RecencyFrequencyMonetary

1
RecencyFrequencyMonetary

• Behavioral Segmentation of Customers

2
RFM vs. PsychographicSegmentation

• RFM
• Focus on customer behavior
• Best for predicting of future customer behavior
• Can be used primarily for existing customers
• Recency, Frequency, Monetary

• Psychographic
• Focus on customer characteristics
• Best for understanding customer behavior
• Can be used for prospects and existing customers
• Gender, Age, Income, Interests, Lifestyle

3
Recency Best Predictor of Response

• Exact Quintiles
• Step 1 Determine how to measure recency based on
  customer decision process.
• Step 2 Sort customers by most recent business
  decision.
• Step 3 Assign a 5 to the top quintile, a 4 to
  the next 20, and so forth.
• Step 4 Realize recency best predictor of
  response to a promotion.

• Hard Coding
• Step 1 Determine how to measure recency based on
  customer decision process.
• Step 2 Select exact time intervals for each of 5
  groupings (5 given to purchase in the last 3
  months, 4 to purchase from 3-6 months ago, and so
  forth).
• Step 3 Sort customers according to recency.

4
Frequency Possibly Best Predictor of Loyalty

• Step1 Use same customer behavior as selected for
  recency.
• Step 2 Decide on a meaningful time interval
  (visits to solon per year, shopping at Vons per
  month, e-mails sent per day etc.)
• Step 3 Sort customers according to frequency.
• Step 4 Assign a 5 to the top quintile, a 4 to
  the next 20, and so forth.
• Step 5 Monitor this behavior to determine
  retention patterns and possible defection.
• Step 6 Realize that frequency is not as
  predictive as recency for promotion but may be
  better at gauging loyalty.

5
Monetary Directly Related Customer Worth

• Step 1 Decide how to track spending patterns
  (total bill at resort per visit, room rate,
  dinners, golf etc. at resort per visit).
• Step2 Determine meaningful time interval for
  tracking and calculation (week, month, year).
• Step 3 Sort customers by spending amount per
  time interval.
• Step 4 Assign a 5 to the top quintile, a 4 to
  the next 20, and so forth.
• Step 5 Realize that spending amount is not as
  predictive as recency, or frequency except for
  bigger ticket items.

6
Classifying Customers by RFM Codes

• Step 1 Sort by Recency Code and assign numbers
  as before (5,4,3,2,1).
• Step 2 Sort by Frequency within each Recency
  Code (start with the 5s in recency, than 4s and
  so forth). Then assign numbers to frequency
  quintiles as before within each recency code.
• Step 3 Sort by Monetary within Recency and
  within Frequency (start with the 5s for recency
  and frequency). For example, a 551 would indicate
  that among the most recent and frequent
  purchasers this customer is at the bottom in
  terms of amount spent.
• Step 4 Generate 5 X 5 X 5 125 RFM Cells.
• Step 5 Sort all 125 cells in descending order
  (555,554,553, 111).

7
Business to Business RFM

• Step 1 Same procedures as for end-consumers but
  classification now is 5 x 2 x 2 20 cells
  (recency, frequency, and monetary respectively).
• Step 2 Sort by code in descending order.
• Step 3 Consider sending sales force to the top
  20 and direct mail to some portion of the
  remaining 80.

8
Break-Even Analysis

• Promotional offers need to be tested according to
  proper sampling procedures (to be discussed later
  in semester).
• Break-Even Response Rate
• Break-Even Index (BEI)
• Use BEI to identify winning RFM cells.
• Economics suggests promoting only to winning cells