A Deeper Path to

Personalization
How RFM segmentation can deliver more
personalized engagement
 01 - 02

CONTENTS
Introduction

03
Where Traditional Customer
Segmentation Falls Short

04 - 06
RFM Segmentation: What it
Means for Your Business Growth

07 - 12
Reimagining Engagement
Using RFM Segmentation
• Creating more effective segments based on RFM score
• Developing personalized engagement strategies

13 - 15
How a Billion-dollar Global
Security Software Brand
Executes RFM Segmentation

15
A New Era of Personalization

16
What To Read Next

16
About Moengage
We are living in the golden age of product communications, and relevant engagement,
management. The opportunity has never been across all touchpoints, whether offline or online
greater for B2C product managers to make a [web, mobile, email, social media, chat,
massive impact in their roles to drive loyalty, and more].
optimize campaign performance, and
increase revenue.
MoEngage is fully committed to helping brands
Product managers that are excelling today have move away from a campaign-centric focus and
truly become customer-centric. They’re to a customer-centric future. With our vision to
obsessed with uncovering deeper insights into be the world's most trusted customer
their customers - and they make these insights engagement platform, product managers are
easily available, dynamically updated, and empowered to deliver personalized and
instantly accessible to everyone. meaningful experiences across the customer
journey with speed and ease.
They have a rich understanding of their customer
journeys, likes, dislikes, preferences, lifestyle With this eBook, we are excited to share
choices, and their propensity to valuable learnings on how to deepen your
churn/buy/remain loyal. And most importantly, understanding of your customers to build
they have the ability to act on these stronger customer engagement and accelerate
insights, instantly. business growth.

They then use these deep insights to deliver

personalized recommendations, contextual

Nalin Goel, SVP Product at MoEngage

Nalin Goel is the SVP Product at MoEngage, the leading insights-led

customer engagement platform.

www.moengage.com

01
Every Product Manager will face a version of this What I like most about Recency, Frequency, and
dilemma: You want to create experiences that Monetary (RFM) Analysis is that it offers a
feel like they’re made just for your ideal deeper path to segmentation based on
consumer. At the same time, you want to extend customer behaviors (i.e. how often they
that feeling to a diverse consumer base with very purchase, what they purchase, or what their
different demographic and average annual spend is). That type of
behavioral characteristics. information is key to accelerating revenue
growth by increasing the lifetime value of
How do you square this circle? How do you existing customers as well as finding the right
reconcile the contradiction between broad new customers.
appeal and specific relevance? The answer
is segmentation. In this guide, you’ll learn how to deploy RFM
Analysis to execute insights-driven
Segmentation is a well-known field and there are segmentation at a granular level. And the best
many ways to go about it. One common trap is to part is that you don’t need anything other than
bucket customers based on one-size-fits-all your current data.
qualities that have little to do with customer
behavior from their product and brand It’s been a pleasure partnering with MoEngage,
experiences. While market research and and I hope this guide helps you improve your
demographic information each play a role in key performance indicators in a short period
segmentation; it’s not enough. of time.

Carlos Gonzalez de Villaumbrosia,

CEO at Product School

Carlos Gonzalez de Villaumbrosia is the CEO at Product School, the

global leader in Product Management training.

www.productschool.com

02
WHERE TRADITIONAL CUSTOMER
SEGMENTATION FALLS SHORT
Experienced marketers know that customer segmentation relies heavily on the human aspect
segmentation is a critical first step in establishing of hypothesizing - such as making assumptions
who and how we should be engaging based on general demographics. So it can often
our customers. result in incorrectly weighted factors and
pre-existing biases. The analysis ends up being
Customer segmentation helps us understand our generic and non-personal, and our targets are
customers better, at least on a surface level. By nothing more than customer IDs within a
developing sample segments, we can make given segment.
predictions for a given population. But it is still
difficult to accurately forecast the tendencies of
niche customer groups or individual customers on
a deeper level. This is because traditional

Traditional Segmentation Methods RFM Model of Customer Value

Built on consumer understanding and research Built on historical transactions between

studies commissioned by the business the customer and the business

Uses demographic and psychographic variables Uses R, F and M variables of customer data

Analyzes representative sample sets Analyzes the entire population

Requires careful selection of a representative sample No need to create curated sample sets

Dependent on skilled researchers Dependent on efficient and accurate data

Runs the risk of human error No scope for human error

Requires subjective evaluation which runs Objective nature of evaluation so less

the risk of including biases scope for biases

03
RFM SEGMENTATION: WHAT IT MEANS
FOR YOUR BUSINESS GROWTH

Changing customer needs in the digital era are When we do RFM, personalized segmentation
forcing brands to adopt a more personalized, right, we are able to answer more insightful
insights-led approach to segmentation. Insights questions, such as:
gleaned from data help us build one-to-one
engagement across channels and touchpoints • How can we use the data to test new marketing
with each customer. We use these insights to messages and content to increase conversions?
identify the right customers to engage with and
execute more targeted campaigns that minimize • Is it possible to grow brand value, customer
errors, improve ROI, and delight customers with satisfaction, and engagement based on
relevant, in-the-moment messages. An RFM our segmentation?
segmentation approach like this taps into the data
on a deeper level, and at scale, to forecast • Can we find new opportunities to upsell,
customer preferences and engage at precisely the cross-sell and monetize new
right time. marketing opportunities?

04
This advanced approach to segmentation helps marketers develop multiple levels of insights:

Customer Insights Journey Insights Campaign Insights

Which customer segments are

What’s the customer’s Where are you losing your engaging well and which aren’t
favorite category? users in the customer journey? for a given campaign

Which set of customers are How are your customers Are we sending too many
loyal, price-sensitive or navigating your app/website? notifications to our customers? If
need attention? yes, to which ones?

Which set of customers are Which customer cohorts are What’s the long-term retention
going to churn, convert, or NOT converting? of customers who engage
go dormant? with campaigns?

When is the best time to Which customer cohorts have What’s the uplift on retention,
deliver a message for bad retention? conversions, or engagement
a customer? against control groups?

THE IMPORTANCE OF RFM SEGMENTATION

When traditional segmentation falls short of a marketer’s needs, a compelling strategy to employ is
what’s known as “RFM” or “Recency, Frequency, and Monetary value” to generate more valuable
customer insights.

These three metrics are strong indicators of a customer’s actual data-based behavior. The frequency and
monetary value affect a customer’s lifetime value, while recency affects retention – a strong measure
of engagement.

RFM criteria are defined as follows:

• Recency: the number of days since a customer’s last purchase. Typically, this value is defined in days.
For example, if the customer’s last order was 42 days ago, then their recency is “42.”

• Frequency: the number of orders placed in a given time period. If a customer has placed seven orders
over the course of one year, then their frequency is “7.”

• Monetary Value: the total amount of money spent by a customer over a given time period. For
example, a customer that has placed five orders of $50 each over the course of one year would have a
monetary value of $250.

05
 INSIGHTS GATHERED USING RFM MODELLING
Frequency
Recency Monetary

How recently a user How frequently a How much a

interacted with the user interacts user spends
website/app
Ex: Total number of Ex: Customer
Ex: Days since last days when a Lifetime Value
purchase/visit purchase/visit
was done

RFM segmentation is a highly effective tool RFM allows us to ask (and answer) deeper
because actual customer data, not assumptions questions than traditional segmentation, such as:
fundamentally drive it. It relies on historical or
behavioral data analysis to predict customer • Who are my best customers?
behavior and preference, and divides customers
into homogeneous groups. Instead of relying on a • Which customers have the potential to buy at
representative sample, the entire customer a higher price point?
population can be considered in the analysis. As
we rely more on aggregate data, we remove the • Which customers are most likely to churn
subjectivity from traditional segmentation or lapse?
methods. Each customer’s interaction with a
brand’s product can be analyzed to help create • Which customers can be converted
better insights and opportunities for engagement through promotions?
and create a potential for increased lifetime
value (LTV). • Which customers are likely to be loyal in the
near future?

06
REIMAGINING PERSONALIZATION
USING RFM SEGENTATION
There is an established step-by-step approach A customer may have a score of five (most
that companies can use to implement a successful valuable) for recency because they have
RFM segmentation strategy. These steps bring purchased within the last two days. That
together the data required to accurately segment customer may score a two in frequency because
customers and empower contextual, it is only their second purchase. Finally, they may
insights-driven engagement strategies. Of course, score a one in monetary value because their total
the first step is to collect and collate the relevant purchases to date are very low. Once customers
data: customer lists, purchase histories, lifetime have all been scored across the three metrics,
value, and other relevant information. This data these values can be combined to create an
will be used to establish the scales that will create aggregate score.
the appropriate consumer buckets.
However, companies could choose a custom
Leveraging an AI-driven customer engagement weighing scale based on priorities, such as
platform with built-in RFM analysis will do this increasing the importance of monetary value by
hard work for you and help surface the right 50 percent to identify highly valued customers
insights. Companies that utilize RFM with better spending potential. With these
methodology rank all of their customers for each scores, teams can assign all customers to an
of the three key categories: recency, frequency, appropriate consumer bucket that will help
and monetary value. inform which strategies will be most effective.
The end results in a more personalized look at
The scoring is a simple one to five scale, where customers and customer segments, driven by
one is the least valuable and five is most valuable. empirical data and behavioral insights. This
With a five-point scoring scale, the top 20% of allows brands to create valuable new customer
customers will receive a five, the next 20% will groups to market to.
receive a four, the next 20% a three, and so on.

07
A. CREATING MORE EFFECTIVE SEGMENTS BASED ON RFM SCORE

Customers who are showing similar behavior on R, F, M scores, and RFM aggregate scores, are grouped
into the same RFM buckets or segments. These segments are identified and named with respect to
customer behavior. Here’s the four-step framework to create your segments.

Step 1: Collect and collate relevant data/values

Customer Recency Value Frequency Value Monetary Value

#1 10 1 5500

#2 8 4 800

#3 4 7 2400

#4 3 11 9700

#5 1 2 4300

Step 2: Set the RFM scales

Score Recency Scale Frequency Scale Monetary Scale

1 Last bought a year ago Bought once a year Less than 1000

2 Last bought 9 months ago Bought thrice a year 1001 - 3000

3 Last bought 6 months ago Bought 5 times a year 3001 - 5000

4 Last bought 3 months ago Bought 7 times a year 5001 - 7000

5 Last bought 1 month ago Bought 10 times a year 7001 - 10000

Step 3: Assign scores

Customer Recency Score (Value) Frequency Score (Value) Monetary Score (Value)

#1 1 (10) 1 (1) 4 (5500)

#2 2 (8) 3 (4) 1 (800)

#3 3 (4) 4 (7) 2 (2400)

#4 4 (3) 5 (11) 5 (9700)

#5 5 (1) 2 (2) 3 (4300)

08
Step 4: Label segments
RFM Model | User Transition

Price sensitive About to Sleep Lost Hibernating

Recent Users Needs Attention Loyal Potential

Customers Loyalist

Can’t Lose Them Champions

At one end of the spectrum, we have the most Other customers should be targeted quickly to
active and valuable customers, such as: avoid losing their future business. Some of these
will be categorized as:
• Champions - Those who visited most recently,
visited most often, and spent the most. • Needs Attention - Those who spent a good
amount at one point in time but have not
• Loyal Customers - Those who visited recently, visited recently.
visited often and spent a great amount.
• About To Sleep - Those who are scoring below
• Potential Loyalists - Those who spent a average recency, frequency, and monetary
significant amount of money. values, and are apt to disappear.

• Can’t Lose Them - Those who spent a great The final group may take some additional analysis
amount of money and visited often but and focus if the brand wants to regain their
not recently. attention again or increase their spending:

Beyond the obviously profitable customers, there • Price Sensitive -Those who visited recently,
are also those that have the potential to become and also often, but do not spend much
more loyal: (indicating a hesitancy based on price).

• Recent Customers - Those who visited most • Hibernating - Those who have not visited
recently, but not as often and have not recently, nor have made many valuable
spent much. purchases. For E-commerce brands, this could
be a customer who buys from you during the
• Promising Customers - Those who are average holiday season, but not throughout the year.
in their recency, frequency, and monetary scores,
but still have the potential to buy more and • Lost - Those who have the lowest recency,
more often. frequency, and monetary scores.

09
B. DEVELOPING PERSONALIZED ENGAGEMENT STRATEGIES
Once customers have been assigned to the appropriate category, engagement strategies can be
developed and targeted for each group, such as:

• Champions - Reward these customers. They promote your products, can be early adopters for your
new launches, and are brand ambassadors

• Loyal Customers - These customers are responsive to your promotions. The brand should suggest
higher-value products and ask the customer for reviews.

User Properties User Behavious Custom Segments

Has executed Extended_subscription at least

1 times in the last 30 days

with attributes: Paid Subscriber

+ Add Attribute Hey Jon, Review us!

**Tracked when a user visits a web page. Thanks for upgrading your subscription.
How about rating our anti-virus on
Google Play? Review us today and get
AND OR
15% off on your next upgrade.

User Properties User Behavious Custom Segments

Loyal customer x is

Plus plan member

NOTE: The segmentation is a blend of different categories of data like — User Behavior: extended_subscription in the last 30 days |
User Attributes: Paid Subscriber | User properties: <name of the customer> who is Plus plan member

• Potential Loyalist - Engage these customers with long-term offers like loyalty programs or
membership rewards. Suggest other categories of products that may be of relevance to them.

• Recent Customers - For new and more recent customers, make their onboarding experience smooth
and provide assistance as needed.

10
 User Properties User Behavious Custom Segments

Has executed New_paid_subscription at least

1 times in the last 30 days

with attributes: Paid Subscriber

Hey Dany, Welcome Onboard!
+ Add Attribute Thanks for subscribing to our paid plan.
Would you want us to give you a small
**Tracked when a user visits a web page.
demo on how you can protect your
devices better?

AND OR

User Properties User Behavious Custom Segments

New customer x is

Basic plan member

NOTE: The segmentation is a blend of different categories of data like | User Behavior: paid_subscription in the last 30 days |
User Attributes: Paid Subscriber | User properties: <name of the customer> who is Basic plan member

• Promising - Make them loyal by creating brand awareness and giving free trials or access to
specific promotions.

• Needs Attention - The brand should bring back these customers by providing limited-time offers and
recommending content, products or solutions based on behavioral or purchase history.

User Properties User Behavious Custom Segments

Has executed Free_subscription at least

1 times in the last 30 days

with attributes: Free Subscriber Hey Arya,

+ Add Attribute Become our Paid member today!
We are giving an exclusive 10% off on our
**Tracked when a user visits a web page.
basic plan to you. The offer will last for
only 7 days so grab your discount today.

AND OR

User Properties User Behavious Custom Segments

Free customer x is

Free member

NOTE: The segmentation is a blend of different categories of data like | User Behavior: free_subscription in the last 30 days |
User Attributes: Free Subscriber | User properties: <name of the customer> who is Free member
11
• About to Sleep - These customers will be lost if not reactivated. Recommend popular products to them
while providing discounts for memberships.

• Price Sensitive - As customers continue to look for the best deal, the brand should recommend the
highest rated products and send deals or discount offers to re-engage.

• Can’t Lose Them - Listen to these customers’ feedback, suggest newer products, and provide reasons
for them to stick with your brand.

• Hibernating - Continue to make recommendations for products in relevant categories, and provide
personalized offers. Make sure to communicate more often during the periods these customers are most
active with your brand.

User Properties User Behavious Custom Segments

Has executed Free_subscription at least

1 times in the last 45 days

with attributes: Free Subscriber

+ Add Attribute
Hey Sam,
**Tracked when a user visits a web page.
Here's the link our latest blog on
Freemium vs Paid Model and why you
should opt for our paid subscription.
AND OR

User Properties User Behavious Custom Segments

Free customer x is

Free member

NOTE: The segmentation is a blend of different categories of data like | User Behavior: free_subscription in the last 45 days |
User Attributes: Free member | User properties: <name of the customer> who is Free member

• Lost - Make your presence known through unique and ongoing campaigns, try to re-engage them with
content or products they previously engaged with in the past.

12
HOW A BILLION-DOLLAR GLOBAL
SECURITY SOFTWARE BRAND
EXECUTES RFM SEGMENTATION
This case study details one brand’s journey Scoring buckets were created using a scale of 1
implementing a test using the RFM to 3, 4, or 5 based on required granularity, with 1
segmentation strategy. being the highest. Each bucket was then labeled
based on scoring:
Before getting started with RFM segmentation,
this brand had to define the RFM variables for
their brand. • Loyal customers receive scores of 1-1-1
across the board (the highest scores in each
category)

1. Recency was defined by the • Potentially loyal customers receive a 3-4/5

difference between the current date
and the last date their customer was
seen on the app.

2. Frequency was defined as the

OUTCOMES FROM THE
client’s initial subscription length. RFM MODEL
3. Monetary value was defined by the
customer’s subscription type and its Using the segmentation described above, this
value in US dollars. brand was able to divide its customer base into
eight categories. More than 92% of their
customers fell into just four of these categories.
The brand’s hypothesis was that by segmenting its Accordingly, the security software brand used
audiences into distinct homogeneous groups this data to focus its efforts on Needs Attention,
based on their RFM scores, they would better Hibernating Customers, Recent Customers, and
understand customer behavior in their mobile Loyal Customers. With these new strategies, the
app, which features are most engaging, and the brand expects to see 1.5% conversions, i.e., email
potential to convert and develop a addition and at least a 3% click-through rate
strong relationship. (CTR) rate.

13
Based on scoring, messaging campaigns would be tweaked based on groups/cohorts to decide which
message, content, and number of campaigns to run. Messages would vary between educational
messages, promotional messages, and upsell/cross-sell campaigns. Older campaigns could also be
re-used for certain new groups identified in the RFM model. The results of this short-term experiment far
exceeded the original goals and expectations:

CTR CONVERSION

Needs 1.89%
Attention
10.57%

CTR CONVERSION

8.16% 3.27%
Hibernating

CTR CONVERSION

Recent 2.07%
Customer
18.34%

CTR CONVERSION

Loyal 2.04%
Customer
22.45%

14
Overall click-through rates exceeded the brand’s Another example is of a push notification sent to
goals by 5-18%. Additionally, conversion rates the loyal customer segment that asked them to
were up in all four areas, showcasing the value of add on a second email for security monitoring.
the RFM segmentation model. The bottom line
here is that by utilizing RFM segmentation, the
brand was able to perform a highly data-driven
analysis to produce better insights that led to
better click-through and conversion rates.
Add in a second email for
Based on the data and focused segmentation, the monitoring
brand was able to revise messaging on a more
Find out with always-on
personalized level. For instance, below is a cyber monitoring
screenshot of a push message sent to ‘low
Get Started
scoring’ segments to promote Dark
Web monitoring:

Moving forward, the brand can now run

optimized customer lifecycle campaigns with
Is your email on the dark web? personalized messages and offers for each
Find out with always-on individual, maximizing the return of their
cyber monitoring marketing efforts.
Get Started

A NEW ERA OF PERSONALIZATION

We’re in the digital age of engagement, and customers are able to guide themselves through their own
buyer journey like never before. With all of that online activity going on, it is critical for marketing teams
to use tactics like insights-driven RFM segmentation to read the behavioral data and understand
customers on a deeper level. RFM adds the analytical component to insights-led engagement and
creates a more personalized interaction with each and every customer.

15
WHAT TO READ NEXT
We’re in the digital age of engagement, and customers are able to guide themselves through their own
buyer journey like never before. With all of that online activity going on, it is incumbent on marketing
teams to employ tactics like insights-driven RFM segmentation to read the behavioral data and
understand customers on a deeper level. RFM adds the analytical component to insights-led
engagement and creates a more personalized interaction with each and every customer.

Watch it now
Read More

Read More

ABOUT
MoEngage is an insights-led customer engagement platform, built for the customer-obsessed marketer.

With AI-powered customer journey orchestration, personalization capabilities, and in-built analytics,
MoEngage enables hyper-personalization at scale across mobile, email, web, SMS, and
messaging channels.

Over 1,000+ brands across 39+ countries use MoEngage to send 80 billion messages to over 1 billion
consumers every month. We serve enterprises such as Deutsche Telekom, Samsung, Ally Financial,
Vodafone, and McAfee along with internet-first brands such as Flipkart, Ola, OYO, and Bigbasket use
MoEngage to orchestrate efficient customer engagement.

MoEngage was recognized as a Leader in the Gartner Magic Quadrant 2020 for Mobile Marketing
Platforms and a Strong Performer in The Forrester Wave™ Report for Cross-Channel Campaign
Management (Independent Platforms), Q3 2021. G2 also named MoEngage a Leader in the Fall 2020
Grid® Report, and the #1 Mobile Marketing Platform in the Spring 2021 Momentum Report.

Get a Demo of MoEngage Today

16