customer relationship management- crm 3.pdfcustomer relationship management- crm day 3: measuring...
TRANSCRIPT
drvkumar.com© Dr. V Kumar 1
Dr. V KumarRichard and Susan Lenny Distinguished Chair Professor of Marketing,
Executive Director, Center for Excellence in Brand & Customer Management,
Director of the Ph.D. Program in Marketing
J. Mack Robinson College of Business
Georgia State University
Customer Relationship Management- CRM
Day 3: Measuring and Maximizing Customer Lifetime Value (Part I)
drvkumar.com© Dr. V Kumar 2
Customer Loyalty and CLVMeasuring CLV (Contd.)
drvkumar.com© Dr. V Kumar
Case Study – Revenue Analyses“Wyndham ByRequest” Member versus Non-member comparisons
Q1 2001 Q2 2001 Q3 2001 Q4 2001
New Members added 25,000 50,000 55,000 50,000
Members
Total nights 104,491 119,624 127,505 141,669
Total stays 36,116 40,554 45,049 46,197
Total guests 26,616 30,811 35,433 36,955
Total revenue
produced
$12,403,161 $14,090,471 $15,099,372 $14,830,528
Rev/member 466.00 457.31 426.13 401.31
Non-members
Total nights 1,548,501 1,918,634 1,429,460 1,040,763
Total stays 406,536 441,865 392,469 338,459
Total guests 352,977 383,051 336,263 291,399
Total revenue
produced
$163,294,362 $164,863,345 $130,685,357 109,404,522
Rev/Non-member 462.62 430.39 388.64 375.45
3
drvkumar.com© Dr. V Kumar
Co-relation Between Guest Membership,Loyalty and Revenues
• Only 1 in 5 Guests across business, conference and leisure segments return to Hilton because of program membership in a given year
• Wyndham spent approximately US $ 50 Million in FY 2001 to attract 180,000 new members to ―Wyndham ByRequest‖ Loyalty Program
Did Wyndham’s aggressive campaign result in proportional increase in retention and revenue?
4
drvkumar.com© Dr. V Kumar
Brand Preferences and Business Travel Decision Process
Hotel Selection Decision Elements
Influential in %
2002 2001 2000 1999
Location 91 89 83 86
Previous experience with hotel 89 94 84 86
Value for the price 82 86 76 72
Reputation of hotel/chain 82 80 74 71
Room rate 72 81 75 62
Recommendation of friend/associate 69 70 68 68
Likelihood of upgrade to better accommodation 63 N/A N/A N/A
Brand name 59 53 62 66
Gives both airline mileage and points 51 52 N/A N/A
Airline mileage 37 44 26 25
Recommendation of travel agent 34 35 40 34
Frequent-guest points 29 42 31 26
5
Source – The YP&B Yankelovich Partners National Business Travel Monitor (2000 and 2002)
drvkumar.com© Dr. V Kumar
Total Membership/Program Characteristics by Brand
6
Brand program Established Total
MembersC
How Points
are Earned
based on
Dollar Value
Approximate
Number of
Participating
Properties
Number of
Points
redeemed for
Free Three-Day
Vacation D
Marriott
Rewards
1983 15.5 10 points per
dollar spent A
2,100 50,000
Starwood
Preferred Guest
1986 15 2 points per
dollar spent B
700 21,000
Hilton HHonors 1987 12.7 10 points per
dollar spent
2,100 65,000
Hyatt Gold
Passport
1987 10 5 points per
dollar spent
363 36,000
A – Only room rate at Courtyard, Fairfield Inn and Springhill Suites
B – 3 points for higher-status members
C – In millions
D – Estimated Source - HBR
Dollars Spent
for Free
Three- Day
Vacation
5,000
10,500
6,500
7,200
drvkumar.com© Dr. V Kumar
Segmenting Customers Based On Past And Future Profitability
7
Rising Stars
ACTION:
- Invest to deepen relationship
- Identify specific up-sell / cross-sell opportunities
- Cultivate attitudinal loyalty
True Loyalists
ACTION:
- Cultivate attitudinal loyalty
- Invest to nurture / defend / retain
- Reward proactively
Total Misfits
ACTION:
- No relationship investment
- Aim to extract profit from every transaction by
migrating the customer to low cost channels
Falling Angels
ACTION:
- Identify specific up-sell / cross-sell opportunities
- Transact through low-cost channels
- Optimize (Minimize) Marketing costs
Historical Profits
(PCV)
Fu
ture
Pro
fita
bilit
y
(CLV
)
Low High
Low
High
drvkumar.com© Dr. V Kumar
The Managerial Implications
• Understand loyalty diversity and profitability diversity
• Manage loyalty and profitability simultaneously
• Make forward-looking customer investment
decisions, including projected customer profitability
8
drvkumar.com© Dr. V Kumar
Customer Loyalty and CLV
9
drvkumar.com© Dr. V Kumar 10
drvkumar.com© Dr. V Kumar
Customer Lifetime Value (CLV)
DEFINITION
• Customer lifetime value is defined as the sum of cumulated cash
flows—discounted using the Weighted Average Cost of Capital
(WACC)— of a customer over his or her entire lifetime (three years in
most cases) with the company.
11
Approach to Measurement
Recurring
Costs
Recurring
RevenuesNet
Margin
Expected Number
of Purchases over
next 3 years
Accumulated
Margin
Acquisition
Costs
Customer
Lifetime
Value- X
-
Gross
Contribution
Margin
Marketing
Costs-
drvkumar.com© Dr. V Kumar 12
E
N
H
A
N
C
E
D
D
A
T
A
C
O
L
L
E
C
T
I
O
N
C
U
S
T
O
M
E
R
&
S
U
P
P
L
Y
S
I
D
E
F
A
C
T
O
R
S
IMPROVED
ACQUISITION
IMPROVED
RETENTION
Dynamic
Customer
Management
Based on
CLV
IMPROVED
SHAREHOLDER
VALUE
CLV For Better Customer Management & Improved Shareholder Value
drvkumar.com© Dr. V Kumar
Why CLV?
• Forward looking metric unlike other traditional measures (that include past contributions to profit).
• Helps marketers to adopt the right marketing activities ‗today‘ to increase ‗future‘ profitability.
• It can be used to include prospects, not just current customers.
• It is the only metric that incorporates into one, all the elements of revenue, expense and customer behavior that drive profitability.
• It enforces the focus on the customer (instead of products) as the driver of profitability.
13
drvkumar.com© Dr. V Kumar
Understanding the different components of CLV
14
drvkumar.com© Dr. V Kumar
Gross Contribution Margin (GC)
15
Gross Contribution
Margin (GC)
Revenue obtained
from a customer
in a given purchase
Cost of goods
sold=
_
drvkumar.com© Dr. V Kumar
Marketing Costs (M)
Development and Retention costs
- costs of programs to increase the value of existing relationships
- costs of loyalty or frequent buyer programs
- costs of campaigns to ‗win back‘ former customers
- costs of servicing customer accounts
16
Time period (i)
This refers to the natural "lifetime" of customers
For most businesses, it is reasonable to expect customers to return for many
years (‗N‘ number of years).
drvkumar.com© Dr. V Kumar
Discount Rate (d)
• Profits you receive from your customers come in over several years
• Money received in future years is not worth as much today as money received today.
• To estimate the value of future money, we must discount it by a certain percentage.
• This equates the future money to current profits.
17
What is a good discount rate?
- Depends on general interest rate.
- Normally proportional to the treasury bill rate.
- In other words proportional to the interest that banks pay on savings
accounts.
drvkumar.com© Dr. V Kumar
CLV Measurement Scenarios
Scenario 1
• One Purchase per Year
• Eg:
Scenario 2
• Multiple Purchases per Year
• Eg:
18
drvkumar.com© Dr. V Kumar
Measuring Customer Lifetime ValueCase I- P&C Insurance Company
19
td )1(
trN
1t*M-GC CLV
•Sales take place once a year
•Annual allocation of marketing resources
•Customer Retention rate common over time
•Revenues per customer constant over time
N = Planning horizon in years
GC = Annual gross contribution margin.
M = Marketing costs per year.
d = Annual discount rate (appropriate for marketing investments).
r = Annual retention rate
drvkumar.com© Dr. V Kumar
Measuring Customer Lifetime ValueCase I- P&C Insurance Company
20
10
1]})2.1/()75[(.{*000,5000,26
tCLV
tt
= $34,681.00
Numerical Example
Consider an online insurance company trying to estimate CLV for a
customer base of 1000 customers.
Annual promotional expense = $5,000
Annual retention rate = 75%
Planning period = 10 years
Annual Gross Contribution = $26,000
Discount rate = 20%
drvkumar.com© Dr. V Kumar
• Non-Contractual Setting
• Multiple Purchases per year
• Information available on past revenues
• Decision to invest in the relationship depends on – Is the agent still active?
– Is the relationship profitable?
21
Measuring Customer Lifetime ValueCase II – Independent Agents
tT
ditACt
itAlivePi
GC
1
1*
1)( of NPV
ACit = Average gross contribution margin in year t based on all prior purchases
d = discount rate
i = agent
t = current time period
T is the number of years
P(Alive) is the probability that agent i is active
drvkumar.com© Dr. V Kumar 22
24.127$81.5543.71
)12.1/1(*100*7.0)12.1/1(*100*8.0
1
1*
1)( of NPV
21
2
t
ditACt
itAlivePiGC
Measuring Customer Lifetime Value (Contd.)Case II – Independent Agents
t
ditACt
itAlivePi
GC
1
1*
1)( ofNPV
2
ACit = 100 for Year 1 and 100 for year 2
d = discount rate (12% on a yearly basis)
i = agent
Number of years = 2
P(Alive) is the probability that agent i is active = 0.8 for year 1 and 0.7 for year 2.
drvkumar.com© Dr. V Kumar
Are there any limitations with this approach?
23
drvkumar.com© Dr. V Kumar
• Non-Contractual Setting
• Multiple Purchases per year
• Information available on past purchase behavior, revenues,
and channels of contact made to distributor
• Decision to allocate resources depends on
– How active is the customer?
– Which channel is the most receptive channel for the customer?
– Does the revenue from the customer offset the marketing
investments?
24
Measuring Customer Lifetime ValueCase III – Individual Customer Level
drvkumar.com© Dr. V Kumar 25
Measuring Customer Lifetime ValueCase III – Individual Customer Level
CLV= NPV of GC – NPV of Marketing Cost
• Since purchases can occur multiple times within a year, the gross contribution from sale is discounted to present value as is occurs
• Similarly, as the marketing cost gets expended, it is discounted to PV
• The costs are computed for each channel using the frequency of contacts in each channel and the unit cost of contact in each channel
• The future periods considered, say three years, could include multiple purchase occasions
drvkumar.com© Dr. V Kumar 26
i
i
T
t frequencyt
ti
d
GC
1
,
1
Measuring Customer Lifetime Value (Contd.)Case III – Individual Customer Level
NPV of Gross
Contribution
Margin
Where:
GCi,t = Gross contribution margin from customer i in purchase occasion t
frequencyi = 12/expinti,
expinti = expected inter purchase time for customer i
d is the discount rate for money
Ti is number of purchases made by distributor i, until the end of the planning period
• Since purchases can occur multiple times within a year, the gross contribution from sale is discounted to present value as it occurs
drvkumar.com© Dr. V Kumar 27
Measuring Customer Lifetime Value (Contd.)Case III – Individual Customer Level
NPV of
Marketing Cost
Where:
ci,m,l = unit marketing cost, for customer i in channel m in time period l
xi,m,l = number of contacts to customer i in channel m in time period l
d is the discount rate for money
n is the number of years to forecast, and
l
lmilmimn
l d
xc
)1(
* ,,,,
1
• The costs are computed for each channel using the frequency of contacts in each channel and the unit cost of contact in each channel
drvkumar.com© Dr. V Kumar
.
)1(
*
1
,,,,
11
,
l
lmilmimn
l
T
t frequencyt
ti
itd
xc
d
GCCLV
i
i
28
Where,
CLV = Customer Lifetime Value
GCi,t = Gross contribution from customer i in purchase occasion t
ci,m,l = unit marketing cost, for customer i in channel m in time period l
xi,m,l = number of contacts to customer i in channel m in time period l
frequencyi = 12/expinti,
expinti = expected inter purchase time for customer i
d is the discount rate for money
n is the number of years to forecast, and
Ti is number of purchases made by distributor i, until the end of the planning period
GC Marketing Cost
Measuring Customer Lifetime Value (Contd.)Case III – Individual Customer Level
drvkumar.com© Dr. V Kumar
)}15.1
20*3()25*3()
15.1
15*60()10*60{(
15.1
2380
15.1
3100
15.1
2200
15.1
200025.115.0
i
CLV
29
.
)1(
*
11
,,,,
11
,
l
lmilmimn
l
T
t frequencyt
ti
id
xc
d
GCCLV
i
i
Let us consider for customer i:
Prediction window (l) = 2 years
Frequency of purchase = 2 purchases per year
Unit cost of salesperson = $60
No. of sales contacts by salesperson = 10 in year 1 & 15 in year 2
Unit cost of telesales = $3
No. of contacts through telesales = 25 in year 1 & 20 in year 2
Interest (Discount) Rate = 15%
= $6583
Measuring Customer Lifetime Value (Contd.)
Case III – Individual Customer Level
drvkumar.com© Dr. V Kumar
Measuring CLV- Summary
• Contractual Vs. Non-contractual
• Single Vs. Multiple Purchases per year
• Calculation of retention rate
• Predict customer activity
• Calculation of gross contribution margin
30
drvkumar.com© Dr. V Kumar 31
CUSTOMER LIFECYCLE :
•Measurement/
Forecasting
•ROI
•Shopping Basket
•Measurement
•Prediction
•Purchase Sequence
•Enrichment
•Customer Loyalty Programs
•Forecasting
•Prevention
•Win back
OPTIMAL RESOURCE ALLOCATION
Product/Service Call MgmtMarCommPricing
Return on Marketing Investments
Acquisition Retention Attrition
Value
Time
Strategic
Impact
Traditional CLV Curve
Can we influence this graph?
drvkumar.com© Dr. V Kumar
Where do we go from here?
32
drvkumar.com© Dr. V Kumar
Case StudyImplementing the CLV metric
33
drvkumar.com© Dr. V Kumar
Background
• Fashion retailer having retail stores across USA
34
Develop a suitable metric to
measure and manage customer level
profitability
Identify the right metric to manage
customer loyalty
Challenges
drvkumar.com© Dr. V Kumar
Step 1: Identifying the Drivers of Loyalty
35
The retailer used several measures to identify loyal customers:
- Regularity of Purchase
- Frequency of Purchase
- Tenure
Regularity Frequency RFM Tenure
CLV r= - 0.09 r= 0.17 r= 0.19 r= 0.44
N 172,688 470,932 470,932 470,932
Question: Do these measures of loyalty drive profitability?
Result:
Except for tenure, the traditional metrics of loyalty showed poor correlation with loyalty.
drvkumar.com© Dr. V Kumar
Step 2: Measuring CLV
36
Where:
GCi,t = Gross contribution from customer i in purchase occasion t
ci,m,l = unit marketing cost, for customer i in channel m in time period l
xi,m,l = number of contacts to customer i in channel m in time period l
frequencyi = 12/expinti,
expinti = expected inter purchase time for customer i
r = the discount rate for money
n = is the number of years to forecast
Ti = total number of purchases made by customer i
The lifetime value is computed for each customer using this formula:
.
)1(
*
11
,,,,
11
,
l
lmilmim
n
l
T
t frequencyt
tiit
r
xc
r
GCMCLV
i
i
drvkumar.com© Dr. V Kumar
Step 3: Scoring & Segmenting the Customers
37
Customers are rank-ordered into deciles and
segmented based on the distribution of CLV across the deciles
297
92
1
1690
(16) (29) (104)(6)22.356
-200
0
200
400
600
800
1000
1200
1400
1600
1 2 3 4 5 6 7 8 9 10
Decile based on CLV
($)
High CLV
Medium CLV Low CLV
drvkumar.com© Dr. V Kumar
Step 4: Identifying the drivers of CLV
38
Top Drivers of CLV
$ Spent in other channels (Multi-channel shopping) (+)
Tenure (+)
$ Spent in Product A (+)
Cross-Buying (+)
$ Spent in Product B (+)
Lifetime Returns (∩)
Amount of Returns ($)
CLV
Score
($)
X
B
drvkumar.com© Dr. V Kumar
Step 5: Interpreting The Impact of The Drivers
39
Lifetime Returns
$ Spent on Product B
Cross-Selling
$ Spent on Product A
Tenure
$ Spent in
other channels
A 15% increase in cross-channel spending by customers in the top 2 CLV
deciles results in 31% increase in their CLV for PRC stores.
Similar interpretation holds for the remaining variables illustrated below.
31%
14%
12%
11%
4%
-3%
-5% 0% 5% 10% 15% 20% 25% 30% 35%
% Change in CLV
drvkumar.com© Dr. V Kumar 40
Step 6: Profile Analyses
Develop profile analyses for the High and Low CLV Customers
Gender: Female
Age: 35-54 years
Marital Status: Married
Presence of Children
Estimated Household
Income: $125,000+
Stays closer to retailer
Loyalty Card Member
Mail Order Shopper
Shops frequently in
upscale stores
Typical High CLV Customer
Gender: Male
Age: 25-34 years
Marital Status: Single
Presence of no children
Estimated Household
Income: < $50,000
Stays further away from
retailer
Not necessarily a Loyalty
Card Member
Single Channel Shopper
Typical Low CLV Customer
drvkumar.com© Dr. V Kumar
Step 7: Measuring Store Profitability
41
Store
Store
Revenue ($)
Revenue
Rank
Expected store
profitability
based on CLV
Profitability
Rank
1 20,196,138 1 3,304,942 4
2 11,870,392 2 1,731,856 9
3 9,761,732 3 -4,471,439 14
4 8,705,402 4 7,635,066 1
5 6,487,945 5 2,579,805 7
6 6,314,190 6 -5,816,329 15
7 5,085,694 7 4,660,984 2
8 4,510,125 8 2,759,272 6
9 4,357,225 9 2,526,916 8
10 4,311,662 10 3,577,310 3
11 4,189,061 11 185,066 12
12 3,880,850 12 1,031,893 11
13 3,856,373 13 1,117,543 10
14 3,777,840 14 3,053,192 5
15 3,757,544 15 -348,419 13
drvkumar.com© Dr. V Kumar
Firm Value
42
Firm Level
Store Level
Customer
Level
Product
Level
drvkumar.com© Dr. V Kumar 43
CLV Based Marketing Strategies
drvkumar.com© Dr. V Kumar 44
The Wheel of Fortune Strategies Used for Maximizing CLV
Acquiring
Profitable
Customers
Customer
Selection
Preventing
Attrition of
Customers
Referral
Marketing
Strategy
Linking Investments in
Branding to
Customer Profitability
Pitching the Right Product,to the
Right Customer, at
the Right Time
Managing Loyalty and
Profitability
Simultaneously
MEASURING& MAXIMIZING
CUSTOMER
LIFETIME VALUE
Linking CLV to Shareholder
Value
Product
Returns
Future of
Customer
Management
Cross - Buy
Source: Kumar, V., “Managing Customers for Profits”, The Wharton School Publishing
Optimal
Allocation of
ResourcesManaging
Multi-channel
Shoppers
Interaction
Orientation
drvkumar.com© Dr. V Kumar
Customer Selection
To detect and target
customers/distributors based
on their value potential as
opposed to other traditional
metrics.
At the least, ROI is doubled.
45
Objective: Marketing Metric Outcome:
drvkumar.com© Dr. V Kumar
“The top 15 to 20 percent of a bank's customers
represents 130 to 200 percent of the bank's
bottom line profitability."
The First Manhattan Consulting Group
December 1997
46
drvkumar.com© Dr. V Kumar
An Empirical Illustration Showing How CLV Outperforms Other Customer Selection Metrics
47
% of
Cohort
(Selected
from
top)
Using the first 30 months of data to predict the next 18 months of purchase
behavior
Customer Lifetime
Value
Share of Wallet RFM PCV
5 Average
Revenue
489,541 255,885 308,698 357,265
Gross Profit 146,862 76,766 92,609 107,719
Variable Costs 1,270 620 1,051 790
Net Profit 145,592 76,146 91,558 106,389
10 Average
Revenue
265,664 105,394 185,557 186,124
Gross Profit 79,699 31,618 55,667 55,837
Variable Costs 751 588 775 610
Net Profit 78,948 31,030 54,892 55,227
*The reported values are in dollars and are cell medians. Gross profit is residual revenue after
removing cost of goods sold. In general for the firm that provided the database, the cost of
goods sold is around 70%. Hence gross profit is equal to revenue*0.3.
drvkumar.com© Dr. V Kumar
CLV Outperforms All Customer Selection Metrics
48
Summary of gains
Customer
Selection (Top 5% of cohort)
Increase in Net Profits
per customer(Calculated using the first 30 months of
data to predict the next 18 months of
purchase behavior)
Total increase in
Net Profits(For 2000 customers)
Break-even The cost of developing Lifetime
Value model (about $500K) can be
easily recovered from increase in net
profits from less than:
Lifetime Value
Versus SOW
$69,446 $139 Million 8 customers
Lifetime Value
Versus RFM
$54,034 $108 Million 10 customers
Lifetime Value
Versus LTD
$39,203 $78.4 Million 13 customers
drvkumar.com© Dr. V Kumar 49
The Wheel of Fortune Strategies Used for Maximizing CLV
Acquiring
Profitable
Customers
Customer
Selection
Preventing
Attrition of
Customers
Referral
Marketing
Strategy
Linking Investments in
Branding to
Customer Profitability
Pitching the Right Product,to the
Right Customer, at
the Right Time
Managing Loyalty and
Profitability
Simultaneously
MEASURING& MAXIMIZING
CUSTOMER
LIFETIME VALUE
Linking CLV to Shareholder
Value
Product
Returns
Future of
Customer
Management
Cross - Buy
Source: Kumar, V., “Managing Customers for Profits”, The Wharton School Publishing
Optimal
Allocation of
ResourcesManaging
Multi-channel
Shoppers
Interaction
Orientation
drvkumar.com© Dr. V Kumar
Managing Loyalty And Profitability Simultaneously
Segmentation based
approach to guide marketing
actions to loyal
customers/distributors.
Efficient allocation of resources
resulting in increase in
profitability.
50
Objective: Marketing Metric Outcome:
drvkumar.com© Dr. V Kumar
“55% of consumers who rate their primary
financial institution high on service quality report
they are „very likely‟ to consolidate their business
with one provider.”
- Study by Peppers & Rogers Group,
Financial Service Marketing Magazine
& Roper Starch, 2002.
51
drvkumar.com© Dr. V Kumar
Evolving Dominant Logic in Customer Loyalty
52
1 RFM = Recency, Frequency & Monetary Value; PCV = Past Customer Value; SOW = Share of Wallet.
9
8
7
6
5
4
3
2
1
No
ExtensiveMinimalTechnology & Analytics
usage
CLVRFM, PCV, SOW1Metrics used
Link loyalty to profitability,
Influence behavioral loyalty,
Cultivate attitudinal loyalty
Build market share, increase
revenues, Build behavioral loyalty
through repeat purchase or usage
Program Objective
Tangible + ExperientialTangibleReward Type
Reactive + ProactiveReactiveReward Mechanism
Multiple (usually made possible through
partners and alliances)
Minimal Reward Options
Personalized and relevant, aimed at influencing
specific behavioral change or attitudinal
gratification
Standard and uniform aimed at
repeat purchase
Rewarding Scheme
Customized, based on type of usage or type of
spend
Standardized, based on usage or
spend
Program Type
Customer levelAggregate levelOperationalization Level
Evolving Loyalty Programs:
Customer Centric
Earlier Loyalty Programs:
Program Centric
Dimension
drvkumar.com© Dr. V Kumar
Managing Loyalty & ProfitabilityRethink Customer Segmentation
53
BUTTERFLIES
•Good fit of company offering and
customer needs
•High profit potential
•Action:–Aim to achieve transactional satisfaction, not
attitudinal loyalty
–Milk the accounts as long as they are active
–Key challenge: cease investment once
inflection point is reached
STRANGERS
•Little fit of company offering and customer
needs
•Lowest profit potential
•Action: –No relationship investment
–Profitize every transaction
BARNACLES
•Limited fit of company offering and
customer needs
•Low profit potential
•Action: –Measure size and share-of-wallet
–If share-of-wallet is low, specific up and cross-
selling
–If size of wallet is small, strict cost control
TRUE FRIENDS
•Good fit of company offering and
customer needs
•Highest profit potential
•Actions: –Consistent intermittently spaced
communication
–Achieve attitudinal and behavioural loyalty
–Delight to nurture/defend/retain
High
Profitability
Low
Profitability
Short-term
Customers
Long-term
Customers
drvkumar.com© Dr. V Kumar 54
The Wheel of Fortune Strategies Used for Maximizing CLV
Acquiring
Profitable
Customers
Customer
Selection
Preventing
Attrition of
Customers
Referral
Marketing
Strategy
Linking Investments in
Branding to
Customer Profitability
Pitching the Right Product,to the
Right Customer, at
the Right Time
Managing Loyalty and
Profitability
Simultaneously
MEASURING& MAXIMIZING
CUSTOMER
LIFETIME VALUE
Linking CLV to Shareholder
Value
Product
Returns
Future of
Customer
Management
Cross - Buy
Source: Kumar, V., “Managing Customers for Profits”, The Wharton School Publishing
Optimal
Allocation of
ResourcesManaging
Multi-channel
Shoppers
Interaction
Orientation
drvkumar.com© Dr. V Kumar 55
How To Optimize Resource Allocation And Maximize CLV?
drvkumar.com© Dr. V Kumar
Optimal Resource Allocation
How many contacts to be
made to which
customer/distributor and
through what channel?
Increase in ROI due to
decrease in Marketing cost
56
Objective: Marketing Metric Outcome:
drvkumar.com© Dr. V Kumar
Designing the Optimal Marketing Spend Across Channels
Comprises of the following 3 models:
57
Inter-purchase time modelAt what time intervals is a customer likely to spend?
Customer/Distributor Cash Flow ModelWhat is the future pattern of the cash flows for each customer/distributor,
based on the purchases likely to be made and the marketing costs likely to be incurred ?
Optimal Resource Allocation ModelHow do we apportion limited marketing resources across all the customers/distributors?
drvkumar.com© Dr. V Kumar
What are the Drivers?
Inter-purchase Time Model
• Customer Characteristics
– Switching Costs
– Involvement
– Past Behavior
• Supplier specific Factors
– Level of rich modes
– Level of standardized modes
– Inter-contact time
Cash Flow Model
• Customer Characteristics
– Lagged Contribution Margin
– Establishment Size
– Industry Category
– Total Quantity of Purchases
• Supplier specific Factors
– Total Marketing Communication
58
drvkumar.com© Dr. V Kumar
Optimal Resource Allocation
59
• CLV provides decision support system for allocating the ‗right‘ level of
resources to the ‗right‘ customer so as to maximize profits.
• Example: For a customer exhibiting low share of wallet and low lifetime
value, the resources deployed across various communication channels
would be modified as:– Reduce spending limit to $433 (from $819)
– Decrease frequency of face to face meetings to once every 12.5 months (from 4.5 months)
– Decrease frequency of direct marketing to once every 12.6 days (from 9.7 days)
This will lead to a substantial increase in profits to $12030 from $7435
per customer exhibiting similar characteristics
drvkumar.com© Dr. V Kumar
Database and Analysis Variables
60
Database
B2B high technology customer database
Customer Purchase History
Purchase Timing
Marketing Strategy variables
Premium Service Level
Number of contacts in,
– Face-to-Face Channel
– TeleSales channel
– Direct Mail Channel
– TeleWeb Channel
Number of customer initiated contacts
Customer size and sales
Cross-Buying
Number of Returns
• Analysis Sample
Cohort of Customers who made their first purchase in the first quarter of 1997
Effective Sample Size: N=216
drvkumar.com© Dr. V Kumar
Process Algorithm-Overall Analysis
61
PHASE III
PHASE II
PHASE I
Develop a customer
spending model based on
past revenues and
firmographics
Explain variation in
activity using buyer and
seller time varying
covariates
Inter-purchase time model
Predict the activity status for the next business
cycle using a generalized gamma model and
computing expected time until next purchase
Customer Cash Flow model
Compute Net Present Value of Future Profits
based on predictions from expected cash flows
Implement Optimal Strategy
Optimal Resource Allocation model
Generate optimal cost allocation rules based on
the profit function using Genetic Algorithms
Customer ProfileCustomer Purchase
History
Customer Contact
History
drvkumar.com© Dr. V Kumar 62
Bought in
the next 12
months
Did not Buy in
the next 12
months
Total
Expected to buy
in next 12 months85% 15 % n=128 (100%)
Not expected to
buy in the next 12
months
12 % 87 % n= 88 (100%)
Results from Phase I -Hit Rate of the Inter-purchase time model
drvkumar.com© Dr. V Kumar 63
Results from phase II -The Customer Cash Flow Model
Mean Standard
Deviation
Minimum Maximum
Observed 50,199 23,850 -171 1,810,881
Predicted 57,729 24,538 -178 1,897,257
a All reported values are in $ and are rounded off to the nearest integer
Comparison of descriptive statistics between observed cash flow and
Predicted cash flow in the holdout samplea
drvkumar.com© Dr. V Kumar
Combining Phase II and Marketing Costs
• Phase II provides the revenue generated for each individual
customer
• Combining this with the marketing costs provides the profit
computation
• The Profits thus computed are rank ordered by Deciles
• A lift of 7 times is obtained for the first Decile
64
drvkumar.com© Dr. V Kumar
Profit Function
65
1 2 3 4 5 6 7 8 9 10
Decile
0
100
200
300
400
500
(in
th
ou
sa
nd
s)
Ave
rag
e N
et P
rese
nt V
alu
e o
f P
rofit
s
Mean Profit = $55,229
drvkumar.com© Dr. V Kumar
Improvement Over the Currently usedShare of Wallet Approach
66
• Share of Wallet is defined as the ratio of the total customer spending with the firm to the
total category spending(the firm plus its competitors) for that customer. Share of Wallet can
be considered a measure of loyalty
• The Share of Wallet approach yields the following classification of customers and their
average profits
• However a cross analysis of Share of Wallet and Customer Value indicates that a superior
approach can be adopted by identifying more responsive and profitable customers who may
have escaped attention when only the Share of Wallet approach is followed
$19,490
(n=113)
$94,437
(n= 103)
Share of Wallet
Low High
Average Profit
drvkumar.com© Dr. V Kumar
Customer Value versus Share of Wallet: Distribution of Customers
67
Low SOW High SOW
High Value
Cell II – BUTTERFLIES
Avg Profit = $ 35,317 (N= 61)
Cell IV - TRUE FRIENDS
Avg Profit = $ 201,695 (N= 48)
Low Value
Cell I - STRANGERS
Average Profit = $ 925 (N= 52 )
Cell III - BARNACLES
Average Profit = $ 830 (N= 55 )
Customer
Value
Share of WalletThe improved method results
in the following matrix
•Customers are assigned to the low share of wallet segment if their score for share of wallet is less than 0.5.
•The rest of the customers are assigned to the high share of wallet segment.
•Customer Value is calculated using the NPV function.
•Customers are assigned to the low lifetime value segment if their net present value of expected profits
is less than the median value of expected profits from all the customers in the sample.
•The median expected profits in our sample is $1,156.00.
drvkumar.com© Dr. V Kumar
Extent of the Improvement
• We note that the improved method allows us to identify customers
who have a low share of wallet with the firm, but provide high value
(with an average profit of $35,317 versus an average profit of
$19,490 when only share of wallet is used to target customers)
• These customers would have been totally ignored if the uni-
dimensional approach based on share of wallet had been followed
68
drvkumar.com© Dr. V Kumar
Actionable Strategy
69
Customer Value
High
Low
IVII
IIII
Low HighShare of Wallet
Hence this approach is able to identify customers in the low share of wallet
segment who merit marketing investments due to their higher profitability
Marketing action should aim to move customers from Cell III to Cell IV
and from Cell II to Cell IV as a means to improve ongoing profitability.
Cost minimization strategy should be employed to transact with customers
in Cell I
STRANGERS BARNACLES
BUTTERFLIESTRUE FRIENDS
drvkumar.com© Dr. V Kumar
Average Number of Touches/Establishment(B to B Firm)
70
FO
CU
S
PR
IMA
RY
AC
TIV
AT
ION
AC
QU
ISIT
ION
HIG
H
EM
ER
GIN
G
CO
MP
ET
ITIV
E
0
50
100
150
Avg
. T
ou
ch
es
1H03 ACTUAL RECOMMENDEDF
OC
US
PR
IMA
RY
AC
TIV
AT
ION
AC
QU
ISIT
ION
EM
ER
GIN
G
CO
MP
ET
ITIV
E
0
10
20
30
40
50
60
70
Avg T
ouches
Extend reach by moving touches away from Focus and Primary, and toward Activation and Acquisition
Average touch based on total touches to establishments receiving at least 1 proactive touch
Excludes very small business (1-99 enterprise employee size)
drvkumar.com© Dr. V Kumar
Optimal Resource Allocation Matrix
71
Cost Reduction($):
Currently Spending $960
Optimal Spending Limit $2220
Face to Face Meetings:
Currently meets once every 6.6 months
Optimal meeting interval is 2.2 months
Direct Mail/Telesales:
Current Interval is 4.82 days
Optimal Interval is 1.9 days
Profits:
Current Profit is $109,364
Optimal profit is $178,092
Cost Reduction($):
Currently Spending $2340
Optimal Spending Limit $2690
Face to Face Meetings:
Currently meets once every 2.5 months
Optimal meeting interval is 1.5 months
Direct Mail/Telesales:
Current Interval is 6.3 days
Optimal Interval is 2.3 days
Profits:
Current Profit is $534,888
Optimal profit is $905,224
Cost Reduction($):
Currently Spending $720
Optimal Spending Limit $570
Face to Face Meetings:
Currently meets once every 4.5 months
Optimal meeting interval is 12.5 months
Direct Mail/Telesales:
Current Interval is 9.7 days
Optimal Interval is 12.6 days
Profits:
Current Profit is $7435
Optimal profit is $12,030
Cost Reduction($):
Currently Spending $1770
Optimal Spending Limit $1470
Face to Face Meetings:
Currently meets once every 2.4 months
Optimal meeting interval is 10 months
Direct Mail/Telesales:
Current Interval is 8.4 days
Optimal Interval is 8.3 days
Profits:
Current Profit is $10,913
Optimal profit is $28,354
High
Low
Low
HighShare of Wallet
Customer
Lifetime
Value
drvkumar.com© Dr. V Kumar
Managerial Applications of the Framework
• The proposed algorithm enables the firm to take specific marketing actions
for a larger set of customers in the customer list, for a given budget
• The firm can use the framework to assess the return on its marketing
investments and hence bring accountability to marketing actions
• The algorithm is used to provide early indications to the firm about a
customer‘s transition through various phases in his/her life-cycle
(exploration, evaluation, maturity, and decline)
72
drvkumar.com© Dr. V Kumar
Harrah‟s
73
“Technology that helps
Harrah’s keep track of
customer spending and
behavior will continue
to be key to Harrah’s
growth … This thing is the
heart and soul of
Harrah’s.”
Gary Loveman, CEO
Harrah’sSource: ―New Harrah‘s CEO plans Growth Strategy,‖ Peter Henderson, Reuters Update, January 9, 2003
drvkumar.com© Dr. V Kumar
Harrah‟s Business Strategy
74
Established a Brand Identity Based on….
Knowing their
customers wellRewarding customer
loyalty
Giving customers
great service
Source: Academic Case Study. ―Harrah‘s High Payoff from Customer Information, Hugh J. Watson and Linda Volonino 11/01
drvkumar.com© Dr. V Kumar
How well do they know their customers?
75
“Harrah’s has 19 million loyalty
club members worldwide,
and it can identify an individual
by name, spending
habits, and interactions with the
casino-hotel’s restaurants,
web site, stores and
entertainment facilities.”
Source: ―Upping the Profit Ante,‖ Reid A. Paul, HT Magazine, April 2001.
drvkumar.com© Dr. V Kumar
It‟s in the cards!
76
Minimum annual
value of cardholder tier
Diamond: $5,000
Platinum: $ 1,500
Gold: 0
Examples of card-holder data
collected• Basic information such as age, sex, home
address
• Betting patterns; casino-game preferences;
length of play
• Frequency of visits
Growth of Harrah’s customer databaseMillions of customers
1995 1996 1997 1998 1999 2000 2001 2002
5.3
9.9
11.9
13.8
19.0
23.025.0
26.6
Share of Harrah’s 2001 cardholder revenue
by change in cardholders average value’
from 2000 to 2001,percent
40
20
14
16 10
Increased
Value > 3x
Increased
Value < 1x
Increased
Value 1-2x
Did not change or
Decreased value
Increased Value
2-3x
drvkumar.com© Dr. V Kumar
BCP Telecommunicacoes
77
Leading wireless service provider operating
in two regions of Brazil;
1.9 million subscribers @ $700 million / Yr Rev
How BCP Uses DW and Customer Analytics:
• Segmenting customers for more effective campaigns
•Understanding each campaign’s effectiveness
• Knowing and acting on behavioral differences and
preferences across customer base
• Reducing acquisition costs by channel
• Ensuring relevant and timely communications
to each customer
• Knowing what offers a customer will most prefer
• Predicting & pro-actively intervening to prevent churn
drvkumar.com© Dr. V Kumar
LTV Calculation in Class
78
drvkumar.com© Dr. V Kumar
Let‟s go back to the question:
Let‘s look at this exercise!!
79
Dollar value of
the first
purchase
Customer
Lifetime Value
drvkumar.com© Dr. V Kumar
Assignment 5: Lifetime Value of Customers
• Mail order catalog firm
80
Total Customers
7953
Initial purchase<$50
4657
5 years observations
5 years observationsInitial purchase>=$50
3296
Manager: Should I mail catalog to all customers?
drvkumar.com© Dr. V Kumar
Lifetime Value of Customers (Contd.)
81
LTV Pt (Qt t )
d tt 0
n
(Dt Rt )
d tt1
n
A
Pt = the probability of purchase in period t
Qt = the quantity purchased in period t
= the margin on purchases in period t
dt = the discount rate , where d= ( 1+ ( interest rate x risk factor))
Dt = costs to develop the relationship in period t
Rt = cost to retain the customer in period t
A = initial acquisition cost
n = the number of periods
tInclude initial
purchase and
acquisition cost
drvkumar.com© Dr. V Kumar
Decomposition of the Computation
82
LTV Pt (Qt t )
d tt 0
n
(Dt Rt )
d tt1
n
A
•New Customer Acquisition costs
•Average cost to obtain prospect name = $0.10
•Average cost to send initial catalog = $0.75
•Average response rate = 2.3 %
•Number of catalogs
mailed annually to each
acquired customer = 5
•Cost of mailing a catalog
to a customer = $0.75
Tables 1,2,3
Average Margin
on orders =42%
•5. Annual interest rate = 20%
•6. Risk Factor = 1
•d= ( 1+ ( interest rate x risk factor))
Past data: Prob.=1
drvkumar.com© Dr. V Kumar 83
Lifetime Value of Customers -IllustrationTable 1: <$50 Initial
Purchase Cohort
# Orders 5
0 3837
1 626
2 141
3 38
4 10
5 3
6
7
8 2
9
10
11
12
13
Table 3: Initial and Repeat Order $ Amounts by Year and Group
$ amount of $ amount of
Group initial order repeat order-
Year 5
Initial order N 4657.00 820.00
<$50 Minimum -35.00 4.38
Maximum 50.00 484.91
Mean 31.84 53.63
Std. deviation 9.97 39.31
× =0
=626
=282
=114
=40
=15
=0
=0
=16
=0
=0
=0
=0
=0
×
# of
orders ××1093 $ / order53.63 42% = 24,619,39________
(1+0.2)5 =9,887.31
t
ttt
d
QP )( for t=5
________
4657=2.12
drvkumar.com© Dr. V Kumar 84
Lifetime Value of Customers –Illustration (Contd.)
5
1
)(
tt
tt
d
RD11.2148
)2.01(
)75.0*5(5
1
t
t A 36.956524657*%3.2
)75.01.0(*4657
Initial purchase<$50
46571 2 3 4 5
A
Pt (Qt t )
d tt 0
5
(Dt Rt )
d tt1
5
LTV
drvkumar.com© Dr. V Kumar
Lifetime Value of Customers (Contd.)
85
LTV Pt (Qt t )
d tt1
n
(Dt Rt )
d tt1
n
Pt = the probability of purchase in period t
Qt = the quantity purchased in period t
= the margin on purchases in period t
dt = the discount rate , where d= ( 1+ ( interest rate x risk factor))
Dt = costs to develop the relationship in period t
Rt = cost to retain the customer in period t
A = initial acquisition cost
n = the number of periods
t Does not include
initial purchase and
acquisition cost
drvkumar.com© Dr. V Kumar
Questions
• What is the average lifetime value of a customer in each of the two groups
(Group1- # of Customers with initial purchase < $50, Group2- # of
Customers with initial purchase >=$50 )? Please calculate for case 1:
including initial purchase and acquisition cost; and case 2: excluding initial
purchase and acquisition cost. Which do you think is the right number for
CLV? Is there a difference in your conclusion?
• Is the decision to mail all catalogs to all customers justified in the light of the
above analysis?
• What other methods of grouping these customers can be considered that
will help us differentiate customers based on their value?
• What can we predict in terms of behavior in the coming year? What
additional analysis would we need?
86
drvkumar.com© Dr. V Kumar
Case Assignment 1: Exotic Apparel
Background:
– Exotic Apparel: Successful retail chain selling high-end women‘s wear and
premium foot wear.
– Revenue: $152 million in 2003; five –year annual revenue growth rate: 6.5%
Question of the Managing director:
87
Past Purchase BehaviorFuture profitability of a
customer
drvkumar.com© Dr. V Kumar
Case Assignment 1: Exotic Apparel
88
Research Areas:
• Determine the Customer Lifetime Value (CLV) of each customer.
• Analyze the profiles of customers segmented based on CLV.
• Determine the profitability of each store.
• Define and measure the extent of loyalty for each customer based on traditional
metrics (Regularity of purchase, Frequency of purchase, RFM score and Tenure of
the customer).
• Determine if customers selected on the basis of traditional metrics are the right
choice from the view of long-term profitability. Check to see if CLV is a better
indicator of long-term profitability than the traditional metrics.
• Determine what factors drive profitability and compare them with the factors that drive
loyalty.
drvkumar.com© Dr. V Kumar
Case Assignment 1: Exotic Apparel
89
Request Data
Calculate Metrics
• Frequency of purchases made by every customer based on each customer’s purchase history.
• Total revenue contributed by each customer through each channel of purchase until the current date.
• Revenue contributed by each customer by product category.
• Average dollar spent per transaction by each customer.。。。。
• Unique number that identified a customer in the database across all the transaction that he/she makes.
• Demographics describing each customer (Age, gender, occupation, income, marital status, number of children, location, distance of residence from a retail store, etc.)
• The shopping behavior exhibited by customers 。。。。
•Lifetime revenue accrued from web sales and factory outlets
•Tenure of the customer
•Dollars spent on high end women’s apparel
•Degree of cross buying
•Dollars spent on Home Furnishings
•Returns made during the lifetime。。。。。。
•Target high spenders in web sales and factory outlets
•Deploy marketing tactics to encourage cross-buying
•Identify customers exhibiting high incidence of net returns and devise strategies to discourage that
behavior
•Encourage customers who spend in high end women‘s apparel
•Target customers exhibiting high duration.
Research
Team
Identify Drivers of CLV
drvkumar.com© Dr. V Kumar
Case Assignment 1: Exotic Apparel
90
Questions:
• By focusing only on profitable customers based on CLV, is the company doing
the right thing from a long-term business point of view? What are the other
decisions that the company can take based on CLV?
• Comment on the strategy recommended by the research team. Are there any
strategies that you would add or remove from this list?
• How different would the recommended strategy be if traditional loyalty metrics
were used instead of CLV?
• It seems intuitive that the first five variables listed have a positive correlation
with CLV. Why do you think that the amount of product returns has an inverted
U-shaped relationship with CLV?
drvkumar.com© Dr. V Kumar
Case Assignment 1: Exotic Apparel (Contd.)
How to Answer?
• This assignment is a comprehensive project that covers the materials
taught in the whole course. After each class, it is suggested to go
back to this case to see how the new concepts and methods can be
applied to solve the questions in this case study.
– Example, RFM will be covered on Day 2 and CLV concept will be
covered on Day 3, and CLV-based management on Day 3 and Day
4, etc.
• The questions require the students to understand the metrics that the
“research team” used to solve the problem, and how they get the
strategic recommendations from those metrics. Most questions
should be answered in essay format, but try to make your point as
clear as possible.
91