CONTENT

Chapter I: Overview Of RFM Customer 360........................................................................ 4

1.1. Introducing Customer 360 ...................................................................................... 4
1.2. Benefits of Customer 360 analysis .......................................................................... 4
1.3. Fundamental concepts in RFM Customer 360 analytics ........................................... 5
1.4. Benefits of RFM Customer 360 analysis ..................................................................... 5
Chapter II: Determine Customer Groups ............................................................................. 6
2.1. Define RFM .................................................................................................................... 6
2.1.1. Data separation ........................................................................................................... 6
2.1.2. Build RFM SCORE scoring framework and classification in SQL. ..................... 7
Chapter III. Customer Analysis ........................................................................................... 10
3.1. Overview of customer segments ................................................................................. 10
3.2. Understand your customers ....................................................................................... 11
3.3. Revenue growth solution for the last quarter of the year ........................................ 12
CODE SQL REFERENCE ................................................................................................... 14

2
 Chapter I: Overview Of RFM Customer 360
1.1. Introducing Customer 360
Customer360 is a customer-based analytics strategy transaction data set (Transaction Data);
website interaction data or application (Interaction Data); consumer behavior and customer needs
(Behavioral Data) and data about customer age and demographics (Demographics Data)

1.2. Benefits of Customer 360 analysis

A Customer 360-degree view provides a comprehensive understanding of customers and
their interactions with a company
1. Create More Strategic Sales and Marketing Campaigns:
• A 360-degree customer view extends the capabilities of sales and marketing teams.
By leveraging existing data, these teams can design more strategic campaigns.
2. Understand Customers and Their Behaviors:
• With a 360-degree view, businesses gain insights into customer tastes and behaviors.
• Healthcare providers, for example, can understand not only a patient’s medical
history but also their habits and behaviors. Are regular annual exams scheduled?
What preventative measures can be taken?
3. Deliver More Personalized Shopping Experiences:
• Perhaps the most crucial benefit is the ability to provide personalized shopping
experiences.
• By using aggregated data, businesses can enhance customer experiences throughout
their lifetime shopping journey.
• For instance, a bank aware of a customer’s new dependents planning to attend college
can send personalized savings account offerings.

4
1.3. Fundamental concepts in RFM Customer 360 analytics
- RFM Customer 360 is a customer analysis method based on the RFM (Recency - Frequency -
Money) model.
- This method helps businesses evaluate the value of each customer based on their level of
interaction with the business and offer different approach strategies to optimize customer
care.

Recency ® Measure the time since the customer's

last transaction. The recency of the
transaction can reflect the customer's
current interaction with the business.

Frequency Measure customer purchase frequency

(F) over a certain period of time. The
number of transactions shows the level
of customer engagement.

Monetary Measures the total value of transactions

(M) that customers have made. This value
can reflect the importance of customers
to the business.

1.4. Benefits of RFM Customer 360 analysis

- Better understanding of customers: RFM analysis helps businesses have an overview of their
customers, including purchase frequency, order value and the most recent time the customer made
a purchase. This helps businesses better understand customer needs and come up with appropriate
business strategies.
- Classify customers and optimize marketing campaigns: After ranking each customer according
to RFM criteria, we will have a list of each customer group. This helps optimize your marketing
strategy budget, improve customer interactions, and increase revenue.
- Reduce marketing costs: RFM analysis helps businesses focus on potential customer groups and
devise appropriate marketing strategies. This helps businesses reduce marketing costs and
optimize their resources.

5
 Chapter II: Determine Customer Groups
2.1. Define RFM
2.1.1. Data separation
Description of data sheets
In this case, we have two tables that need to be analyzed: Customer_Transaction table,
Customer_Registered table

Image 2.1. Customer_Transaction

Image 2.1. Customer_ Registered

Customer_transaction Table: records transactions of customers from June 2022 to August
2022. Each row on the table is each transaction task. The Customer_transaction table has 4 columns of
data. Each transaction will be identified by [Transaction_ID]. Each transaction records information:
Which customer [CustomerID], purchase date [Purchase_Date], how much money [GMV].
Customer_Registered Table: records customers who have registered for membership cards.
In this case, we will use the information of the Customer_Registered table to calculate the customer's
contract age (Lifetime) and then divide it by FREQUENCY and MONETARY for the purpose of fair
comparison between customers on the same time period. The purpose is to compare performance
between customer groups

6
2.1.2. Build RFM SCORE scoring framework and classification in SQL.
Tool: MS SQL (Source Code)
How to do:
- Process OLTP data to OLAP from Customer_Transaction table
- Use join to join two tables Customer_Registered and Customer_Transaction to get information
- Use functions such as SUM, DATEDIFF, WINDOWFUNCTION, CASE WHEN to process
➢ Step 1: Calculate Recency – Frequency – Monetary

Convention for calculating Recency – Frequency – Monetary

RECENCY Time period since the customer's last purchase (use
September 1, 2022 to calculate)
FREQUENCY Total number of days a customer purchases/Lifetime
MONETARY Total amount a customer has purchased/Lifetime

Customer_transaction table after being processed

➢ Step 2: Score RFM according to IQR in SQL

Divide the R-F-M data into 4 equal parts (quartiles) labeled from 1 to 4. In this case, we
convention that the larger the label, the better the value (Label=4 is best). For Recency, the higher its
value, the worse it is, indicating that customers have not come back to buy for a long time.

7
 Customer_transaction table after calculating RFM according to IQR
➢ Step 3: Classify customer groups based on RFM in SQL

8
 ➢ Step 4: Customer definition

Customer RFM score groups Characteristic

segmentation
Champions 444 These are the new customers who buy frequently and spend
the most. These customers are very loyal, willing to spend
generously and will likely make another purchase soon.
Loyal 344 These are customers who spend at an average level but buy
Customers very frequently.
Potential 442, 441, 431, 433, 432, Customers who have recently transacted, spent on average
Loyalist 423, 342, 341, 333, 323 a sizable amount of money, and have made multiple
purchases.
Recent 422, 421, 412, 411, 311 Customers who have recently purchased, have a low cart
Customers value, and do not purchase frequently.
Promising 424, 413, 414, 415, 314, These are customers who have recently purchased products
313 and have large purchasing power but not often.
Customers 443, 434, 343, 334, 324 These are customers who have a fairly high purchasing
Needing frequency and cart value and have not returned to purchase
Attention recently.
About To Sleep 331, 321, 312, 221, 213 These are customers who have not made a purchase for a
long time, previously purchased with low frequency and
low cart value.
At Risk 243, 242, 234, 224, 143, These are customers who have not returned for a long time
142, 134, 133, 124 and have made frequent purchases with a fairly average
shopping cart value.
Can’t Lose 144, 214, 114, 113 These are customers who have not returned for a long time
Them and have made regular purchases, with a very large
shopping cart value.
Hibernating 332, 322, 231, 241, 233, These are customers who have not returned for a long time
232, 223, 222, 132, 123, and have weak purchasing power (low purchasing
122, 212, 211 frequency and low shopping cart value).
Lost 111, 112, 121, 131, 141 As customers who have not returned for a long time, their
purchasing frequency and shopping cart value are also very
low.

9
 Chapter III. Customer Analysis
3.1. Overview of customer segments
Tool: PowerBI
The store's total revenue reached 2.46 billion VND with a customer base of 112.66 (thousand) people.

Image 3.1. Business situation

Image 3.1. Business situation

- According to the statistics table showing in 3 months, we can see that the TOP 5 revenue
belongs to the customer groups Hibernating, Champion, Potential Loyalist, At Risk,
Loyal Customers
- However, if we look at the report, we can see that although the Hibernating Customer
group brings in TOP 1 revenue for the business, when compared to the purchasing frequency
of customers in the Loyal Customers and Champion groups. These groups show that the
purchasing frequency of this group is once a month, while the purchasing frequency of
customers belonging to the Hibernating group has a low frequency.
- The "At Risk" customer group brings in the top 5 revenue, so there needs to be a
campaign to care for them to keep them buying because customers have a very high risk of
switching to rival companies.
- In addition, the purchasing frequency of new customer groups is very high, showing that the
campaign was quite successful in attracting new customers to buy regularly. However, the
revenue brought in was not as expected. It is necessary to retain these customers, offer many
incentive programs for them, and need to convert them into potential customers.

10
 Through this report, it is necessary to find a strategy to increase the purchasing frequency of the
Hiberating group and pull the number of customers from this group to the Champion/ Loyal
Customers group.
3.2. Understand your customers

TOTAL OF CUSTOMERS BY SEGMENTATION

TOTAL OF REVENUE BY SEGMENTATION

Looking at the charts, we can see that the number of customers of the business is distributed
as follows:

11
- The number of customers of the Hibernating group is 31.67% compared to the remaining customer
groups and has the largest number of customer groups. Therefore, businesses that want to increase
sales need to have a strategy to care for and stimulate this group of customers so that they can return
to use the service more often and have a plan to convert this group of customers to Loyal or Potential
- Although the "Champion Customers” group brings in high revenue and ranks in the top 2, the
number of customers of this group only accounts for 7.86%. In addition, the "Loyal Customers" group
only accounts for 5.18% even though this group brings in the top 5 revenue. This shows that customer
care is not good, and promotion and marketing programs are not effective so that results in a relatively
low percentage of this customer group, while the "At Risk" and "About to sleep" customer groups
have a high percentage of customers and there has been no conversion among the groups.
3.3. Revenue growth solution for the last quarter of the year
➢ Increase conversion rate and revenue of Potential and Hibernating groups to Loyal segment:
- For customers in the Hibernating group:
• Generate attention and stimulate repeat visits by sending email notifications
about special offers, new products, or useful information. Create a special
marketing campaign to keep them coming back.
• Survey and feedback: Ask them why they haven't made a purchase recently.
Based on their feedback, you can adjust your marketing strategy to better
suit their needs.
- For customers in the Potential group:
• Create a good customer experience: Ensure that customers receive good
service, from ordering to delivery and after-sales. Address all their questions
and feedback quickly and professionally.
• Incentives and promotions: create a rewards program specifically for regular
customers. For example: Bonus points, discounts for next purchases, free
gifts.
➢ Maintain Champion and Loyal customer base.
- Understanding customers:
• Champion Customers: Continue to create positive experiences for them. Ask
them about their opinions and wishes to improve the service.
• Loyal customers: Continuously monitor their behavior. Make sure that they
still feel satisfied and that their needs are met.
- Make a good first impression:
• Champion Customers: Continue to provide high value and quality. Keep them
feeling like they made the right decision in choosing you.
• Loyal Customers: Make sure their first experience with, you is positive. They
will remember this and continue to support you.
- Offers and promotions:
• Champion Customers: Give them special offers, gifts, or exclusive
experiences. Show appreciation for them.
• Loyal customers: Build a reward program specifically for them. For example:
Bonus points, discounts, gifts.
- Interact and communicate regularly:
• Champion Customers: Contact them regularly via email, text, or call. Ask
about the situation and provide new information.
• Loyal Customers: Send notifications about new products, promotions, and
special events. Maintain interaction via social networks.
- Monitor and evaluate performance:

12
• Champion Customers: Measure their purchase frequency and interactions. Make sure
they don't get "lost".
• Loyal customers: Track RFM index (Recency, Frequency, Monetary value). Optimize
your marketing campaign based on this data.

13
CODE SQL REFERENCE
-- Tính các giá trị Recency, Frequency, Monetary
SELECT cr.Contract,
DATEDIFF(DAY, MAX(CAST(Purchase_Date AS DATE)), '2022-09-01') AS
RECENCY,
ROUND(CAST(COUNT(Purchase_date) AS FLOAT) /
CAST(DATEDIFF(YEAR, CAST(created_date AS DATE), '2022-09-01')
AS FLOAT), 2) AS FREQUENCY,
ROUND(CAST(SUM(GMV) AS FLOAT) /
CAST(DATEDIFF(YEAR, CAST(created_date AS DATE), '2022-09-01')
AS FLOAT), 2) AS MONETARY,
ROW_NUMBER()OVER(ORDER BY (MAX(CAST(Purchase_Date AS DATE)))DESC) AS
RN_RECENCY,
ROW_NUMBER()OVER(ORDER BY (ROUND(CAST(COUNT(DISTINCT Purchase_date)
AS FLOAT) /
CAST(DATEDIFF(YEAR,
CAST(created_date AS DATE), '2022-09-01') AS FLOAT), 2))) AS RN_FREQUENCY,
ROW_NUMBER()OVER(ORDER BY SUM(GMV)) AS RN_MONETARY
INTO #a
FROM Customer_Transaction ct
JOIN Customer_Registered cr ON ct.CustomerID = cr.ID
WHERE CustomerID != 0
GROUP BY cr.Contract, created_date;
-- Tính điểm RFM
SELECT Contract, RECENCY, FREQUENCY, MONETARY,
CASE
WHEN RN_RECENCY >= (SELECT MIN(RN_RECENCY) FROM #a) AND
RN_RECENCY < (SELECT CAST(COUNT(RN_RECENCY) * 0.25 AS INT)
FROM #a) THEN 4
WHEN RN_RECENCY >= (SELECT CAST(COUNT(RN_RECENCY) * 0.25 AS INT)
FROM #a) AND
RN_RECENCY < (SELECT CAST(COUNT(RN_RECENCY) * 0.5 AS INT)
FROM #a) THEN 3
WHEN RN_RECENCY >= (SELECT CAST(COUNT(RN_RECENCY) * 0.5 AS INT)
FROM #a) AND
RN_RECENCY < (SELECT CAST(COUNT(RN_RECENCY) * 0.75 AS INT)
FROM #a) THEN 2
ELSE 1

END AS R,
CASE
WHEN RN_FREQUENCY >= (SELECT MIN(RN_FREQUENCY) FROM #a) AND
RN_FREQUENCY < (SELECT CAST(COUNT(RN_FREQUENCY) * 0.25 AS
INT) FROM #a) THEN 1
WHEN RN_FREQUENCY >= (SELECT CAST(COUNT(RN_FREQUENCY) * 0.25 AS
INT) FROM #a) AND
RN_FREQUENCY < (SELECT CAST(COUNT(RN_FREQUENCY) * 0.5 AS
INT) FROM #a) THEN 2
WHEN RN_FREQUENCY >= (SELECT CAST(COUNT(RN_FREQUENCY) * 0.5 AS
INT) FROM #a) AND
RN_FREQUENCY < (SELECT CAST(COUNT(RN_FREQUENCY) * 0.75 AS
INT) FROM #a) THEN 3
ELSE 4

END AS F,

14
 CASE
WHEN RN_MONETARY >= (SELECT MIN(RN_MONETARY) FROM #a) AND
RN_MONETARY < (SELECT CAST(COUNT(RN_MONETARY) * 0.25 AS
INT) FROM #a) THEN 1
WHEN RN_MONETARY >= (SELECT CAST(COUNT(RN_MONETARY) * 0.25 AS
INT) FROM #a) AND
RN_MONETARY < (SELECT CAST(COUNT(RN_MONETARY) * 0.5 AS INT)
FROM #a) THEN 2
WHEN RN_MONETARY >= (SELECT CAST(COUNT(RN_MONETARY) * 0.5 AS
INT) FROM #a) AND
RN_MONETARY < (SELECT CAST(COUNT(RN_MONETARY) * 0.75 AS
INT) FROM #a) THEN 3
ELSE 4

END AS M
INTO #RFM
FROM #a;
--Mapping các nhóm khách hàng
SELECT *,
CONCAT(R,F,M) AS 'RFM',
CASE
WHEN CONCAT(R,F,M) IN ('111', '112', '121', '131', '141') THEN
'Lost'
WHEN CONCAT(R,F,M) IN ('332', '322', '231', '241', '233', '232',
'223', '222', '132', '123', '122', '212', '211') THEN 'Hibernating'
WHEN CONCAT(R,F,M) IN ('144', '214', '215', '115', '114', '113')
THEN 'Can’t Lose Them'
WHEN CONCAT(R,F,M) IN ('243', '242', '234', '224', '143', '142',
'134', '133', '124') THEN 'At Risk'
WHEN CONCAT(R,F,M) IN ('331', '321', '312', '221', '213') THEN
'About To Sleep'
WHEN CONCAT(R,F,M) IN ('443', '434', '343', '334', '324') THEN
'Customers Needing Attention'
WHEN CONCAT(R,F,M) IN ('424', '413', '414', '415', '314', '313')
THEN 'Promising'
WHEN CONCAT(R,F,M) IN ('422', '421', '412', '411', '311') THEN
'Recent Customers'
WHEN CONCAT(R,F,M) IN ('442', '441', '431', '433', '432', '423',
'342', '341', '333', '323') THEN 'Potential Loyalist'
WHEN CONCAT(R,F,M) = '344' THEN 'Loyal Customers'
WHEN CONCAT(R,F,M) = '444' THEN 'Champion'
ELSE 'New Customers'

END AS SEGMENTATION
FROM #RFM

15