Transcript RFM

Figure 4
Recency-Frequency-Monetary Value (RFM) Matrix
Hi recency
Hi monetary value
Hi frequency
Example #1 RFM
Scoring scheme:
Buy interval
<1 month = 5
1-3 months = 4
3-6 months = 3
6-9 months = 2
>9 months = 1
frequency
>40 = 5
30-40 = 4
20-29 = 3
10-19 = 2
<10 = 1
total buy
>$50k = 5
$20k-$50k = 4
$10k-$20k = 3
$5k-$10k = 2
<$5k = 1
Example #1 RFM
Customer
ID
last buy*
A
05/08
B
01/08
C
11/07
D
07/08
E
01/07
F
03/08
G
05/07
H
07/08
*vs. now, 07/08
frequency
12
18
35
1
6
40
5
20
total $
6k
9k
40k
1k
5k
90k
2k
9k
RFM score
422 422
322 322
244 442
511 511
122 211
355 553
111 111
532 532
Convention
to rearr
Most
valuable
F
H
D
C
A
B
E
G
Rank
low=
recovery
Example #2 RFM
1) Data preparation
Old variable
recoded (or create a new variable)
Recency
If
Last order placed w/in past 3 months:
Last order w/in past 6 months
Last order w/in past 9 months
Last order w/in past year
Last order w/in past 2 years
then:
20 points
10
5
3
1
Frequency
#purchases over past 2 years x 4 points,
max = 20points (i.e., if #purchases x 4 >20, reset to =20)
Monetary Value
$spent over past 2 years x .10 (max = 20 points)
2) Weights (judgment)
Recency score: 5
Frequency
3
Monetary value 2
3) Multiple variables by weights and sum to get “final weighted RFM scores” for targeting good customers
Example #2 RFM
Step 1) Data preparation
The old variables are recoded (creating new vars)
Recency
If
Last order placed w/in past 3 months:
Last order w/in past 6 months
Last order w/in past 9 months
Last order w/in past year
Last order w/in past 2 years
then:
20 points
10
5
3
1
Frequency
#purchases over past 2 years x 4 points,
max = 20points (i.e., if #purchases x 4 >20, reset
to =20)
Monetary Value
$spent over past 2 years x .10 (max = 20 points)
Step 2) Weights (judgment)
Recency score:
Frequency
Monetary value
5
3
2
Step 3) Multiple variables by
weights and sum to get “final
weighted RFM scores” for
targeting good customers