ess.
The RFM concept is the following:
• Customers who ordered recently are more likely to order again than those who ordered in a less recent period
• Customers who ordered frequently are more likely to order again than those who ordered less frequently
• Customers who ordered a higher monetary value (spent more) are more likely to order again than those who ordered a lower monetary value, and it has been tested heavily in the catalog businesses.

Recency, frequency and monetary value, form the basis of database marketing.

Frequency is often a powerful predictor of response, but it is seldom as powerful as Recency. Recency is the most powerful predictor and the easiest to define. There may be alternative ways to measure frequency. One should test alternative measures to figure out which is best suited. This can be done by testing the response on these alternatives.

Recency enables the prediction of future value, while frequency and monetary value enable the estimation of the current value. The combination of the 3 dimensions (RFM) allows the combined analysis of current and future Customer value.

The sequence R->F->M reflects a decreasing series of predictive power (however, this is not always true for the ‘M’ predictor: if one tries to promote an expensive product, M has propably increased predictive power).

In case there is no monetary value attached to an interaction (e.g. a web visit or a complaint), the analysis may be limited to recency & frequency (RF analysis).

The higher the RFM score, the more probable it is for a Customer to respond to a marketing program. This fact has been clearly confirmed in practice.

Why is this fact actionable ? Because if one classifies Customers to groups according to the RFM score, she/he can expect each distinct RFM group to have substantially different response to an offer, from the rest (especially if the number of groups is limited). Therefore she/he can focus only on certain Customer groups which are expected to respond highly,

In order to develop Customer Intelligence, a business needs to be able to measure its performance in the maintenance of profitable customer relationships. Customer intelligence attempts to define customer behaviour and then look for variances in that behaviour. The business rules which apply to the Customer relationship, need to be defined first. Based on these rules relevant measurements & goals can be defined. Therefore, a business needs to systematically answer the following questions:

When is a party (an individual or a business) considered a prospect ? (define the sales pipeline stages for each Customer segment, e.g. lead, prospect, stage at which a proposal is submitted, an order is placed etc). When is a party a new customer ? (1 order, after 2 orders ?) Which is the Customer lifecycle ? Which events mark stages of the lifecycle (1st order, 2nd order, service call, billing inquiry, complaint, etc) ? When is a party no longer a Customer – when is the Customer lifecycle ended ? What is a Customer LifeCycle? Customers begin interacting with a business, and over time, either decide to continue this interaction, or end it. At any point in this LifeCycle, the Customer is either becoming more or less likely to continue interacting. If data from these interactions are captured (purchases, visits, complaints etc.) this data can be used to predict where the Customer is in their LifeCycle. By predicting that, one can focus on Customers most likely to buy, and try to "save" valuable (or profitable) Customers who have declining interest (an info-driven focused approach), instead of wasting money on Customers unlikely to continue interacting (‘blind’ unfocused approach).

In many cases the answer to the above questions (business rule definitions) is not so simple. How can one be sure that a Customer is no longer a Customer, in the retail market. In subscription based service markets the answer is probably easier to give: a party is a Customer, as long as a subscription to a service is active.

Concepts like latency, recency, RFM (recency - frequency - monetary value) are applied many decades, in order to identify the active Customers and achieve higher response rates to the Customer retention & loyalty efforts. In order to apply these techniques, one has to develop a database which stores all Customer contact history on all channels (or CTPs – Customer touch points).

Latency

Latency refers to the average time between Customer activity events (e.g. orders, use of services, visit on web sites). There are alternative ways to estimate the average time between events in order to determine latency. For example if one has captured the dates of a Customer’s orders, she/he can derive the intervals between orders: Time between 1st & 2nd order : 70 days, Time between 2nd & 3rd order : 50 days, latency can be determined to be 60 days (the average time between orders in this sequence of three orders). The trend in ‘time between events’, can also be analysed in order to evaluate the dynamics of a Customer’s behavior. An increasing ‘time between orders or web visits’ is not a good sign (this type of analysis is equivalent to frequency analysis which shall be described below). On the other hand if the event is related to customer dissatisfaction (e.g complaints), increasing ‘time between complaints ’ is a good sign. How can latency be used to develop a simple customer retention program. One has to estimate latency for a Customer group and then measure the days since last event for each customer in that group. When this measurement reaches the latency estimate or exceeds it for a specific Customer, one has to act in order to influence that Customer’s behavior. By offering a discount, this Customer is encouraged to continue interacting with the business.

Recency concept

It has been proven in practise that there is a recency effect in Human behaviour (relevant theoretical studies have been performed in the past but the empirical evidence is also sufficient). This effect viewed in the Customer behaviour analysis field, leads to the following concept: The more recently a Customer has ordered a product or used a service from a certain Business, the more likely it is she/he will purchase a product or use a service again from that Business. In other words, Customers who have transacted with a Business recently are more likely to transact with it again, than Customers who have transacted with that Business less recently.

The recency metric could be defined as the number of days/weeks/months (the scale is relevant to the business/product), since the last relevant transaction occurred (the definition of recency is key to its successful use - testing of alternatives may be needed). Central in this analysis is the dimension of time. ‘How long since a Customer event happened’, is key to understand past and predict future customer behaviour. However the exact scale of time which is relevant in each business/product has to be identified (e.g. in retail sales a second purchase may take place a few months after the first purchase, in home loans a second purchase may never take place). If the Customer Lifecycle is understood (for a specific Business / product), the recency effect can be used to produce actionable ideas. Customer contact history can be used to identify standard Customer lifecycle stages (identify types of milestone events, stages and average stage duration).

Given that a relatively more recent Customer is more likely to buy again, than a less recent one, the former has relatively higher Customer value. This is strictly true if we compare Customers of similar purchase value. We notice here the fact that we reach a comparative conclusion. Recency is widely acknowledged as the most powerful predictor of future behaviour. The future behavior under analysis may be any of the following: Subscriptions to services, purchases, usage of services, visits, complaints In order to perform a simple recency analysis, you have to divide the recent past in few (2 or 3) time periods. Each period should be relevant to the estimated lifecycle stage duration ( e.g. estimated average time between 1st and 2nd purchase).

Applications of recency analysis are:
• Customer retention programs (identify groups with decreased purchasing recency or identify groups with very recent complaints)
• In a large product portfolio, identify products which achieve high customer loyalty (high recency according to the defined cycle) and products that do not
• Management & optimization of campaign plans (adapt offers according to recency groups e.g. with a discount ladder)
• Identify campaign response rates per used recency group, in a test campaign
• Adjust estimation of Customer value ranking, according to recency
• Identification of high value Customers

Recency concepts are applied in TV shopping or on-line shopping (e.g. on Amazon.com when you buy one book, a message appears saying ‘If you make a second order within 90 minutes order shipment will be combined’) The concept of recency and its uses have only meaning and business value when adopted to the specific context of a business and its products. Recency metrics should be applied only to the same product, since different products have different characteristics (customer lifecycles, comparative product price (monetary value)). In order to analyse Customer value for all products, more complex metrics are needed (see RFM).

RFM (recency - frequency - monetary value) customer scoring techniques

RFM is a technique analysing the three dimensions of Customer activity:
• Recency: when was the last customer interaction
• Frequency: how frequent was the Customer in its interactions with the business
• Monetary value of the interactions

The sort of interaction (or Customer contact event) can vary according to the market and the analysis goals. Usually it involves Customer orders or service usage (e.g. usage of a credit card, usage of telecom services), but it can also involve faults, complaints, web site visits, registration to services, or any other event of importance to the business.
The RFM concept is the following:
• Customers who ordered recently are more likely to order again than those who ordered in a less recent period
• Customers who ordered frequently are more likely to order again than those who ordered less frequently
• Customers who ordered a higher monetary value (spent more) are more likely to order again than those who ordered a lower monetary value, and it has been tested heavily in the catalog businesses.

Recency, frequency and monetary value, form the basis of database marketing.

Frequency is often a powerful predictor of response, but it is seldom as powerful as Recency. Recency is the most powerful predictor and the easiest to define. There may be alternative ways to measure frequency. One should test alternative measures to figure out which is best suited. This can be done by testing the response on these alternatives.

Recency enables the prediction of future value, while frequency and monetary value enable the estimation of the current value. The combination of the 3 dimensions (RFM) allows the combined analysis of current and future Customer value.

The sequence R->F->M reflects a decreasing series of predictive power (however, this is not always true for the ‘M’ predictor: if one tries to promote an expensive product, M has propably increased predictive power).

In case there is no monetary value attached to an interaction (e.g. a web visit or a complaint), the analysis may be limited to recency & frequency (RF analysis).

The higher the RFM score, the more probable it is for a Customer to respond to a marketing program. This fact has been clearly confirmed in practice.

Why is this fact actionable ? Because if one classifies Customers to groups according to the RFM score, she/he can expect each distinct RFM group to have substantially different response to an offer, from the rest (especially if the number of groups is limited). Therefore she/he can focus only on certain Customer groups which are expected to respond highly, o

ecency, RFM (recency - frequency - monetary value) are applied many decades, in order to identify the active Customers and achieve higher response rates to the Customer retention & loyalty efforts. In order to apply these techniques, one has to develop a database which stores all Customer contact history on all channels (or CTPs – Customer touch points).

Latency

Latency refers to the average time between Customer activity events (e.g. orders, use of services, visit on web sites). There are alternative ways to estimate the average time between events in order to determine latency. For example if one has captured the dates of a Customer’s orders, she/he can derive the intervals between orders: Time between 1st & 2nd order : 70 days, Time between 2nd & 3rd order : 50 days, latency can be determined to be 60 days (the average time between orders in this sequence of three orders). The trend in ‘time between events’, can also be analysed in order to evaluate the dynamics of a Customer’s behavior. An increasing ‘time between orders or web visits’ is not a good sign (this type of analysis is equivalent to frequency analysis which shall be described below). On the other hand if the event is related to customer dissatisfaction (e.g complaints), increasing ‘time between complaints ’ is a good sign. How can latency be used to develop a simple customer retention program. One has to estimate latency for a Customer group and then measure the days since last event for each customer in that group. When this measurement reaches the latency estimate or exceeds it for a specific Customer, one has to act in order to influence that Customer’s behavior. By offering a discount, this Customer is encouraged to continue interacting with the business.

Recency concept

It has been proven in practise that there is a recency effect in Human behaviour (relevant theoretical studies have been performed in the past but the empirical evidence is also sufficient). This effect viewed in the Customer behaviour analysis field, leads to the following concept: The more recently a Customer has ordered a product or used a service from a certain Business, the more likely it is she/he will purchase a product or use a service again from that Business. In other words, Customers who have transacted with a Business recently are more likely to transact with it again, than Customers who have transacted with that Business less recently.

The recency metric could be defined as the number of days/weeks/months (the scale is relevant to the business/product), since the last relevant transaction occurred (the definition of recency is key to its successful use - testing of alternatives may be needed). Central in this analysis is the dimension of time. ‘How long since a Customer event happened’, is key to understand past and predict future customer behaviour. However the exact scale of time which is relevant in each business/product has to be identified (e.g. in retail sales a second purchase may take place a few months after the first purchase, in home loans a second purchase may never take place). If the Customer Lifecycle is understood (for a specific Business / product), the recency effect can be used to produce actionable ideas. Customer contact history can be used to identify standard Customer lifecycle stages (identify types of milestone events, stages and average stage duration).

Given that a relatively more recent Customer is more likely to buy again, than a less recent one, the former has relatively higher Customer value. This is strictly true if we compare Customers of similar purchase value. We notice here the fact that we reach a comparative conclusion. Recency is widely acknowledged as the most powerful predictor of future behaviour. The future behavior under analysis may be any of the following: Subscriptions to services, purchases, usage of services, visits, complaints In order to perform a simple recency analysis, you have to divide the recent past in few (2 or 3) time periods. Each period should be relevant to the estimated lifecycle stage duration ( e.g. estimated average time between 1st and 2nd purchase).

Applications of recency analysis are:
• Customer retention programs (identify groups with decreased purchasing recency or identify groups with very recent complaints)
• In a large product portfolio, identify products which achieve high customer loyalty (high recency according to the defined cycle) and products that do not
• Management & optimization of campaign plans (adapt offers according to recency groups e.g. with a discount ladder)
• Identify campaign response rates per used recency group, in a test campaign
• Adjust estimation of Customer value ranking, according to recency
• Identification of high value Customers

Recency concepts are applied in TV shopping or on-line shopping (e.g. on Amazon.com when you buy one book, a message appears saying ‘If you make a second order within 90 minutes order shipment will be combined’) The concept of recency and its uses have only meaning and business value when adopted to the specific context of a business and its products. Recency metrics should be applied only to the same product, since different products have different characteristics (customer lifecycles, comparative product price (monetary value)). In order to analyse Customer value for all products, more complex metrics are needed (see RFM).

RFM (recency - frequency - monetary value) customer scoring techniques

RFM is a technique analysing the three dimensions of Customer activity:
• Recency: when was the last customer interaction
• Frequency: how frequent was the Customer in its interactions with the business
• Monetary value of the interactions

The sort of interaction (or Customer contact event) can vary according to the market and the analysis goals. Usually it involves Customer orders or service usage (e.g. usage of a credit card, usage of telecom services), but it can also involve faults, complaints, web site visits, registration to services, or any other event of importance to the business.
The RFM concept is the following:
• Customers who ordered recently are more likely to order again than those who ordered in a less recent period
• Customers who ordered frequently are more likely to order again than those who ordered less frequently
• Customers who ordered a higher monetary value (spent more) are more likely to order again than those who ordered a lower monetary value, and it has been tested heavily in the catalog businesses.

Recency, frequency and monetary value, form the basis of database marketing.

Frequency is often a powerful predictor of response, but it is seldom as powerful as Recency. Recency is the most powerful predictor and the easiest to define. There may be alternative ways to measure frequency. One should test alternative measures to figure out which is best suited. This can be done by testing the response on these alternatives.

Recency enables the prediction of future value, while frequency and monetary value enable the estimation of the current value. The combination of the 3 dimensions (RFM) allows the combined analysis of current and future Customer value.

The sequence R->F->M reflects a decreasing series of predictive power (however, this is not always true for the ‘M’ predictor: if one tries to promote an expensive product, M has propably increased predictive power).

In case there is no monetary value attached to an interaction (e.g. a web visit or a complaint), the analysis may be limited to recency & frequency (RF analysis).

The higher the RFM score, the more probable it is for a Customer to respond to a marketing program. This fact has been clearly confirmed in practice.

Why is this fact actionable ? Because if one classifies Customers to groups according to the RFM score, she/he can expect each distinct RFM group to have substantially different response to an offer, from the rest (especially if the number of groups is limited). Therefore she/he can focus only on certain Customer groups which are expected to respond highly,

he following concept: The more recently a Customer has ordered a product or used a service from a certain Business, the more likely it is she/he will purchase a product or use a service again from that Business. In other words, Customers who have transacted with a Business recently are more likely to transact with it again, than Customers who have transacted with that Business less recently.

The recency metric could be defined as the number of days/weeks/months (the scale is relevant to the business/product), since the last relevant transaction occurred (the definition of recency is key to its successful use - testing of alternatives may be needed). Central in this analysis is the dimension of time. ‘How long since a Customer event happened’, is key to understand past and predict future customer behaviour. However the exact scale of time which is relevant in each business/product has to be identified (e.g. in retail sales a second purchase may take place a few months after the first purchase, in home loans a second purchase may never take place). If the Customer Lifecycle is understood (for a specific Business / product), the recency effect can be used to produce actionable ideas. Customer contact history can be used to identify standard Customer lifecycle stages (identify types of milestone events, stages and average stage duration).

Given that a relatively more recent Customer is more likely to buy again, than a less recent one, the former has relatively higher Customer value. This is strictly true if we compare Customers of similar purchase value. We notice here the fact that we reach a comparative conclusion. Recency is widely acknowledged as the most powerful predictor of future behaviour. The future behavior under analysis may be any of the following: Subscriptions to services, purchases, usage of services, visits, complaints In order to perform a simple recency analysis, you have to divide the recent past in few (2 or 3) time periods. Each period should be relevant to the estimated lifecycle stage duration ( e.g. estimated average time between 1st and 2nd purchase).

Applications of recency analysis are:
• Customer retention programs (identify groups with decreased purchasing recency or identify groups with very recent complaints)
• In a large product portfolio, identify products which achieve high customer loyalty (high recency according to the defined cycle) and products that do not
• Management & optimization of campaign plans (adapt offers according to recency groups e.g. with a discount ladder)
• Identify campaign response rates per used recency group, in a test campaign
• Adjust estimation of Customer value ranking, according to recency
• Identification of high value Customers

Recency concepts are applied in TV shopping or on-line shopping (e.g. on Amazon.com when you buy one book, a message appears saying ‘If you make a second order within 90 minutes order shipment will be combined’) The concept of recency and its uses have only meaning and business value when adopted to the specific context of a business and its products. Recency metrics should be applied only to the same product, since different products have different characteristics (customer lifecycles, comparative product price (monetary value)). In order to analyse Customer value for all products, more complex metrics are needed (see RFM).

RFM (recency - frequency - monetary value) customer scoring techniques

RFM is a technique analysing the three dimensions of Customer activity:
• Recency: when was the last customer interaction
• Frequency: how frequent was the Customer in its interactions with the business
• Monetary value of the interactions

The sort of interaction (or Customer contact event) can vary according to the market and the analysis goals. Usually it involves Customer orders or service usage (e.g. usage of a credit card, usage of telecom services), but it can also involve faults, complaints, web site visits, registration to services, or any other event of importance to the business.
The RFM concept is the following:
• Customers who ordered recently are more likely to order again than those who ordered in a less recent period
• Customers who ordered frequently are more likely to order again than those who ordered less frequently
• Customers who ordered a higher monetary value (spent more) are more likely to order again than those who ordered a lower monetary value, and it has been tested heavily in the catalog businesses.

Recency, frequency and monetary value, form the basis of database marketing.

Frequency is often a powerful predictor of response, but it is seldom as powerful as Recency. Recency is the most powerful predictor and the easiest to define. There may be alternative ways to measure frequency. One should test alternative measures to figure out which is best suited. This can be done by testing the response on these alternatives.

Recency enables the prediction of future value, while frequency and monetary value enable the estimation of the current value. The combination of the 3 dimensions (RFM) allows the combined analysis of current and future Customer value.

The sequence R->F->M reflects a decreasing series of predictive power (however, this is not always true for the ‘M’ predictor: if one tries to promote an expensive product, M has propably increased predictive power).

In case there is no monetary value attached to an interaction (e.g. a web visit or a complaint), the analysis may be limited to recency & frequency (RF analysis).

The higher the RFM score, the more probable it is for a Customer to respond to a marketing program. This fact has been clearly confirmed in practice.

Why is this fact actionable ? Because if one classifies Customers to groups according to the RFM score, she/he can expect each distinct RFM group to have substantially different response to an offer, from the rest (especially if the number of groups is limited). Therefore she/he can focus only on certain Customer groups which are expected to respond highly,

on ( e.g. estimated average time between 1st and 2nd purchase).

Applications of recency analysis are:
• Customer retention programs (identify groups with decreased purchasing recency or identify groups with very recent complaints)
• In a large product portfolio, identify products which achieve high customer loyalty (high recency according to the defined cycle) and products that do not
• Management & optimization of campaign plans (adapt offers according to recency groups e.g. with a discount ladder)
• Identify campaign response rates per used recency group, in a test campaign
• Adjust estimation of Customer value ranking, according to recency
• Identification of high value Customers

Recency concepts are applied in TV shopping or on-line shopping (e.g. on Amazon.com when you buy one book, a message appears saying ‘If you make a second order within 90 minutes order shipment will be combined’) The concept of recency and its uses have only meaning and business value when adopted to the specific context of a business and its products. Recency metrics should be applied only to the same product, since different products have different characteristics (customer lifecycles, comparative product price (monetary value)). In order to analyse Customer value for all products, more complex metrics are needed (see RFM).

RFM (recency - frequency - monetary value) customer scoring techniques

RFM is a technique analysing the three dimensions of Customer activity:
• Recency: when was the last customer interaction
• Frequency: how frequent was the Customer in its interactions with the business
• Monetary value of the interactions

The sort of interaction (or Customer contact event) can vary according to the market and the analysis goals. Usually it involves Customer orders or service usage (e.g. usage of a credit card, usage of telecom services), but it can also involve faults, complaints, web site visits, registration to services, or any other event of importance to the business.
The RFM concept is the following:
• Customers who ordered recently are more likely to order again than those who ordered in a less recent period
• Customers who ordered frequently are more likely to order again than those who ordered less frequently
• Customers who ordered a higher monetary value (spent more) are more likely to order again than those who ordered a lower monetary value, and it has been tested heavily in the catalog businesses.

Recency, frequency and monetary value, form the basis of database marketing.

Frequency is often a powerful predictor of response, but it is seldom as powerful as Recency. Recency is the most powerful predictor and the easiest to define. There may be alternative ways to measure frequency. One should test alternative measures to figure out which is best suited. This can be done by testing the response on these alternatives.

Recency enables the prediction of future value, while frequency and monetary value enable the estimation of the current value. The combination of the 3 dimensions (RFM) allows the combined analysis of current and future Customer value.

The sequence R->F->M reflects a decreasing series of predictive power (however, this is not always true for the ‘M’ predictor: if one tries to promote an expensive product, M has propably increased predictive power).

In case there is no monetary value attached to an interaction (e.g. a web visit or a complaint), the analysis may be limited to recency & frequency (RF analysis).

The higher the RFM score, the more probable it is for a Customer to respond to a marketing program. This fact has been clearly confirmed in practice.

Why is this fact actionable ? Because if one classifies Customers to groups according to the RFM score, she/he can expect each distinct RFM group to have substantially different response to an offer, from the rest (especially if the number of groups is limited). Therefore she/he can focus only on certain Customer groups which are expected to respond highly,

ess.
The RFM concept is the following:
• Customers who ordered recently are more likely to order again than those who ordered in a less recent period
• Customers who ordered frequently are more likely to order again than those who ordered less frequently
• Customers who ordered a higher monetary value (spent more) are more likely to order again than those who ordered a lower monetary value, and it has been tested heavily in the catalog businesses.

Recency, frequency and monetary value, form the basis of database marketing.

Frequency is often a powerful predictor of response, but it is seldom as powerful as Recency. Recency is the most powerful predictor and the easiest to define. There may be alternative ways to measure frequency. One should test alternative measures to figure out which is best suited. This can be done by testing the response on these alternatives.

Recency enables the prediction of future value, while frequency and monetary value enable the estimation of the current value. The combination of the 3 dimensions (RFM) allows the combined analysis of current and future Customer value.

The sequence R->F->M reflects a decreasing series of predictive power (however, this is not always true for the ‘M’ predictor: if one tries to promote an expensive product, M has propably increased predictive power).

In case there is no monetary value attached to an interaction (e.g. a web visit or a complaint), the analysis may be limited to recency & frequency (RF analysis).

The higher the RFM score, the more probable it is for a Customer to respond to a marketing program. This fact has been clearly confirmed in practice.

Why is this fact actionable ? Because if one classifies Customers to groups according to the RFM score, she/he can expect each distinct RFM group to have substantially different response to an offer, from the rest (especially if the number of groups is limited). Therefore she/he can focus only on certain Customer groups which are expected to respond highly, or adjust the offering in a way to achieve high response from many targeted groups. This can be achieved by offering a more attractive deal (e.g. a higher discount) to the lower RFM (or RF) groups (which are less likely to respond), than to the higher RFM group(s), in order to achieve a satisfactory response from more than one group.

Copyright 2006, Kostis Panayotakis