Analytical CRM
PRESENTED BY-
ABHISHEK DUBEY
BALRAM
VAIBHAV GUPTA
JAY KUMAR NAHATA
Topics covered in the presentation
■ Segmentation and selection
■ Retention and cross-sell analysis
■ Effects of marketing activities
■ Reporting results
Every customer is unique
Every customer is unique
One-to-one marketing justifies this fact
Concept not realizable
Problems in approaching customers separately
Reasons
Capacity and customer insight to do this are
lacking
Benefits do not appear to justify the costs
Segmentation
Segmentation
Dividing customers into
Distinct homogenous groups
Can be approached in specific ways
Common aspects
Needs and wants
Purchase behavior
Communication consumption
Segmentation criteria
Used to divide up the market
Basis to identify or assign into groups
Groups are not profiled at this time
Buyer
User
Individual
Organization
Kotler
1. Geographic
Postal code
2. Demographic
Age / Sex
3. Socio-economic
Social class / Education / Income
Kotler
4. Behavior
Purchase
Communication
5. Psychographic
Lifestyle
6. Buying motives & purchase
considerations
Anna
35 years old Drives a minivan
Married Works at home
Two children Subscribes to Sunday
newspaper, buys
Owns home occasional single-copy
Newcomer
Home value
$175,000
Interests:
$100,000+ HH
Food, Wine
income
Parenting
Reads Business
Week Travels out of state
& Sports Illustrated three times a year,
Plays tennis and golf international travel
once a year
Bonoma & Shapiro
1. Demographic
Industrial classification / Co. size / location
2. Operating variables
Technology / User status / Customer support
3. Purchasing approaches
Organization / Nature of relationships maintained
Bonoma & Shapiro
4. Situational factors
Urgency / Specific applications / Order size
5. Personal characteristics
Values and norms of employees
Risk attitude of customers
Loyalty to company
Segmentation technique
Markets can be segmented in a number of
ways
Basis to identify or assign into groups
Choice
Number
Specific criteria
Kotler
1. Measurability
Size / Purchase power/ Characteristics
2. Substantial
Segments are large and profitable
Largest homogenous group for a tailored
marketing program
3. Accessibility
Segments can be reached and served effectively
Kotler
4. Differentiable
Conceptually distinguishable
Respond to different marketing stimuli
5. Actionable
Effective programs can be made for attracting and
serving customers
Cluster solution
10
0
0 1 2 3 4 5 6 7 8 9 10
Clustering
The art of finding groups in data
Objective
Gather items from a database into sets according to
(unknown) common characteristics
Much more difficult than classification since the classes
are not known in advance (no training)
Profiling a segment
How to develop
Accurate
Complete
Current
Unique customer
profile
Important
Identification
Enriching profiles
Developing Marketing service concepts for each
segment
Segmentation
Aids in formulation of marketing strategy
Goal
Approach customer groups in differentiated manner
They become more satisfied and loyal
They spend more with the supplier
Marketing service
Differentiation
Product range
Service
Price
Communication
Service concept can gain in quality
Who is the person ?
What motivates them ?
Segmentation research
Objective
Who do we want to approach for a certain
marketing campaign activity
Segmentation research
Three techniques
RFM : Recency Frequency Monetary Value
CHAID : Chi Squared Automated Interaction
Detection
CART : Classification And Regression Trees
Recency Freq Monetary Value
Background
Historic behavior is better predictor than purchase
intention and attitude
Customers who spend most during certain period
may not do so during a new marketing campaign
Recency Freq Monetary Value
Scores on three aspects
Last purchase date
Purchase frequency
Amount spent
Last purchase
Scores ( weight 5 )
< 3 months : 20
3 – 6 months : 10
6 – 9 months : 3
> 12 months : 1
Purchase freq / Amount
Purchase frequency ( weight 3 )
No. of purchase in last 2 – 4 years x 4
Maximum score of 20
Amount spent ( weight 2 )
10% of sum of purchase amounts in the last 2 years
Maximum score of 20
RFM Example
RFM
Advantage
Past behavior is more accurate predictor
Uses transaction database
Disadvantage
Best buyers selected for promotion
They will experience excessive ‘Mail’ pressure
2. CHAID
CHAID
Chi-squared Automated Interaction Detection
Produces a tree diagram
Creates categorization within each group
CHAID
N = 240000
Response 4.36%
0-30 Min 30-60 Min
N = 80000 N = 160000
128% 86%
Registration Duration Folder Share
< 12 Years > 12 Years < 50% > 50%
N = 55000 N = 25000 N = 60000 N = 100000
150% 80% 105% 75%
CHAID
3. CART analysis
CART
Classification & Regression Trees
Produces a tree diagram
Segment A is separated from Segment B
Segmentation variables applied to sub-divide
Retention and cross-sell analysis
Retention & Cross selling
Retention & Cross selling
Retention question
Which customers run an increased risk of ending
relationship
Cross selling
Customers may be stimulated to buy another
product
Why retention ?
Customer satisfaction ?
Why retention ?
Retention
Opportunity
Greatest gains may be realized
Closing back door costs less than enticing
customers with designer front door
Benefits
More than a cross-sell or a deep-sell exercise
Life time value is secured
Retention
Meaning
Holding on to the customers
Determinant
Definition of former or current customers
Does someone become a departing customer at the
moment he / she no longer buys a certain product
Eg. Fairness cream
Retention
MarketWhysTM
The Disposition Scale
INSIST The only brand consider using Proportion
of Preferrers
PREFER One of the brands I prefer using
INTEND Not one of the brands I prefer but would like to
try it
ACCEPT
I’d use it in certain circumstances only
NO OPINION
Heard of it but don’t know much about it
NOT AWARE
REJECT Never heard of it
Would never willingly buy it
Example - FAL
Leakage analysis
Last
bght /
Last 12 Last 6 Last Last 12
Base : All random resp (503) mnts mnts Bought MOUB mnts
Any Fair & Lovely 90% 88% 77% 76% 85%
Fair & lovely Multi vitamin 75% 72% 61% 60% 81%
Fair & lovely Ayurvedic 27% 20% 11% 10% 43%
Fair Ever 12% 7% 4% 4% 37%
Fair One 12% 6% 4% 3% 31%
Fair One Plus Five 1% 0% 0% 0% 25%
Ponds white beauty 14% 9% 4% 4% 26%
Vicco Turmeric 16% 11% 5% 3% 29%
Any Garnier 10% 8% 5% 5% 54%
Fair & lovely Skin clarity 5% 3% 2% 2% 38%
Fair & lovely Active sun-block 3% 1% 1% 1% 31%
Emami natural fair 2% 1% 0% 0% 9%
Share of usage
Last Bought Vs Next FAL FAL Multi Fair
Purchase Ayurvedic Vitamin Ever Fair One Garnier
Base:All 53 303 21 80 31
FAL Ayurvedic 91 1 0 1 3
FAL Multi Vitamin 9 96 0 3 3
Fair Ever 0 0 95 1 0
Fair One 4 3 0 98 13
Garnier 2 0 0 0 87
• FAL Multivitamin is successful in
attracting users of other brands.
Variables
Need of reliable data
Only name, address etc. will not be enough
Intense interaction required
During relationship, need is to think of data that will
predict churn
Data availability and quality plays a critical
role
Reasons for not continuing use of Fair One
Base: Respondents who have used in the last 1
37
yr, but not bought on last occasion
Product Issues 62%
Skin remains dry 27%
Doesn't make skin soft 5%
Smell is not good 5%
Oiliness remains on face 5%
Face is not clean after using it 3%
Has more bleach 3%
Gives patch look -
Cream stick on the faces -
Cream is not soft -
Causes side effects 24%
Causes rashes 14%
Causes wrinkles 5%
Causes side effects 3%
Does not make you fair 3%
Causes Pimples/Acne/Spots -
Have skin problems after using it 3%
Causes spots on face 3%
Others 22%
Satisfied with my regular brand 11%
Causes sweat 5%
New Cream -
Does not suit my skin 5%
Prefer Other Brands Page 47
Investment company
Data requirement
Investment behavior
Number of different products purchased by
customer
Financial climate
Rate of return
Customer satisfaction
Investment objectives
Communication behavior
Examples
Loss in stock market
Investor 1 - Invest in bank FD
Investor 2 – Invests more in equity
Differences
Active & non-active customers
Loyal & disloyal customers
Sedtec
Managing complaints
Cross-selling
Cross-selling
Definition
Sale of products to current customers who are
already purchasing one or more products
Growth phase
Engaging in cross-selling ensures continuity and
further development of relationship
Cross -selling
Opportunity
Reduces likelihood of a customer ending the
relationship prematurely
Benefits
Transaction profit
Boost to relationship
Cross –selling - POV
Product
List of existing customers with history of
purchasing certain offering
Client
When customer contacts organization, it is possible
to estimate which product has best chance of being
cross-sold
Cross-selling over time
More than one of same product in
contact
2 life insurance policies
Two or more of different products in
contact
Home insurance
Life insurance
Howz that ?
Effects of marketing
activities
What response ?
What response ?
Basic question
How do we measure response of
marketing activities
How many people respond to a mail, a phone call
or an invitation on a website ?
Did it have a positive effect on customer value ?
Different responses ?
Marketing activity
Selection of customers critical for
Cross-selling
Retention
Goal : Serve customers
Standardised marketing approach
Sales process analysis
+/- 180 10 3 1
Order
Select 1st Quotation
Data Call for Visit Call
base appointmen Negotiation
t
Sales lead time – 10 weeks
Sales – Transaction perspective
Order
Select 1st Quotation
Data Call for Visit Call
base appointmen Negotiation
t
■ Critical
Logic of step in the sales process
Conversion between the steps through time
Allocation of resources at each step
Marketing or sales cost of transaction
How to increase value
Increase effectiveness of sales /
marketing
Ways to shorten throughput time of the sales
process ?
Eliminate one step to speed up the process ?
Can we make better selection of prospects ?
Will training of sales people help increase
conversion rate ?
Have we applied the appropriate channels ?
Effectiveness of targeting
A step to relationship marketing
Who do we select for a specific sales process ?
Benefits
Allocation of marketing and sales resources to
segments, products, acquisitions of new customers
and retention of existing customers
Effectiveness of targeting
Lifetime value
Each interaction
Takes relationship one step further
Influences future interactions
Direction of relationship
Type of interaction
A welcome call
A complaint
Reward for continuing relationship
Price discount during sales week
Learning organization
Dialogue with customer is important
Knowledge about him
Used in developing relationship
Knowledge created from experience gained
Timing
In a real time market, short response times are a
pre-condition to success
Examples
Low business from long term customers
Additional acquisition efforts required
New prospects
Offer incentives to become customers
Learning
Agent in contact centre
Learning
Service repairman in contact with the
customer
Learning
Learning
Account manager
Learning
Marketer
Learning
Database manager
Learning
Task of management
Encourage learning process and not
punish it
Motivation and inspiration must come from
individual
Agents in call-centre
Unable to convert numerous experience into
knowledge
Share it internally / externally
Mental models
Construction of an image of reality
Bring reality to surface
Hold it up to light
Open for discussion
Images of others should not be renounced
Shared vision
Common answer to ‘why are we doing
things this way’ to be developed
Every department should have a goal to
achieve together
Service employees are not to minimize cost
Salesmen are not to maximize turnover
Finance is not limited to reduction of WC
Team learning
Develop vision and knowledge into
actions
Act collectively and learn from experience
Dialogue
Assumptions removed
Free exchange of information becomes possible
Eg. A slow agent in a call centre may not make
much sales but can give valuable customer
information
Team learning
Life Time Value
What matter is …
Not transaction profit
One transaction is not sufficient
… but lifetime value
Customer’s profit contribution
Very few companies arrive at reliable calculation of
LCV
Issues
Difficult to calculate
Economies and market changes
Loyal customers become less loyal and profitable
Investments designed results in losses
Necessity
More MIS
Insights into factors that bring final results
Lifetime Value
Definition
Net present value of the future contribution by a
customer to the overhead and profit of a company
Accurate calculation
All income and expense are allocated to each
customer
Example
The sum of the future profits yielded is 523
Assuming a discount rate of 10%, the CLV at moment 0 is 398
CLV
Prospects Customers Retained Retained Retained
Customers Customers Customers
$ $ $ $
Discount Factor
Divide by Number of Initial Customers
= Customer Lifetime Value
Lifetime Value
Formal definition:
The total net incoming a company can expect from
a customer
Abstract definition:
Future profit from a customer
OR how much the customer is worth now
Depends on many factors
Subscriber (telecom) vs. visitor (e-commerce)
Business Uses
Numbers are of interest for marketing
For effective decision support need LTV
before and after marketing activities
Retention, incentive allocation, acquisition etc.
Profit estimate
LTV after – LTV before – cost
Calculation
Calculation
Customer turnover
- Discount granted
+ Shipping cost passed on to customer
+ Supplier’s credit
Gross turnover
- Turnover from returns
Net order sales
- Cost of goods / service sold
- Administrative / Physical Order processing cost
- Admin / Physical Order cost of processing returns
- Bad debt expenses
- Cost of acquisition & relationship management
Customer’s contr’ to O/H and profit of organisn’
Example
Difficulties
Insights
Actual expenses
Reliable cost estimates
Acquire, serve & manage relationships
Identify customers
In order to arrive at actual calculation
Intermediaries involved
Resolution
1. Value calculation
Segment level instead of individual customer level
2. Restricted time period
Expected income or expense only for first two or
three years
1. Segments-based approach
1. Segments-based approach
Segment is basic entity in marketing strategy
Based on elaborate research and modeling
Decisions (campaigns, incentives, channels) are done
at segment level
Segments implicitly assumed “homogeneous” for
specific property considered
So for marketing purposes we really only
need LTV at the segment level
2. Time Horizon
2. Time Horizon
Theoretically, the horizon should be
infinite. It is unmanageable in the reality
Long-term relationship is important
Take a long horizon, e.g. 10 years
Short-term relationship is important
Take a small horizon, e.g. 1 year
In the empirical application
Use a horizon of 2 years
Estimating Future Value
This is a forecasting problem
Seasonality effects
Market changes, competitor activity
Common solutions
“Leave to experts”
Use current customer value from current data
Segment-based approach makes it easier
Need only “average” correct forecast
RFM Variables
Recency
the most recent date that the customer has requested
for a service (usually a purchase )
Frequency
The number of time the customer has purchased
Monetary
Total dollar amount that a customer has spent
RFM Variables
3.0
2.5
2.0
Number of
purchases per
year 1.5
1.0
0.5
0.0
1 2 3 4 5
Years as a customer
RFM Variables
$70
$60
$50
Average
PurchasePrice $40
$30
$20
$10
$0
1 2 3 4 5
Years as a customer
RFM Variables
90%
80%
70%
Percentage 60%
Retained from
Previous Year 50%
40%
30%
20%
10%
0%
1 2 3 4 5
Years as a customer
Why satisfaction is important ?
■ Study by Le Beouf:
“The reasons why customers no longer dealt with a
particular supplier”
Benefits of CLV
By knowing the CLV, one can
Focus on groups of customers of equal wealth
Evaluate the budget of a marketing campaign
Measure the efficiency of a past marketing campaign by
evaluating the CLV change it incurred
Focus on the most valuable customers, which deserve
to be closely followed
Neglect the less valuable ones, to which the company
should pay less attention
Pareto’s Principle
Buford Electronics
Details
Electronics company
US $ 350 Mio turnover
Customers segmented into groups
Acquisition and relationship management policies
Customer-Focused Initiatives
Customer Champion
Create a customer champion Director who
is responsible for championing the voice of Customer Value Attache
the customer through the organisation. Nokia product engineer goes on-site with
customer for up to 1 month to learn about
challenges and show how Nokia can add
value
Everyday Life Observation
To gain deeper understanding of customer, send Customer Charter & Advocate
video crews & TV cameras into 80 households Independent customer advocate whose role is to
around world to capture customer daily routines resolve particularly difficult customer and
business problems. Customer charter to improve
customer experience with service.
Customer Success Engineer
Team Customer Partner Experience
Centralised group that diagnoses root causes of Organisation-wide customer and partner
complex customer problems and implements satisfaction index to provide a holistic view of
solutions across business business health and trigger specific corrective
actions where necessary.
Conclusion
End result
Development of relation and value
Keeping customers is less expensive than attracting
new customers
CLV
A customer is an asset that represents value
Cash contribution to profits and O/H
