CRM Chapter 6

Chapter 6:
Customer Analytics Part II

Overview
Topics discussed:
§ Strategic customer-based value metrics
§ Popular customer selection strategies
§ Techniques to evaluate alternative customer selection strategies
V. Kumar and W. Reinartz – Customer Relationship Management: Concept, Strategy, and Tools, 3rd Edition 2
Customer-based Value Metrics
Customer
based
Value
Metrics
Popular Strategic
Size Share of Past

Share of Transition LTV Customer
Of Category RFM Customer
Wallet Matrix Metrics Equity
Wallet Reqt. Value
Strategic Customer-based Value Metrics
§ RFM Method
§ Past Customer Value
§ Lifetime Value Metrics
§ Customer Equity
RFM Method
• Elapsed time since a

customer last placed an
Recency order with the company
• Number of times a
customer orders from
Frequency the company in a RFM Value
certain defined period
• Amount that a customer

Monetary spends on an average
value transaction
§ Technique to evaluate customer behavior and customer value

§ Often used in practice
§ Tracks customer behavior over time in a state-space
RFM Method
§ Example
§ Customer base: 400,000 customers
§ Sample size: 40,000 customers
§ Firm’s marketing mailer campaign: $150 discount coupon
§ Response rate: 808 customers (2.02%)
RFM Method
§ Recency coding
§ Test group of 40,000 customers is sorted in descending order based on the criterion of
‘most recent purchase date’
§ The earliest purchasers are listed on the top and the oldest are listed at the bottom
§ The sorted data is divided into five equally sized groups (20% in each group)
§ The top-most group is assigned a recency code of 1, the next group
a code of 2 until the bottom-most group is assigned a code of 5
§ Analysis of customer response data shows that the mailer campaign got the highest
response from those customers grouped in recency code 1 followed by those in code 2
etc.
RFM Method: Response and Recency
5.00%
4.50%
4.00%
Customer Response %
3.00% 2.80%
2.00%
1.50%
1.05%
1.00%
0.25%
0.00%
1 2 3 4 5
Recency Code (1 - 5)
§ Graph depicts the distribution of relative frequencies of customer groups assigned to

recency codes 1 to 5
§ Highest response rate (4.5%) for the campaign was from customers in the test group who
belonged to the highest recency quintile (recency code =1)
RFM Method: Response and Frequency
3.00%
2.45%
2.50% 2.22%
2.08%
2.00% 1.67% 1.68%

Response Rate %
1.50%
1.00%
0.50%
0.00%
1 2 3 4 5
Frequency Code 1-5
§ Graph depicts the response rate of each of the frequency-based sorted quintiles
§ The highest response rate (2.45%) for the campaign was from customers in the test group
belonging to the highest frequency quintile (frequency code =1)
RFM Method: Response and Monetary Value
2.35%
2.50%
2.05%
1.95% 1.90% 1.85%
2.00%
1.50%
Response Rate %
1.00%
0.50%
0.00%
1 2 3 4 5
Monetary Value Code (1-5)
§ Customer data is sorted, grouped and coded with a value from 1-5
§ The highest response rate (2.35%) for the campaign was from those customers in the test
group who belonged to the highest monetary value quintile (monetary value code =1)
RFM Method: RFM procedure
Last purchase: Group 1 ß R=1

1 day ago
average
ß R=1 response rate
ß R=1
Group 2 ß R=2
average
Step 1 Step 2
Step 3
ß R=2
...
...
...
Group 3 ß R=5
average
Last purchase: ß R=5

320 days ago
RFM Method: Limitations
§ RFM method 1 independently links customer response data with R, F and M values and
then groups customers belonging to specific RFM codes
§ May not produce equal number of customers under each RFM cell since individual metrics
R, F, and M are likely to be somewhat correlated
§ For example, a person spending above average (high M) is also likely to spend more
frequently (high F)
§ For practical purposes, it is desirable to have exactly the same number of individuals in
each RFM cell
RFM Method: Cell Sorting Technique
§ A list of 40,000 test group of customers is first sorted for recency and then grouped into 5
groups of 8,000 customers each
§ The 8,000 customers in each group are sorted based on frequency and divided into five
equal groups of 1,600 customers each - at the end of this stage, there will be RF codes
starting from 11 to 55 with each group including 1,600 customers
§ In the last stage, each of the RF groups is further sorted based on monetary value and
divided into five equal groups of 320 customers each
§ RFM codes starting from 111 to 555 each including 320 customers
§ Considering each RFM code as a cell, there will be 125 cells (5 recency divisions * 5
frequency divisions * 5 monetary value divisions = 125 RFM Codes)
RFM Method: RFM Cell Sorting
M
F
11 131
12 132
R
133
1 13
14 134
2 15 135
41
3 441
42
442
43
4 443
44
444
5 45
445
Customer Sorted Once Sorted Five Sorted Twenty-Five

Database Times per R Times per R Quintile
Quintile
RFM Method: Breakeven Value (BE)
§ Breakeven = net profit from a marketing promotion equals the cost associated with
conducting the promotion
§ Breakeven Value (BE) = unit cost price / unit net profit
§ BE computes the minimum response rates required in order to offset the promotional costs
involved and thereby not incur any losses
§ Example
§ Mailing $150 discount coupons
§ The cost per mailing piece is $1.00
§ The net profit (after all costs) per used coupon is $45,
à Breakeven Value (BE) = $1.00/$45 = 0.0222 or 2.22%
RFM Method: Breakeven Index
§ Breakeven Index (BEI) = ((Actual response rate – BE) / BE)*100
§ Example
§ If the actual response rate of a particular RFM cell was 3.5%
§ BE is 2.22%,
à The BEI = ((3.5% - 2.22%)/2.22%) * 100 = 57.66
§ Positive BEI value à some profit was made from the group of customers
§ 0 BEI value à the transactions just broke even
§ Negative BEI value à the transactions resulted in a loss
RFM Method: Combining RFM codes, breakeven
codes, breakeven index
Cell # RFM codes Cost per mail $ Net profit per saleBreakeven (%) Actual response Breakeven index
($) (%)
1 111 1 45.00 2.22 17.55 690
2 112 1 45.00 2.22 17.45 685
3 113 1 45.00 2.22 17.35 681
4 114 1 45.00 2.22 17.25 676
5 115 1 45.00 2.22 17.15 672
6 121 1 45.00 2.22 17.05 667
7 122 1 45.00 2.22 16.95 663
8 123 1 45.00 2.22 16.85 658
9 124 1 45.00 2.22 16.75 654
10 125 1 45.00 2.22 16.65 649
11 131 1 45.00 2.22 16.55 645
12 132 1 45.00 2.22 16.45 640
13 133 1 45.00 2.22 16.35 636
14 134 1 45.00 2.22 16.25 631
15 135 1 45.00 2.22 16.15 627
16 141 1 45.00 2.22 16.05 622
17 142 1 45.00 2.22 15.95 618
18 143 1 45.00 2.22 15.85 613
19 144 1 45.00 2.22 15.75 609
20 145 1 45.00 2.22 15.65 604
21 151 1 45.00 2.22 15.55 600
22 152 1 45.00 2.22 15.45 595
23 153 1 45.00 2.22 15.35 591
24 154 1 45.00 2.22 15.25 586
25 155 1 45.00 2.22 15.15 582
26 211 1 45.00 2.22 15.65 604
27 212 1 45.00 2.22 15.55 600
28 213 1 45.00 2.22 15.45 595
29 214 1 45.00 2.22 15.35 591
30 215 1 45.00 2.22 15.25 586
31 221 1 45.00 2.22 15.15 582
32 222 1 45.00 2.22 15.05 577
33 223 1 45.00 2.22 14.95 573
34 224 1 45.00 2.22 14.85 568
35 225 1 45.00 2.22 14.75 564
RFM codes versus BEI
1400
1200
1000
800
600 Break-Even Index

BEI
400
200
0
111
132
153
224
245
321
342
413
434
455
531
552
RFM Cell Codes
RFM and BEI
RFM BEI
§ Customers with higher RFM values tend to have higher BEI values
§ Customers with a lower recency value but relatively high F and M values tend to have
positive BEI values
§ Customer response rate drops more rapidly for the recency metric
§ Customer response rate for the frequency metric drops more rapidly than the one for the
monetary value metric
Order of Importance of R, F, and M
§ Technique most often applied in the order of recency, frequency, and monetary value
§ However, the order varies for different industry segments.
§ More accurate order of coding would depend on the rapidity the customer response rate
drops
§ The metric (R, F, or M) for which the response rates decline the quickest is likely the best
predictor of future customer response
§ Hence, should be coded first
Comparison of profits for targeting campaign test
Test Full customer base RFM Selection

Average response rate 2.02% 2.02% 15.25%
# of responses 808 8,080 2,732.8
Average Net profit/Sale $45 $45 $45
Net Revenue $36,360 $363,600 $122,976
# of Mailers sent 40,000 400,000 17,920
Cost per mailer $1.00 $1.00 $1.00
Mailing cost $40,000.00 $400,000.00 $17,920.00
Profits (-$3,640.00) ($36,400.00) $105,056.00
Relative Importance of R, F, and M
§ Regression techniques to compute the relative weights of the R, F, and M metrics

§ Relative weights are used to compute the cumulative points of each customer
§ The pre-computed weights for R, F and M, based on a test sample are used to assign RFM
scores to each customer
§ The higher the computed score, the more likely the customer will be profitable in future
§ This method is flexible and can be tailored to each business situation
Recency Score
§ 20 if within past 2 months, 10 if within past 4 months, 05 if within past 6 months, 03 if within
past 9 months, 01 if within past 12 months, relative weight = 5
Customer Purchase Recency Assigned Weighted

Number (Month) Points Points
1 2 20 100
John 2 4 10 50
3 9 3 15
Smith 1 6 5 25
1 2 20 100
Mags 2 4 10 50
3 6 5 25
4 9 3 15
Frequency Score
§ Points for Frequency: 3 points for each purchase within 12 months; Maximum = 15 points;
Relative weight = 2
Customer Purchase Frequency Assigned Weighted

Number Points Points
1 1 3 6
John 2 1 3 6
3 1 3 6
Smith 1 2 6 12
1 1 3 6
Mags 2 1 3 6
3 2 6 12
4 1 3 6
Monetary Value Score
§ Monetary Value: 10 percent of the $-value of purchase with 12 months; Maximum = 25

points; Relative weight = 3
Customer Purchase Value of Assigned Weighted

Number purchase ($) Points Points
1 $40 4 12
John 2 $120 12 36
3 $60 6 18
Smith 1 $400 25 75
1 $90 9 27
Mags 2 $70 7 21
3 $80 8 24
4 $40 4 12
RFM Cumulative Score
§ Cumulative scores: 249 for John, 112 for Smith and 308 for Mags, indicates a potential
preference for Mags
§ John seems to be a good prospect, but mailing to Smith might be a misdirected marketing
effort
Customer Purchase Total Weighted Cumulative
Number Points Points
1 118 118
John 2 92 210
3 39 249
Smith 1 112 112
1 133 133
Mags 2 77 210
3 61 271
4 37 308
Past Customer Value
§ Computation of Customer Profitability (PCV)

T
§ PCV of customer i = å GCi (t0 -t )n * (1 + d ) t
t =0
§ Where: i = number representing the customer,
t = time index,
δ = applicable discount rate (for example 1.25% per month),
t0 = current time period,
T = number of time periods prior to current period that should be considered,
GCin = gross contribution of transaction of customer in period t
§ Since products / services are bought at different points in time during the customer’s lifetime,
all transactions have to be adjusted for the time value of money
§ Limitations
§ Equation does not consider whether a customer is going to be active in the future and it
does not incorporate the expected cost of maintaining the customer in the future
Spending Pattern of a Customer
Jan Feb March April May
Purchase Amount ($) 800 50 50 30 20
GC 240 15 15 9 6
Gross contribution (GC) = purchase amount X contribution margin
§ PCVi = 6*(1+0.0125)o +9*(1+0.0125)1 +15*(1+0.0125)2 +15*(1+0.0125)3+ 240*(1+0.0125)4

= 302.01486
§ The customer is worth $302.01 expressed in net present value in May dollars
§ Comparing the PCV of a set of customers leads to a prioritization of directing future

marketing efforts
Lifetime Value Metrics
§ Multi-period evaluation of a customer’s value to the firm

§ Lifetime Value (LTV)
Recurring
Revenues
Contribution
margin
Recurring
costs
Lifetime of a
customer
customer Lifetime Profit
LTV
Discount Acquisition
rate cost
Basic LTV Model
T t
! 1 $
§ LTVi = ∑ GCit # &
t=1
" 1+ δ %
§ Where: i = customer,
t = time period,
d = interest (or discount) rate,
GCit = gross contribution of customer i at time t,
T = observation time horizon,
LTVi = lifetime value of an individual customer i at net present value time t=0
Basic LTV Model
§ LTV is a measure of a single customer’s worth to the firm
§ Used for pedagogical and conceptual purposes
§ Information source
§ CM and T from managerial judgment or from actual purchase data.
§ The interest rate, a function of a firm’s cost of capital, can be obtained from financial
accounting
§ Evaluation
§ Typically based on past customer behavior and may have limited diagnostic value for
future decision-making Caution:
If the time unit is different from a yearly
basis, the interest rate d needs to be
adjusted accordingly.
LTV with Splitted Revenues and Costs
T t
§ # 1 &
LTVi = ∑ ( Sit − DCit ) − MCit ) ⋅ % (
t=1
$ 1+ δ '
t = time period,
Sit = sales value to customer i at time t,
DCit = direct costs of products by customer i at time t,
MCit = marketing costs directed at customer i at time t,

LTVi = lifetime value of an individual customer i at net present value time
t=0
LTV with Splitted Revenues and Costs
§ The cost element of this example is broken down into direct product-related costs and
marketing costs
§ Depending on data availability, it can be enhanced by including service-related cost,

delivery cost, or other relevant cost elements
LTV Including Customer Retention Probabilities
t
§ " K
T % " 1 %
LTVi = ((∑$ ∏ Rrk ' GCit $ ' )) − ACi
t=1 # k=1 & # 1+ δ &
t = time period,
Rrt = average retention rate at time t (it is possible to use an individual

level retention probability Rrit but usually this is difficult to obtain),
ACi = costs of acquiring customer i (acquisition costs)
LTV Including Customer Retention Probabilities
§ In this equation the term ∏ Rr is actually the survival rate SRt

k
k=1
§ The retention rate is constant over time and thus the expression can be simplified using the
identity:
t
∏ Rr k = (Rr)t
k=1
LTV with Constant Retention Rate and Gross Contribution
§ Assuming that T à ¥ and that the retention rate and the contribution margin (CM) do not
vary over time: " Rr %
LTVi = GCi $ ' − ACi
# 1− Rr + δ &
Rr
§ The margin multiplier: 1− Rr + δ
§ How long is the Lifetime Duration?

§ For all practical purposes, the lifetime duration is a longer-term duration used
managerially
§ It is important to make an educated judgment regarding a sensible duration horizon in
the context of making decisions
LTV with Constant Retention Rate and Gross Contribution
§ Incorporating externalities in the LTV

§ The value a customer provides to a firm does not only consist of the revenue stream
that results from purchases of goods and services
§ Product rating websites, weblogs and the passing on of personal opinions about a
product or brand co-contribute substantially to the lifetime value of a customer
§ These activities are subsumed under the term word-of mouth (WOM)
Measuring and Incorporating Word-of-Mouth (WOM)
t
" K
T % " 1 %
§ LTVi = ((∑$ ∏ Rrk ' (GCit + nit ACSt ) $ ' )) − ACi
t=1 # k=1 & # 1+ δ &
t = time period,

Rrt = average retention rate at time t,
nit = number of new acquisition at time t due to referrals customer i,

ACSt = average acquisition cost savings per customer gained through
referral of customer i at time t,
ACi = costs of acquiring customer i (acquisition costs)
Customer Value Matrix
Average CRV after 1 year
Low High
Average LTV High Affluents Champions

after 1 year
Low Misers Advocates
This table is adapted from Kumar, Petersen , and

Leone (2007)
LTV
§ Alternative ways to account for externalities

§ The value of a customer’s referrals can be separated from the LTV, for example by
calculating a separate customer referral value (CRV) for each customer
§ A joined evaluation of both metrics helps the management to select and determine
how to develop its customers
§ Information on sales, direct cost, and marketing cost come from internal company
records
§ Many firms install activity-based-costing (ABC) schemes to arrive at appropriate
allocations of customer and process-specific costs
§ Evaluation
§ LTV (or CLV) is a forward looking metric that is appropriate for long-term decision
making
§ It is a flexible measure that has to be adapted to the specific business context of an
industry
Customer Equity
§ Customer equity (CE) = Sum of the LTV of all the customers of a firm
I
CE = ∑ LTVi
i=1
I = all customers of a firm (or specified customer cohort or segment),
LTVi = lifetime value of customer i
§ Indicator of how much the firm is worth at a particular point in time as a result of the firm’s
customer management efforts
§ Can be seen as a link to the shareholder value of a firm
Customer Equity
K
§ Customer Equity Share (CES): CES j = CE j / ∑ CEk
K=1
§ Where: CEj = customer equity of brand j,

j = focal brand,
K = all brands a firm offers
§ Basically the same information as for the LTV is required
§ Evaluation
§ The CE represents the value of the customer base to a company
§ The metric can be seen as an indicator for the shareholder value of a firm
Customer Equity Calculation Example
1 2 3. 4. 5. 6. 7. 8. 9 10. 11.
Year from Sales per Manufa- Manufac- Mktg and Actual Survival Expected Profit per Discounted Total
Acquis-ition Customer cturer turer Servicing Retention Rate Number of Customer Profit per Disctd.
Margin Gross Costs Rate Active per period Customer Profits per
Margin Customer per Manu- per Period Period to
facturer to the
Manufactu- Manufactu
rer -rer
0 120 0.3 36 20 0.4 0.4 400 16 16 6,400

1 120 0.3 36 20 0.63 0.25 250 16 14 3,500
2 120 0.3 36 20 0.75 0.187 187 16 12 2,244
3 120 0.3 36 20 0.82 0.153 153 16 11 1,683
4 120 0.3 36 20 0.85 0.131 131 16 9 1,179
Total 15,006
customer
equity
Popular Customer Selection Strategies
§ Techniques
§ Profiling
§ Binary Classification Trees
§ Logistic regression
§ Techniques to Evaluate Customer Selection Strategies
§ Missclassification Rate
§ Lift Analysis
§ CRM at Work: The Kroger Co.

à Example where a company combines analytical skills, judicious judgment, knowledge
about consumer behavior and careful targeting of customers
Profiling
§ The intuitive approach for customer selection is based on the assumption that the most
profitable customers share common characteristics
§ Based on this assumption the company should try to target customers with similar profiles
to the currently most profitable ones
§ Disadvantage
§ Only customers that are similar to existing ones are considered
§ Profitable customer segments that do not match the current customer base might
be missed
Profiling: Using Profiling for New Customer Acquisition
Response Internal External

Variable Variables Variables
Transactions, demographics, lifestyle Demographics, lifestyle

Individual A
Individual B Current
Individual C Customers
A B C
Individual 1
Individual 2
Potential
Individual 3 Customers
F E D
Binary Classification Trees
§ Used to identify the best predictors of a 0/1 or binary dependent variable

§ Useful when there is a large set of potential predictors for a model
§ Classification tree algorithms can be used to iteratively search the data to find out which
predictor best separates the two categories of a binary (or more generally categorical)
target variable
Binary Classification Trees
§ The algorithm procedure

1. To find out which of the explanatory variables Xi best explains the outcome Y,
calculate the number of misclassifications
2. Use the variable Xi with the lowest misclassification rate to separate the customer
base
3. This process can be repeated for each sub segment until the misclassification rate
drops below a tolerable threshold or all of the predictors have been applied to the
model
§ Evaluation
§ Problem with the approach: prone to over-fitting
§ The model developed may not perform nearly as well on a new or separate dataset
Binary Classification Trees: Classification of Potential Hockey
Equipment Buyers
Male Female Total
Bought Did not Bought Did not Bought Did not

hockey buy hockey buy hockey buy
hockey hockey hockey
Bought 60 1,140 50 1,550 110 2,690

scuba
Did not 1,540 2,860 80 1,320 1,620 4,180
buy
scuba
Total 1,600 4,000 130 2,870 1,730 6,870
Equipment Buyers
§ Possible separations of potential hockey equipment buyers

Step 1
Separation by gender
1,600 bought hockey equipment
Male: 5,600
4,000 did not buy hockey equipment
130 bought hockey equipment

Female: 3,000
Separation by purchase of scuba

equipment 110 bought hockey equipment
Bought: 2,800
1,620 bought hockey equipment

Did not buy: 5,800
Equipment Buyers
§ Classification of hockey buyers by gender

Step 2
Buyer
Yes 1730
No 6870
Total 8600
Gender
Male Female
Buyer Buyer
Yes 1,600 Yes 130
No 4,000 No 2,870
Total 5,600 Total 3,000
Equipment Buyers
§ Classification tree for hockey equipment buyers

Step 3
Buyer
Yes 1730
No 6870
Total 8600
Gender
Male Female
Buyer Buyer
Yes 1,600 Yes 130
No 4,000 No 2,870
Total 5,600 Total 3,000
Bought scuba equipment Marital status
Bought scuba Not bought scuba Married Not married

Buyer Buyer Buyer Buyer
Yes 60 Yes 1,540 Yes 100 Yes 30
No 1,140 No 2,860 No 2,000 No 870
Total 1,200 Total 4,400 Total 2,100 Total 900
Logistic Regression
§ Method of choice when the dependent variable is binary and assumes only two discrete
values
§ By inputting values for the predictor variables for each new customer – the logistic model
will yield a predicted probability
§ Customers with high ‘predicted probabilities’ may be chosen to receive an offer since they
seem more likely to respond positively
Logistic Regression: Examples
§ Example 1: Home ownership

§ Home ownership as a function of income can be modeled whereby ownership is
delineated by a 1 and non-ownership a 0
§ The predicted value based on the model is interpreted as the probability that the
individual is a homeowner
§ With a positive correlation between increasing income and increasing probability of
ownership, one can expect results as
§ Predicted probability of ownership is .22 for a person with an income of $35,000
§ Predicted probability of .95 for a person with a $250,000 income
Logistic Regression: Examples
§ Example 2: Credit Card Offering

§ Dependent Variable - whether the customer signed up for a gold card offer
§ Predictor Variables - other bank services the customer used plus financial and
demographic customer information
§ By inputting values for the predictor variables for each new customer, the logistic model
will yield a predicted probability
§ Customers with high ‘predicted probabilities’ may be chosen to receive the offer since
they seem more likely to respond positively
Linear and Logistic Regression
§ In linear regression, the effect of one unit change in the independent variable on the
dependent variable is assumed to be a constant represented by the slope of a straight line
§ For logistic regression the effect of a one-unit increase in the predictor variable varies along
an s-shaped curve . At the extremes, a one-unit change has very little effect, but in the
middle a one unit change has a fairly large effect
Logistic Regression: Formula
§ y = α + βx +ε
§ Where: y= dependent variable,
x= predictor variable,
α = constant (which is estimated by linear regression and often called

intercerpt),
β = the effect of x on y (also estimated by linear regression),

ε = error term
Logistic Regression: Transformation Steps
p
§ Step 1: If p represents the probability of an event occurring, take the ratio 1 - p
Since p is a positive quantity less than 1, the range of this expression is 0 to infinity
æ p ö
§ Step 2: Take the logarithm of this ratio: logçç ÷÷
è 1 - p ø
This transformation allows the range of values for this expression to lie between
negative infinity and positive infinity
Logistic Regression: Transformation Steps
" p %
§ Step 3: The value z = log $ ' can be considered as the dependent variable and a linear
# 1− p &
relationship of this value with predictor variables in the form z = α + β x + e can be written.
a and β coefficients can be estimated
§ Step 4: In order to obtain the predicted probability p, a back transformation is to be done:

Since logæç p ö÷ = z = α + β x + e à e z = $
" p %
ç ÷ '
è1- p ø # 1− p &
§ Then calculate the probability p of an event occurring, the variable of interest, as
" ez % " 1 %
p=$ z'
=$ −z '
# 1− e & # 1+ e &
Logistic Regression: Interpretation of Coefficients
§ Interpret the coefficients carefully

§ If the logit regression coefficient β = 2.303 e β = e 2.303 = 10
à When the independent variable x increases one unit, the odds that the dependent
variable equals 1 increase by a factor of 10, keeping all other variables constant.
§ Sample of contract extension and sales
Contract extension Number of sales calls
1 2
1 4
0 6
0 2
1 4
1 8
0 3
0 0
0 2
0 5
0 0
1 2
1 8
1 4
Logistic Regression: Interpretation of Coefficients
Odds of Logistic Regression Example
§ It expects that sales calls impact the probability of contract extension. Thus, it estimates the
model:
æ 1 ö
=ç -a - b ( sales calls ) ÷
§ Prob (contract extension) è1+ e ø
§ The resulting estimates are: α= -1.37; β= 0.39
§ Odds of logistic regression example
Sales calls = 0 Sales calls = 1

Odds (exp(α+β *sales 0.253 0.375
calls))
Probability of contract 0.202 0.273

extension
Difference in probabilities 0.071
Logistic Regression: Evaluation
§ Evaluation
§ For logistic regression the effect of a one-unit increase in the predicor variable varies
along an s-shaped curve
§ This means that at the extremes, a one-unit change has a very little effect, but in the
center a one-unit change has a fairly large effect
Techniques to Evaluate Alternative Customer Selection
Strategies
§ A critical step to decide which model to use for selecting and targeting potential customers
is to evaluate alternative selection strategies
§ When several alternative models are available for customer selection, we need to compare
their relative predition quality
§ Divide the data into a training (calibration) dataset (2/3 of the data) and a holdout (test)
sample (1/3 of the data)
§ The models are estimated based on the training dataset
§ Select the model that generalizes best from the training to the data
§ Techniques
§ Misclassification Rate
§ Lift Analysis
Misclassification Rate
§ Estimate the different models on two thirds of the available data

§ Calculate the misclassificaion rate on the remaining third to check for model performance
§ Example
Predicted
1 0 Totals
Observed 1 726 56 782
0 173 504 677
Totals 899 560 1,459
§ The number of mispredictions is the sum of the off-diagonal entries (56+173=229)

§ Misclassification Rate: 229/ 1,459 = 15,7%
LIFT Analysis
§ Lift Charts
§ Lifts charts show how much better the current model performs against the results
expected if no model was used (base model)
§ Can be used to track a model’s performance over time, or to compare a model’s
performance on different samples
§ Lift% = (Response rate for each decile) / (Overall response rate) *100
§ Cumulative lift% = (Cumulative response rate) / (Overall response rate) *100
§ Cumulative response rate = cumulative number of buyers / Number of customers per
decile
Lift Performance Illustration
§ Lift and Cumulative Lift
Number of Number of Number of Response Lift Cumulative

decile customers buyers rate % lift
1 5,000 1,759 35.18 3.09 3.09
2 5,000 1,126 22.52 1.98 5.07
3 5,000 998 19.96 1.75 6.82
4 5,000 554 11.08 0.97 7.80
5 5,000 449 8.98 0.79 8.59
6 5,000 337 6.74 0.59 9.18
7 5,000 221 4.42 0.39 9.57
8 5,000 113 2.26 0.20 9.76
9 5,000 89 1.78 0.16 9.92
10 5,000 45 0.90 0.08 10.00
Total 50,000 5,691 11.38
§ The Decile analysis distributes customers into ten equal size groups
§ For a model that performs well, customers in the first decile exhibit the highest response
rate
Lift Analysis
3.50
3.09
3.00
2.50
1.98
2.00 1.75
Lift
1.50
0.97
1.00 0.79
0.59
0.39
0.50 0.20 0.16 0.08
0.00
1 2 3 4 5 6 7 8 9 10
Deciles
§ Lifts that exceed 1 indicate better than average performance
§ Less than 1 indicate a poorer than average performance
§ For the top decile the lift is 3.09; indicates that by targeting only these customers are
expected to yield 3.09 times the number of buyers found by randomly mailing the same
number of customers
Cumulative Lift Analysis
10
7
Cumulative Lift
6
5
0
0 1 2 3 4 5 6 7 8 9 10
Decile
§ The cumulative lifts for the model reveal what proportion of responders we can expect to
gain from targeting a specific percentage of customers using the model
§ Choosing the top 30 % of the customers from the top three deciles will obtain 68% of
the total responders
Comparison of Models
10
9
8
7
Cumulative Lift
6 Logistic
Past Customer Value
5
RFM
4 No Model
3
2
1
0
0 1 2 3 4 5 6 7 8 9 10
Deciles
§ Logistic models tend to provide the best lift performance

§ The past customer value approach provides the next best performance
§ The traditional RFM approach exhibits the poorest performance
Minicase: United Parcel Service (UPS)- Differentiating
Customer Service According to Customer Value
§ Headquartered in Atlanta, Georgia, UPS is one of the world’s largest package delivery
companies
§ UPS decided to create service levels for managing temperature-sensitive healthcare
products
§ The service levels were an addition to an already existing Temperature True portfolio.
§ The change allowed healthcare companies to choose from tiered service levels based on
the degree of temperature control needed:
§ UPS Temperature True Plus
§ UPS Temperature True Standard
§ UPS Temperature True Saver
§ Service options include:
§ Quality assurance support, 24/7 control tower monitoring, packaging consulting
services, best-in-class technology, proactive monitoring and intervention services, risk
management and contingency planning, and deep regulatory compliance expertise.
Summary
§ The higher the computed RFM score, the more profitable the customer is expected to be in
the future
§ The PCV is another important metric, in which the value of a customer is determined
based on the total contribution (toward profits) provided by the customer in the past after
adjusting for the time value of money
§ The LTV reflects the long-term economic value of a customer
§ The sum of the LTV of all the customers of a firm represents the customer equity (CE)
§ Firms employ different customer selection strategies to target the right customers
§ Lift analysis, decile analysis and cumulative lift analysis are various techniques firms use to
evaluate alternative selection strategies
§ Logistic Regression is superior to past customer value and RFM techniques
International Perspectives: Did you Know?
§ Increased competition in the marketplace has forced companies to become more customer
focused
§ Companies have to find ways to separate their company from the competition
§ Example: SAS and Radica
§ Data analytics and wearable technology

§ Laing O’Rourke
§ Collected data can save company money in medical related costs
§ CRM
§ Wearable technology provides valuable data that can be used to engage in
effective targeted marketing campaigns

CRM Chapter 6

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

CRM Chapter 6

Uploaded by

Copyright:

Available Formats

Chapter 6:

Customer Analytics Part II

Size Share of Past

• Elapsed time since a

• Amount that a customer

§ Technique to evaluate customer behavior and customer value

§ Graph depicts the distribution of relative frequencies of customer groups assigned to

2.00% 1.67% 1.68%

Frequency Code 1-5

Monetary Value Code (1-5)

Last purchase: Group 1 ß R=1

Last purchase: ß R=5

Customer Sorted Once Sorted Five Sorted Twenty-Five

§ Breakeven Index (BEI) = ((Actual response rate – BE) / BE)*100

600 Break-Even Index

Test Full customer base RFM Selection

§ Regression techniques to compute the relative weights of the R, F, and M metrics

Customer Purchase Recency Assigned Weighted

Customer Purchase Frequency Assigned Weighted

§ Monetary Value: 10 percent of the $-value of purchase with 12 months; Maximum = 25

Customer Purchase Value of Assigned Weighted

Smith 1 112 112

§ Computation of Customer Profitability (PCV)

Jan Feb March April May

Purchase Amount ($) 800 50 50 30 20

Gross contribution (GC) = purchase amount X contribution margin

§ PCVi = 6*(1+0.0125)o +9*(1+0.0125)1 +15*(1+0.0125)2 +15*(1+0.0125)3+ 240*(1+0.0125)4

§ Comparing the PCV of a set of customers leads to a prioritization of directing future

§ Multi-period evaluation of a customer’s value to the firm

d = interest (or discount) rate,

GCit = gross contribution of customer i at time t,

T = observation time horizon,

§ LTV is a measure of a single customer’s worth to the firm

§ Used for pedagogical and conceptual purposes

§ CM and T from managerial judgment or from actual purchase data.

d = interest (or discount) rate,

T = observation time horizon,

Sit = sales value to customer i at time t,

DCit = direct costs of products by customer i at time t,

MCit = marketing costs directed at customer i at time t,

§ Depending on data availability, it can be enhanced by including service-related cost,

d = interest (or discount) rate,

T = observation time horizon,

Rrt = average retention rate at time t (it is possible to use an individual

GCit = gross contribution of customer i at time t,

ACi = costs of acquiring customer i (acquisition costs)

§ In this equation the term ∏ Rr is actually the survival rate SRt

§ How long is the Lifetime Duration?

§ Incorporating externalities in the LTV

d = interest (or discount) rate,

T = observation time horizon,

nit = number of new acquisition at time t due to referrals customer i,

Average CRV after 1 year

Average LTV High Affluents Champions

This table is adapted from Kumar, Petersen , and

§ Alternative ways to account for externalities

§ Where: CEj = customer equity of brand j,

0 120 0.3 36 20 0.4 0.4 400 16 16 6,400

§ CRM at Work: The Kroger Co.

Response Internal External

Transactions, demographics, lifestyle Demographics, lifestyle

§ Used to identify the best predictors of a 0/1 or binary dependent variable

§ The algorithm procedure

§ PCVi = 6(1+0.0125)o +9(1+0.0125)1 +15(1+0.0125)2 +15(1+0.0125)3+ 240*(1+0.0125)4