TREND-FACTORED RFM SCORES TO IMPROVE CAMPAIGN PERFORMANCE

• US 20150046223A1
• Filed: 09/22/2014
• Published: 02/12/2015
• Est. Priority Date: 08/08/2013
• Status: Abandoned Application

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Sewak, Mohit
• Application Number:
  US14/492,273
• Publication Number:
  US 20150046223A1
• Time in Patent Office: Days
• US Class Current: 705/7.31
• CPC Class Codes: G06Q 30/0202 Market predictions or forecasting for commercial activities

Abstract

A computer determines an RFM score for each of a plurality of contacts, each RFM score comprising a recency score, a frequency score and a monetary score respectively based on assigning recency values, frequency values, and monetary values of the plurality of contacts to groups. The computer generates trend factors, wherein the trend factors are based on the monetary values and timestamps of each of a plurality of transactions for each contact of the plurality of contacts. The computer generates a mean trend factor, based on the trend factors generated for each of the plurality of transactions for each contact of the plurality of contacts, and determines, by one or more processors, a trend adjusted RFM (tRFM) score based on applying the mean trend factor to at least one of the recency score, the frequency score, and the monetary score, for each contact of the plurality of contacts.

Claims

1. A method for including a trend factor in recency, frequency, and monetary (RFM) scores, the method comprising:

generating an RFM score for each of a plurality of contacts, each RFM score comprising a recency score, a frequency score and a monetary score respectively based on assigning recency values, frequency values, and monetary values of the plurality of contacts to groups;

generating trend factors, wherein the trend factors are based on the monetary values and timestamps of each of a plurality of transactions for each contact of the plurality of contacts;

generating a mean trend factor, based on the trend factors generated for each of the plurality of transactions for each contact of the plurality of contacts; and

determining, by one or more processors, a trend adjusted RFM (tRFM) score based on applying the mean trend factor to at least one of the recency score, the frequency score, and the monetary score, for each contact of the plurality of contacts.

2. The method of claim 1, wherein the step of generating trend factors for each of the plurality of contacts further comprises:

determining, by one or more processor, a first transaction and a second transaction from a plurality of transactions for a contact of the plurality of contacts, wherein the second transaction follows the first transaction; and

generating the trend factors for each of the plurality of contacts, based on a quotient of the monetary value of the first transaction and the monetary value of the second transaction, for each of the plurality of transactions of each contact of the plurality of contacts.

3. The method of claim 1, wherein the trend factors are generated based on a period of time between the timestamps of a transaction and a next transaction for each of the plurality of transactions, for each contact of the plurality of contacts, wherein the period of time for a last transaction is assigned a value.

4. The method of claim 1, wherein, generating a mean of the trend factors for each contact of the plurality of contacts further comprises:

determining a transaction and a next transaction for each of a plurality of transactions; and

generating a weighted mean trend factor as a quotient of;

a summation of a product of a trend factor and a reciprocal of a duration of time between the transaction and the next transaction for each of the plurality of transactions, and

a summation of the reciprocal of the duration of time between the transaction and the next transaction for each of the plurality of transactions.

wherein the reciprocal of the duration of time for a last transaction of the plurality of transactions is assigned a value.

5. The method of claim 1, wherein the mean trend factor is applied to at least one of the recency score, the frequency score, and the monetary score that is most relevant, based on results of previous related campaign activity.

6. The method of claim 1, further comprising:

in response to a lack of historic correlation data of similar campaign activity, applying the mean trend factor to a median weighted component of differential weights for the recency value, frequency value or monetary value of the RFM score.

7. The method of claim 1, wherein the mean trend factor is one of:

an average, a median, or a mode.

8. The method of claim 1, wherein the mean trend factor for each contact of the plurality of contacts is determined as a sum of a best fit calculated slope of a plot of the trend factors of the plurality of transactions, from a least recent transaction to a most recent transaction, and a value of 1.