1

Predictive model assignment – RFM model

University Canada West

BUSI 650: BUSINESS ANALYTICS

CAMPUS-SUMMER22-04: Jul. 11, 2022 - Sep. 25, 2022

Professor Aliosman Pektas

Due: Monday, Aug 15, 2022, 11:59 pm (PDT)

2

Table of Contents
1. ABSTRACT/SUMMARY........................................................................................................................ 3
2. INTRODUCTION:.................................................................................................................................. 4
2.1 BUSINESS CASE/SCENARIO..........................................................................................................................4
2.2 TECHNICAL DETAILS OF RFM MODEL (BACKGROUND).........................................................................4
3. BODY:..................................................................................................................................................... 5
3.1 DATA UNDERSTANDING & DATA PREPARATION FOR RFM MODEL...............................................5
3.2 MODEL DEVELOPMENT.....................................................................................................................6
4. FINDINGS AND DISCUSSIONS............................................................................................................ 7
5. CONCLUSION...................................................................................................................................... 11
6. REFERENCES...................................................................................................................................... 12
7. APPENDIX/APPENDICE..................................................................................................................... 12
3

1. Abstract/Summary

Businesses have been picking up on the utilization of customer data to improve

profitability and enhance customer relationship. This paper entails a descriptive research,
analysis, and application of RFM (Recency, Frequency, Monetary) model with a business
case/scenario. The primary purpose of this paper is to illustrate the significance of utilizing
the RFM data model in real-world business scenarios. Conducting RFM analysis basically
allows the business to segment certain groups of customers based on customer behaviours
along with demographic, geographic and purchasing patterns, or come up with better
strategies to improve sales or engagement in customer retention.

This paper further explains the utilization of RFM data model in a business scenario
of a clothing retail store called SMTE, located in downtown Vancouver, Canada. SMTE is
looking forward to expanding its business by opening another store, but it needs to first
maximise revenues on existing customers to raise capital. In order to do that, the business
needs to identify and segment its existing customers according to the RFM variables.

The paper concludes by stating findings and discussions based on extracted data
which have correlations with grouped customer segments. In addition, there are
recommendations of promotional strategies to grouped segments.
4

2. Introduction:

2.1 Business Case/Scenario

SMTE clothing retail store is a fast-growing business founded by a group of young

entrepreneurs. The clothing store sells various clothing and accessories which are specially
selected from various high-end casual brands that are unique and from exclusive collections
promoted by fashionable and influential celebrities. The value preposition of SMTE clothing
is to make clothing from certain high-end brands like Supreme, Stussy, Anti-Social Social
Club, Bones, etc., be available to try in person in which usually is available upon months of
wait and only online. The target market is from teenagers to millennials that are looking for
stylish and flexing causal wears.

Currently located on Robson Street in the heart of Vancouver downtown, the business
expects to expand in new areas like Metrotown, Marine Drive and Richmond. In order for
expansion to prevail, the business needs to be more profitable to raise more capital. Thus, the
business hires a data analyst to conduct a RFM analysis on the available customer data to
impose better sales & marketing strategies towards the target market.

2.2 Technical Details of RFM Model (Background)

According to Segal (2021), the RFM model was officially known to be introduced in
the 90’s through a published article called “Optimal Selection for Direct Mail” written by
Tom Wansbeek and Jan Roelf Bult. RFM analysis gained its popularity from that article by
many marketers and businesses that they utilize the RFM model to segment customer bases
such as best customers, loyal customers, potential customers or lapsing customers, etc. As
retention of existing customers is often neglected by most businesses, they lose revenues and
often result in high expenditure of marketing new customers (Muntasir, 2022).

RFM model is an analysis tool that is applied for effective customer segmentation and
managing customer relationship based on the following three variables: recency, frequency,
and monetary values (Imani et al., 2022). For instance, each variable can be expressed from a
range of 1-3, 1-5 or 1-10 where the higher number represents for highest/desirable score of
each variable to state certain customer behavior of how recently, how frequently and how
much they spend in monetary value.
-Recency (R) refers to the interval time of latest purchase by the customer. The more recently
it occurred, the higher the potential value of the customer.
-Frequency (F) refers to the number of repetitive purchases made in a certain period by the
customers. The more repetitive the customer purchases, the higher the potential value of the
customer.
-Monetary (M) refers to the amount of money transacted in a certain period. The higher the
value of transaction, the higher the value of the customer.

Through utilization of the model with the right customer data, these quantitative
information can help identify customer groups into respective and unique characteristics such
as biggest purchasers, potential customers, or lapsed customers (Carrasco et al., 2015).
5

3. Body:
3.1 Data Understanding & Data Preparation for RFM model

As mentioned above, RFM model is solely based on analyzing the three variables of
customer data which are recency (last date of purchase), frequency (number of purchases
occurred) and monetary (amount of money spent on purchase). Enriched data is a great added
value to the predictive information and recommendation. To enrich data, the dataset needs
various demographic, geographic or other applicable features to analyze each segment and
correlate better information outcome.

The initial step of RFM model is to collect the purchase history data that contains the
date of customer purchase, total orders and amount of money spent, which then can be sorted
according to how recent the last purchase occurred, how frequently purchased and how much
in total money spent by a certain customer designated with their Customer ID. The raw data
should be filtered and cleaned just like the example data below:

Table 1. An Example Dataset

Note. An Example Dataset

After the focused data is ready, recency value, frequency value and monetary value
can be determined into a desired range of 1 to 5 or 1 to 10 (Range of 1 to 5 is commonly used
where 5 is higher/best value). Ultimately, the RFM score can be concatenated (CONCAD) to
express the value of each customer and which segment he/she belongs to. For instance, the
RFM score of 555 will represent best customers who purchases most recently and often with
great spending power. Hence, they can be grouped a gold customers or VIP, etc. On the other
hand, the RFM score of 111 will represent lapsing customers who purchase for once with
very little spending power. Furthermore, there can also be loyal customers who
intermediately purchases with average spending but not very often. In summary, the
combination of each variable score will imply about the customer spending behavior on the
business products.
6

3.2 Model Development

For this paper, I have developed a RFM model for SMTE clothing retail store to
improve their profitability and retention of existing customers. Based on the RFM scores, I
have segmented the customers into 3 segments with similar RFM scores to gold customers,
silver customers and bronze customers.

Customer Segmentation:
Gold Customers- Customers with RFM scores combination of value 4 and 5 are in this
segment. Gold customers are essentially the best existing customers to make purchases from
the store often and recently and with spending of more than average. These are the ideal
customers the business should retain to generate regular profit.

Silver Customers- Customers with RFM scores combination of value 3 and 4 are in this
segment. Silver Customers are the most loyal customers where they make purchases from
time to time with average to high spending volume.

Bronze Customers- Customers with RFM scores Combination of value 1, 2 and 3 are in this
segment. Bronze customers care the lapsing customers where they may buy once in a while
or even just once with average to low spending volume.

Table 2. Customer Segmentation Criteria.

RFM Score Customer segmentation Description

Any combination of 4-5 Gold The best customers

Any combination of 3-4 Silver The most loyal customers

Any combination of Bronze Lapsing or low potential

1-2-3 customers
7

4. Findings and Discussions

Findings 1:

Graph 1. No. of Customer Segment by Age

A ge
Bronze Gold Silver
7

0
14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

Note. Number of Customer Segment by Age

The graph 1 above illustrates that there is higher demand in the age range between age
20 to 30, and usually from the bronze customers’ segment. Based on this finding, the store
should increase the clothing stocks/items with unique designs that appeal to this age range of
20 to 30 so that they would be willing to spend more and become silver or gold customers.
Based on the number of customers in this age range, there is great potential to maximise
sales, providing with better choices that appeals to them.

Table 3. Supplementary data of age range for customer segments

Note. Supplementary data of age range for customer segments

Recommendation:

The business can conduct email-marketing to a range of customers between 20 to 30

by introducing great offers and deals, as there are email addresses available in the
membership program.
8

Findings 2:

Graph 2. No. of Customer Segment by Location

Locati on

Bronze Gold Silver

30

25

20

15

10

0
Burnaby North Vancouver Others Richmond Surrey Vancouver Downtown

Note. No. of Customer Segment by Location

Graph 2 illustrates that most of the customers that have excellent potential in RFM
scores are from Burnaby.

Table 4. Supplementary data of location for customer segments

Count of
Location Column Labels        
North Vancouver Grand
Location Burnaby Vancouver Others Richmond Surrey Downtown Total
Bronze 24 7 1 10 1 12 55
Gold 8 3 1 1 3 16
Silver 6 1 4 11
Grand Total 38 11 1 15 2 15 82
Note. Supplementary data of location for customer segments

Recommendation:

While the business is on a mission to expand and open another store, they should
consider open the Burnaby district where it is accessible and convenient than traveling to
downtown store. As a profitable outcome, existing customers from Burnaby can visit more
frequently to check out new collections, increasing recency, frequency, and monetary value.
9

Findings 3:

Graph 3. No. of Customer Segment by Gender

Gender
30

25

20

15

10

0
Bronze Gold Silver

F M
Note. No. of Customer Segment by Gender

Graph 3 illustrates there is a close to equal number of silver customers from both male
and female.

Table 5. Supplementary data of gender for customer segments

Count of
Gender Column Labels
Grand
Gender F M Total
2 2
Bronze 7 8 55
Gold 9 7 16
Silver 6 5 11
4 4
Grand Total 2 0 82
Note. Supplementary data of gender for customer segments

Recommendation:

The business should introduce a promotional strategies for silver customers like
cashback programs and rewards to membership for high purchases. This will increase the
interest of silver customers to visit the store more often and spend more, eventually turning
them to gold customers. As there is a close number of sales from male and female customers,
the business should keep up with the demand for both genders.
10

Findings 4:
Graph 4. No. of Customer Segment by Payment

Payment
50
45
40
35
30
25
20
15
10
5
0
Bronze Gold Silver

Debit/Credit Giftcard
Note. No. of Customer Segment by Payment

Graph 4 illustrates that majority of payments are done by debit or credit card, but it is
not essentially an issue. The idea of implementing gift-card payments to the business is to
generate sales from customers when they want to buy gifts to their friends or colleagues for
various occasions.

Table 6. Supplementary data of payment for customer segments

Count of
Payment Column Labels
Debit/ Giftcar Grand
Gender Credit d Total
Bronze 46 9 55
Gold 11 5 16
Silver 8 3 11
Grand Total 65 17 82
Note. Supplementary data of payment for customer segments

Recommendations:

The business should promote great offers and deals on gift-cards as it is a good
strategy to drive sales, based on available given data. There is a 20% payment by gift-cards
which can be a potential to maximise sales to a certain point. Thus, loyal customers can buy it
for other potential customers through gift cards. Basically, the customer who buys the gift
11

card will promote a service, product, or brand that a company offers as part of the gift-giving
process (Hunter, 2014).

5. Conclusion

RFM model is a great analysis tool that can simply identify customer segments to the
purchasing behavior of the that have relationships with categorised consumer segments. It
also creates a basis for suggestions of marketing tactics for desired segmented markets that
can be implemented according.

In the business scenario of SMTE clothing retail store, there shows a great potential to
grow after conducting the RFM analysis.
Based on the findings,
1. the business needs to maximize profits on age range of 20-30 by providing clothing
items that appeal most to them.
2. the business should consider about opening another store in Burnaby where majority
of existing customers are present.
3. the business should keep up with the demand of both male and female customers for
customer retention.
4. the business should promote on selling gift-cards to customers to maximise sales.
12

6. References

Carrasco, R., Blasco, M., & Herrera-Viedma, E. (2015). A 2-tuple Fuzzy Linguistic RFM
Model and Its Implementation. Procedia Computer Science, 55, 1340-1347.
10.1016/j.procs.2015.07.118

Hunter, S. (2014, May 6). How Do Gift Cards Work & Businesses Make Money On Them?
Giftcards.com. https://www.giftcards.com/gcgf/how-do-gift-cards-work

Imani, A., Abbasi, M., Ahang, F., Ghaffari, H. & Mehdi, M. (2022). Customer segmentation
to identify key customers based on RFM model by using data mining techniques.
International journal of research in industrial engineering, 11 (1), 62-76.

Muntasir, M. (2022, March 16). Everything You Need To Know About RFM Analysis.
https://bettermarketing.pub/everything-you-need-to-know-about-rfm-analysis-
d332fc69cdb0

Segal, T. (2021, August 23). Recency, Frequency, Monetary Value (RFM). Investopedia.
https://www.investopedia.com/terms/r/rfm-recency-frequency-monetary-value.asp

7. Appendix/Appendice

Table 1. An Example Dataset

Table 2. Customer Segmentation Criteria.

RFM Score Customer segmentation Description

Any combination of 4-5 Gold The best customers

 13

Any combination of 3-4 Silver The most loyal customers

Any combination of Bronze Lapsing or low potential

1-2-3 customers

Table 3. Supplementary data of age range for customer segments

Table 4. Supplementary data of location for customer segments

Count of
Location Column Labels        
North Vancouver Grand
Location Burnaby Vancouver Others Richmond Surrey Downtown Total
Bronze 24 7 1 10 1 12 55
Gold 8 3 1 1 3 16
Silver 6 1 4 11
Grand Total 38 11 1 15 2 15 82

Table 5. Supplementary data of gender for customer segments

Count of
Gender Column Labels
Grand
Gender F M Total
2 2
Bronze 7 8 55
Gold 9 7 16
Silver 6 5 11
4 4
Grand Total 2 0 82

Table 6. Supplementary data of payment for customer segments

Count of
Payment Column Labels
Debit/ Giftcar Grand
Gender Credit d Total
Bronze 46 9 55
Gold 11 5 16
Silver 8 3 11
Grand Total 65 17 82

Graph 1. No. of Customer Segment by Age

14

A ge
Bronze Gold Silver

7
6
5
4
3
2
1
0
14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

Graph 2. No. of Customer Segment by Location

Locati on

Bronze Gold Silver

30
25
20
15
10
5
0
Burnaby North Others Richmond Surrey Vancouver
Vancouver Downtown

Graph 3. No. of Customer Segment by Gender

30

25 Gender
20

15

10

0
Bronze Gold Silver

F M
15

Graph 4. No. of Customer Segment by Payment

Payment
50
45
40
35
30
25
20
15
10
5
0
Bronze Gold Silver

Debit/Credit Giftcard