect any specific timeframe,
frequency or amounts. Higher values (i.e. 555) are better than
lower values (313). There are 125 total segments created.
23. Why RFM?
Use it to:
– Identify groups of customers ~
segments behave differently
– Target offers by various segments
– An offer to first time “triers” (don’t over do it)
– Promote repeat purchases/donors
– Eliminate non-responsive
24. Why RFM?
Use it:
– Modeling isn’t practical or too expensive
– Small databases
– Gain a high level of understanding
25. RFM versus Predictive Models
Advantage RFM Model
Understandable Easy Black Box
Creation Relatively simple Statistician with PhD
Timing Week Up to 6 weeks
Update As needed Annual
Cost Inexpensive Budget approval
Seasonality Somewhat effective Separate models
Depth All buyers Recent buyers
Breadth 3 variables 100s of variables
Predictive No Yes
26. RFM
Stage 1 Customers are one-time
buyers with positive Expectancy.
Stage 2 Customers are multi-
buyers with positive Expectancy.
Stage 3 Customers are multi-
buyers with negative Expectancy.
Stage 4 Customers are buyers
with negative Expectancy
Expectancy = Recency or Monetary
27. RFM
Stage 1 Customers – “Triers”
Stage 1 Customers are an
appreciating asset. They are new
customers that need to be resold.
They will respond well to early
marketing that features best selling
products with instructions on how to
use them. They trust the use of
testimonials from satisfied
customers. The goal is to move
them to Stage Two.
28. RFM
Stage 2 Customers – “Buyers”
Stage 2 Customers are your most
valuable asset. They are loyal
customers that need to be retained.
They will respond well to frequent
marketing that features your newest
products and bundles. They like to
be rewarded for their loyalty, and
asked for suggestions and
testimonials. The goal is to keep
them in Stage Two
29. RFM
Stage 3 Customers
"High Value, but Falling Potential"
Stage 3 Customers are a
depreciating asset. They are
customers that need to be
reactivated. They aren't responding
to your marketing. Have they moved
or died? Are they shopping
elsewhere? Have they experienced
problems with your company? The
goal is to find out what's wrong, and
fix it if possible.
30. RFM
Stage 4 Customers
"Low Value, and Falling Potential"
Stage 4 Customers are a liability,
not an asset. They are lost
customers that need to be replaced.
They have not responded to your
marketing in a long time. The goal is
to direct resources away from them
toward the retention and acquisition
of more valuable customers. Cut
them loose.
31. RFM Best practices
Build RFM cells with significant quantities
Seasonality impacts R
– Holidays
– Campaigns
– Special offers
Profile top RFM segments
32. RFM Summary
Utilize simple RFM to segment and
understand customers
Develop a Direct Mail marketing plan based
on RFM segmentation
– Target
– Establish a contact frequency plan
– Don’t be afraid to “let go”
Track results
33. Questions, Answers & some GREEN!!!
Gary A Seitz
Executive Vice President
C.TRAC
(440) 572-1000 x140
GSeitz@ctrac.com