Recapturing Lost Customers

Recapturing Lost Customers
Author(s): Jacquelyn S. Thomas, Robert C. Blattberg and Edward J. Fox

Source: Journal of Marketing Research, Vol. 41, No. 1 (Feb., 2004), pp. 31-45
Published by: American Marketing Association
Stable URL: http://www.jstor.org/stable/30162310
Accessed: 23-11-2017 16:02 UTC
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms
American Marketing Association is collaborating with JSTOR to digitize, preserve and

extend access to Journal of Marketing Research
This content downloaded from 193.175.6.21 on Thu, 23 Nov 2017 16:02:55 UTC
All use subject to http://about.jstor.org/terms
JACQUELYN S. THOMAS, ROBERT C. BLATTBERG, and EDWARD J. FOX*
For both academics and practitioners, the dominant focus of customer

relationship management has been customer retention. The authors
assert that customer winback should also be an important part of a cus-
tomer relationship management strategy. Customer winback focuses on
the reinitiation and management of relationships with customers who
have lapsed or defected from a firm. In some cases, firms engage in
extensive efforts to reacquire lapsed customers or defectors, and a corn-
mon tactic is lowering the price to reacquire a customer. This investiga-
tion goes beyond the reacquisition pricing strategy and also examines the
optimal pricing strategy when the customer has decided to reinitiate the
relationship. By simultaneously modeling reacquisition and duration of
the second tenure with the firm, the authors determine that the optimal
pricing strategy for their application involves a low reacquisition price and
higher prices when customers have been reacquired. In addition to pric-
ing strategy, they also discuss the implications of their findings for target-
ing lapsed customers for reacquisition.
Recapturing Lost Customers
Loyalty and retention have been the dominant themes Table 1

among scholars interested in customer relationship manage- ANNUAL CUSTOMER DEFECTION RATES
ment (CRM). Books and articles have been written and busi-
nesses have been developed around the central theme of the Industry Defection Rate
management and maintenance of customer relationships 22%
Internet service providers
(see, e.g., Reichheld 1996). Although progress has been U.S. long distance (telephone) 30%
made in the management of customer relationships, there German mobile telephone market 25%
are still high defection rates. Table 1, which we have adapted Clothing catalogs 25%
Residential tree and lawn care 32%
from the work of Griffin and Lowenstein (2001), shows cus-
Newspaper subscriptions 66%
tomer defection rates across various industries.
Although all aspects of CRM need to be assessed and Adapted from Griffin and Lowenstein (2001).
strategies and tactics developed, an area that has been
largely neglected in the marketing literature is customer
"winback" strategies. Customer winback is the process of opportunity for firms to increase or maintain a customer
firms' revitalizing relationships with customers who have base is the mining and evaluation of the firm's database of
defected. The importance and impact of customer winback defected customers. Stauss and Friege (1999) make this
as a key element in a firm's CRM strategy cannot be under- argument even more convincing in a case study in which
estimated. Research has shown that a firm has a 60% to 70% they find that the net return on investment from a new cus-
chance of successfully repeat-selling to an "active" customer obtained from an external list is 23% compared with
a 214% return on investment from the reinstatement of a
tomer, a 20% to 40% chance of successfully repeat-selling
customer who has defected.
to a lost customer, and only a 5% to 20% chance of suc-
cessfully closing the sale on a brand new customer (Griffin A critical element in the process of firms' recapturing of
and Lowenstein 2001). These statistics suggest that a key lost customers is the assessment of customer profitability.
Customer lifetime value (LTV) is a central profitability met-
ric in analysis of customer relationships; it is typically
*Jacquelyn S. Thomas is an associate professor, Integrated Marketing
Communications Department, Medill School of Journalism (e-mail: defined as the net present value (NPV) of the customer's
jakki@northwestern.edu), and Robert C. Blattberg is the Polk Bros. Pro- profitability throughout the customer-firm relationship
fessor of Retailing, J.L. Kellogg Graduate School of Management (e-mail: (Dwyer 1989). However, when it comes to the recapturing
r-blattberg@kellogg.northwestern.edu), Northwestern University. Edward of lost customers, the second lifetime value (SLTV) of the
J. Fox is W.R. Howell Director of the JCPenney Center for Retail Excel-
customer (Stauss and Friege 1999) is the metric of interest.
lence and Assistant Professor of Marketing, Edwin Cox School of Business,
Southern Methodist University (e-mail: efox@mail.cox.smu.edu). This metric focuses only on the NPV generated after a cus-
tomer has been reacquired. Specifically, Stauss and Friege
Journal of Marketing Research

31 Vol. XLI (February 2004), 31-45
32 JOURNAL OF MARKETING RESEARCH, FEBRUARY 2004
(1999) define SLTV as the future value

teristicof a recaptured
of these cus-
services is that the quantity purchased is typ-
ically to
tomer. They assert that when it comes one unit of the service. Thus,
development demand is typically
and
determined
implementation of a customer recapture by the durationexpecta-
program, of the customer-firm relation-
tions about the potential future value
ship. of a recaptured
This context cus-benefit that rela-
also offers the additional
tionship
tomer should be the guiding factor. durations and the
Logically, lapses are
SLTVclearly identified.
should guide the decisions with respect Although
tothese relationships
which tend to involve contracts, the
customers
should be recaptured and how much should service we analyze allows customers
be spent to reac-to defect at any time
quire them. without a penalty. This service also may change its prices
Some firms engage in extensive efforts to recapture over time. Thus, the context of our research offers the bene-
defected customers or to reactivate lapsed customers. For fit of explicitly identifying relationship duration, but it is not
example, during the long-distance telephone wars, one seg- limited by the pricing and duration constraints that often
ment of customers frequently switched providers. Some cus- characterize contracted services.
tomers switched to benefit from the introductory offer of a In the next section, we review the relevant literature on
competing provider, whereas others simply wanted to solicit pricing and CRM. In subsequent sections, we present a the-
a better offer from the original provider (Marple and Zim- ory of restart customer behavior, detail the modeling frame-
merman 1999). To recapture lost customers, telecommuni- work used to test our theory, describe the data and variables
cations firms engaged in aggressive "come-back" cam- used to estimate the model, present the empirical results,
paigns. When the original provider approached customers to and examine the financial implications of our analysis.
come back, they typically presented them with offers that Finally, we conclude with a general discussion of our analy-
were better than the original offer. Thus, the consumer usu- sis and our study's limitations.
ally benefited. However, reacquisition costs (e.g., reactiva-
SLTV AND WINBACK LITERATURE REVIEW
tion fees, telemarketing efforts, come-back cash incentives)
often caused the SLTV for the reacquired customers to be On the surface, research on dynamic pricing seems an
negative, thereby representing a net loss to the provider appropriate foundation on which to base our research. How-
firms. This example illustrates the importance of SLTV to ever, this research domain provides limited insight into the
the strategy and tactics of customer reacquisition. use of pricing to reacquire and retain lapsed customers. Tra-
The focus of our article is on determining how firms ditional pricing research does not focus on the long-term
should price customers when reacquiring them and how they relationship between the customer and the firm. This is evi-
should price them when they have been reacquired. In prac- denced by the fact that demand models typically used in
tice, the favored approach is to offer restarts lower prices for pricing research do not monitor the churn or retention
the same product. To stimulate purchase activity, behavior of individual customers (for a review, see Mahajan,
Amazon.com offers lapsed customers discounts on their Mueller, and Bass 1993). Kalyanam (1996) shows that dif-
next purchases. Similarly, HoneyBaked, the ham company, ferent ways of specifying demand can lead to different
offers a $10 gift certificate to reactivate customers (Schmid profit-maximizing prices.
1998), and Self Care, a health care products marketer, offers In contrast, the CRM literature focuses on the individual
discounts of up to 25% to customers who have not ordered customer and emphasizes LTV, which is a relatively new
from the company in 18 to 24 months (Kiley 1996). Deter- area and has been the subject of little empirical research,
mining the optimal price for recapturing a customer is only particularly research that addresses customer reacquisition
part of the challenge in firms' customer winback strategy. and SLTV. For example, the research that is most similar to
An equally important decision is how to price when the cus- ours in that it explicitly addresses customer winback is that
tomer has been reacquired. of Stauss and Friege (1999). They explore a concept that
The objective of this research is to assess optimal pricing they term "regain management," which they define as the
strategies for the recapture of lost customers and the man- process of winning back customers who either give notice to
agement of the SLTV of the recaptured customer (i.e., a terminate or have already ended the relationship. Stauss and
restart customer). Specifically, we examine the relationship Friege present a conceptual framework for customer win-
between the price used to recapture the customer and the back that entails regain analysis (i.e., determining which
subsequent price when the customer has been reacquired. customers have defected and why), regain actions (i.e.,
Prior research has shown that the acquisition of a cus- engaging in dialogue with the customer to determine the
tomer affects the future relationship that a firm has with that appropriate regain offer), and regain controls (i.e., the profit/
customer (Thomas 2001). In this context, we examine how loss analysis of the regain actions).
the reacquisition of a customer correlates and influences the Our research addresses the three areas of Stauss and
reinstated customer-firm relationship. A key consideration Friege's (1999) framework in a more rigorous context.
in this analysis is the influence of the customer's relation- Unlike Stauss and Friege, we perform a statistical analysis
ship with the firm before reactivation. Using observable of defected and restart customers to determine drivers of
characteristics about the prior relationship (e.g., tenure, pric- reacquisition, we model both the reacquisition and the reten-
ing, duration of lapse), we also draw conclusions about tion of a lapsed customers, and we focus explicitly on pric-
which lapsed customers are the most profitable targets for ing as a means to recapture customers. Although our data do
reacquisition. not enable us to determine why the lapse occurred, prior
We explore these questions in the context of services such research has shown that unfavorable price perceptions have
as newspapers and magazine subscriptions and organiza- a direct effect on a customer's intention to switch providers
tional memberships (e.g., health clubs). A common charac- (Keaveny 1995).
Recapturing Lost Customers 33
Building on the framework introduced

ticularly relevant. The first is critical for targeting; that is
Friege (1999), Griffin
which lapsedand Lowenstein
customers should the firm target for reacquis (
general outline for winning back
tion? The second issue is lost
critical for making cust
offer decision
practices from industry, they
that is, what are the optimal highlight
reacquisition and follow-upt
(after reacquisition)
highly trained winback teams prices?
and customer
tems. However, counter to the mantra of
they assert that notTargeting Decision
all customers shou
According to Griffin andapproach
A popular Lowenstein, firm
for making targeting decisions in
SLTV, segment customers on
direct marketing firms the
is recency, basis
frequency, of
and monetar
ate customers in those segments
value analysis. Although weaknesses into deter
this approach have
defected. Assessments of SLTV and rationale for defection
been identified, a widely held belief among direct marketer
guide the targeting and offers made in the regain process. is that customers who have bought most recently and mor
Using case analyses, Griffin and Lowenstein find that the often and have the highest monetary value are more likely t
duration of a customer's lapse and the way that customer respond favorably to subsequent offers (Hughes 1996). Th
was acquired affect SLTV estimates. Our research incorpo- belief is consistent with other research findings. For exam
rates and extends their insights into the domain of pricing ple, Schmittlein and Peterson (1994) find that in a brokerag
strategies for customer winback. context, customers who make fewer transactions (i.e., pur
Unlike the research we have mentioned, the majority of chase less frequently) are most likely to terminate their rel
other CRM research does not address customer winback
tionship with the firm. Boulding, Kalra, and Staelin (1999
directly. However, a few researchers have addressed pricing find that, more generally, consumers' prior experience wit
issues related to CRM. Bolton and Lemon (1999) show thata service affects their subsequent attitudes and assessments
customers' use of two continuously provided services partly of that service.
depends on prices. Bolton, Kannan, and Bramlett (2000) Bolton, Kannan, and Bramlett (2000) find that experience
find that a price gain (i.e., a decrease in price) has a signifi-
with a product, measured by the number of prior transac
cant impact on repatronage, but a price loss (i.e., a price
tions, is positively associated with a higher likelihood o
increase) does not. This finding is consistent with that ofrepatronage. It is noteworthy that this result holds regardless
Krishnamurthi, Mazumdar, and Raj (1992), who show thatof whether repatronage is measured either in terms of the
for customers who switch often among consumer packaged- decision to stay or to terminate the relationship or in term
goods brands, price gains have a larger impact on brand
of how much to use the service. They explain this finding b
choice than do price losses. relating a customer's prior experience to repatronage inten
Another common theme in the CRM literature is the
tions and to the customer's desire to maintain the status quo
degree of price sensitivity among loyal customers. PreviousSpecifically, they assert that prior experience drives cus-
researchers have asserted that loyal customers are willing tomer
to expectations and intentions. They argue that inten
pay higher prices (Reichheld and Sasser 1990). Reinartz andtions are strongly related to the actual decision because cus
Kumar (2000) test this assertion empirically. Using atomers strive to maintain the status quo; however, they als
median split, they segment customers into long- versus find that a customer's level of satisfaction can moderate this
short-life customers and then examine the mean price paid result.
by the two segments. They determine that long-life cus- Although the preceding articles are unique, the consistent
tomers actually pay a lower mean price than do short-life theme in the results is that previous experience affects cus
customers.1 Although Reinartz and Kumar never actuallytomer behavior. For lapsed or defected customers, experi-
estimate price sensitivity, they draw inferences about the ence can be measured by their prior tenure (which we refe
price sensitivity of customers who have longer relationships
to as Tenure 1 or first tenure) with the firm. Consistent with
with the firm compared with that of customers whose
prior research, we assert the following:
tenures are shorter. Our research is distinct in that we model
customer reacquisition and duration as a function of price. H : For restart customers, the length of their first tenure with
firm is positively related to (a) the reacquisition probability
Thus, we actually estimate price sensitivity. In addition, our
and (b) their subsequent tenure in a reinitiated relationship.
focus is not on long- versus short-life customers but on the
pricing strategy for restart customers who may have eitherAlthough the normative beliefs of direct marketers about
long or short prior relationships with the firm. which customers to target for repurchasing may hold in
many cases, in the context of this research, it is important to
TOWARD A THEORY OF RESTART CUSTOMER
acknowledge that the relationship was terminated. The fina
BEHAVIOR
ity of this termination is not known at the time when the firm
evaluates potential reacquisition candidates. This raises the
There are two basic issues that we address in our theory
question, How does the amount of time elapsed since the
about restart customer behavior: (1) the nature and influence
last the
of the prior relationship on customer reacquisition and purchase (or the length of the lapse) affect the chance
of reacquisition
subsequent relationship that evolves and (2) the responsive- and the nature of the relationship that ma
ensue?
ness of restart customers to price. These two issues are par-
Prior research has asserted that with the passage of time,
customers adapt to the new level of service provided by the
switched-to
1 An inquiry with Reinartz and Kumar revealed that this result held even firm (Ganesh, Arnold, and Reynolds 2000). In
when they accounted for basic differences in product categories. addition, customers who have switched to a new firm after
having experienced another firm'sofservice exhibit

supply and demand higher
assert that higher prices lead to lower
demand
levels of loyalty and repeat patronage to the(e.g., Einhorn 1994). This general
switched-to firm principle can be
than do patrons of the firm who had
directly never
applied experienced
to the reacquisition price offer.
another provider (Ganesh, Arnold, and Reynolds 2000).
H4: The reacquisition rate is higher if the price offered is lower.
Logically, the longer the time since the last purchase, the
more likely a lapsed customer is Consistent
to have withengaged
economic theory, a newReinartz and Kumar
service or simply to have developed new
(2000) behaviors.
show that On pay
long-life customers the lower mean prices
basis of this reasoning, we believethan thatdo short-life
the priorcustomers; research
theirs was one of the first CRM
suggests the following hypothesis: articles to demonstrate this unique result. However, it is
worth acknowledging that this conclusion is not based on a
H2: The greater the time since last purchase, the lower is the
statistical estimate of price sensitivity but on a median split
likelihood of customer recapture.
of customers based on relationship duration.
Note that this hypothesis is still well We usefounded eventraditional
caution in applying if the economic theory
customer has not switched to a new to theprovider,
price responsebecause
of reacquired itcustomers,
is because
consistent with status quo bias, that economics
is, antypically focuses on price
exaggerated pref- responsiveness in
erence for the current state or discrete
inaction (Samuelson
transactions. In contrast, CRM and
considers the long-
Zeckhauser 1988). term effects of price on customer relationships. Because of
To gain insight into the relationship between
the long-term lapse
perspective, dura- to price may be
responsiveness
tion and the duration of the reinitiated relationship
more complex. For example, Bolton(whichand Lemon (1999) use
we refer to as Tenure 2 or second tenure),
the concept of we paymentreference
equity (i.e., thethe
customer's percep-
theory of cognitive dissonance (Festinger
tion of fairness with 1957).
respect toA the key
exchange of payment
premise of dissonance theory is that dissonance,
for service usage) to discussor consumers'
lack ofuse of services.
fit between two elements (e.g., attitudes
Their research shows and thatbehaviors,
consumers seek to maintain pay-
behavioral decisions and commitments),ment equity gives rise
in a service to pres-
relationship and adjust usage lev-
sures to eliminate or reduce the dissonance. This can be con- els in response to price changes. Specifically, Bolton and
ceptualized as a drive to obtain consistency. Lemon find that service usage levels may increase as price
For any lapsed customer, the decision to reinitiate a rela- increases in order to maintain equity in the relationship.
tionship with the firm creates dissonance relative to the However, the implied positive relationship between price
customer's prior state (i.e., inactivity) with the firm. Disso- and usage contradicts both the long history in economics
nance can vary in magnitude and is moderated by the of the negative effect of price on demand and the com-
importance or intensity with which attitudes are held pelling empirical evidence across many disciplines. Thus,
(Eagly and Chaiken 1993; Festinger 1957). Applying this on the basis of weight of evidence, we hypothesize the
idea to our research, we argue that longer lapses may rep- following:
resent more extreme attitudes. Therefore, the decision to
H5: The second tenure is longer if the retention prices are lower.
reinitiate a relationship after a long lapse results in a
greater amount of dissonance relative to the decision to Although we derive H5 from the history of pricing research,
reinitiate a relationship after a shorter lapse. Cohen (1960) we believe that it is important to test this hypothesis and
asserts that the greater the amount of dissonance, the either to reinforce the similarity between CRM and a
stronger are the attempts to reduce it. This implies that cus- transaction-oriented business perspective or to highlight the
tomers who have had longer lapses make stronger attempts unique perspective of CRM.
to reduce the dissonance between their prior behavior (i.e., To understand further how price affects reacquisition and
the lapse) and their decision to reengage in a relationship repatronage, we revisit the issue of prior experience. Bould-
with the firm. ing, Kalra, and Staelin (1999) show that customers give dif-
To reduce or resolve the dissonance that occurs after a ferent weights to prior experience and current experience.
decision, a person attempts to engage in postdecision pro- They find that as customers gain confidence or experience
cessing that reinforces the new decision that has been madewith a product, they weight their prior assessment of a given
(Festinger 1957, 1964). In the case of reacquired customers, service more heavily than they do new information about the
the current decision is the choice to reengage in a relation-service. Thus, it might be expected that when restart cus-
ship with the firm. In this context, the second tenure can be tomers assess their reinitiated relationship with the firm,
evidence of the reacquired customer's degree of reinforce-they reflect on their prior relationship with the firm and give
ment processing. Thus, because customers with longer significant consideration to that assessment. This behavior is
lapses make stronger attempts to reinforce the reacquisition consistent with the existence of reference prices in con-
decision, they will have longer subsequent relationshipssumer decision making.
than customers with shorter lapses. The reference price literature enables us to generalize that
consumers use prior prices in the formation of reference
H3: Longer lapse durations are positively associated with longer
second tenures. prices and that reference prices have a significant impact on
demand (Kalyanaram and Winer 1995). The specific combi-
nation of prior prices and process by which internal refer-
Offer Decision
ence prices are formed remains an open issue (for a review,
The most basic question about the price offer decision is, see Kalyanaram and Winer 1995). For customer winback,
How will restart consumers respond to price? General laws the logical reference point is the customer's prior relation-
ship Dwyer (1989) outlines alternative

with the firm. Consistent withdiscrete Markov
these models ar
H4, we propose the for customer migration, assuming an always-a-share sce-
following:
nario, and for customer retention, assuming a lost-for-good
H6: The difference between the reacquisition p
scenario. When the initial share of the customer is zero (i.e.,
the last price paid before lapse (i.e., reacq
customers
minus last price in prior are relationship)
identified before having developednegativ
a relation-
ship with the firm), the models can be interpreted as incor-
probability of reacquisition.
porating customer acquisition.
It is notable that this hypothesis
Pfeifer and Carraway (2000) and Rust,allowsZeithaml, and for
ence to be either positive or negative.
Lemon (2000) generalize Dwyer's (1989) approach by
Further drawing on reference
incorporating both migration and price litera
retention. Of particular
that differences between the
interest, Rust, Lemon, new
and Zeithaml price
(2001) allow for thean
influence how restartinvestigation customers
of other decision variables, such assess
as price, by
This assertion is supported by asarguments
modeling transition probabilities a function of covariates.
assessments of value,Usingwhich directly
alternative approaches affect
such as selection models or
tinuity, are based on differences relative
decision calculus, researchers have explicitly modeled both to
(e.g., Bolton 1998; acquisition
Thaler 1985).
and retention More
simultaneously (Berger and Nasr- sp
and Colgate (2001) show that price percept
Bechwati 2001; Blattberg and Deighton 1996; Thomas
affect customer retention. Consistent with t
2001). Such approaches are drawing increasing interest
we assert the following hypothesis:
from researchers in the modeling of customer relationships.
Another CRM literature stream assumes that customers'
H7: The difference between the current price
last price paid probability
before of purchasing
lapse again (given
(i.e., that they do not
current pr
price in prior explicitly terminate
relationship) their relationships withaffects
negatively the firm)
the second tenure.
depends not only on their current state but also on their pur-
chase history. Schmittlein, Morrison, and Columbo (1987)
H6 and H7 can have significant implications for the offer and
and Schmittlein and Peterson (1994) use a stochastic mod-
targeting decisions. If supported, these hypotheses suggest
eling framework, the Pareto/NBD, for this purpose. Reinartz
that the price paid in the prior relationship anchors cus-
and Kumar (2000) dichotomize the continuous probability
tomers' perceptions and guides their subsequent behavior in
predictions of the Pareto/NBD model and, using individual
the reinitiated relationship. This implies that firms should
customer cost information, are able to estimate LTV. The
target customers for reacquisition and make the offer based
Pareto/NPD allows for continuous rather than discrete time,
on prices paid before the lapse.
but it does not explicitly model price or other decision vari-
The assertion that customers respond to differences in
ables as covariates.
price raises the question of whether their response is the
Given this history, we focus on five important character-
same for gains (e.g., price decreases) as it is for losses (e.g.,
istics of our investigation in making our modeling choices.
price increases). Consistent with prospect theory (Einhorn
First, we focus exclusively on the second lifetime (we term
and Hogarth 1981; Kahneman and Tversky 1979; Thaler
this "single spell," in line with the statistics literature).
1985), another finding of the reference price literature is that
Although we may lose some generality by not adapting an
customers respond more to losses than gains (Kalyanaram
and Winer 1995). However, CRM researchers have observed always-a-share model, our focused approach is necessary to
address the four remaining characteristics, thereby enabling
a different effect with respect to price. Bolton and Lemon
us to provide managerially useful insights about the second
(1999) find that gains in price (i.e., decreases) have a larger
lifetime. Second, because our hypotheses relate to the
impact on usage amount than do losses in price (i.e.,
effects of price, we model price explicitly using covariates.
increases). Similarly, Bolton, Kannan, and Bramlett (2000)
Third, because many of our hypotheses relate to second
find that gains in price have a larger effect than do losses on
tenure duration and because the data set we analyze lacks
both the decision to stay in a relationship and the usage
consistent decision intervals, we select a continuous-time
level. Given the solid support for both sides of the issue, we
specification. Fourth, our hypotheses for customer reacqui-
refrain from making predictions about the impact of gains
sition and duration differ, which requires us to model the
versus losses with respect to prices, and we defer to the
empirical results. two as distinct processes. Fifth, the data set includes many
customers whose relationship with the firm exceeds the
MODEL DEVELOPMENT period we observe, so we must allow for censoring.
Given these requirements, we specify a hazard model.
Modeling Customer Relationships Bolton (1998) uses a hazard model in a CRM context to
The growing CRM literature includes several modeling examine the relationship between customer satisfaction and
approaches (for a review, see Jain and Singh 2002).customer
Some of duration. Hazard models are well established in
these articles model SLTV within the broader context of statistics, can incorporate covariates, and are adaptable to all
LTV. This implicitly assumes that the firm can regain lapsedtypes of censoring. Furthermore, hazard models have been
customers ("always a share") instead of customers beingshown to be well suited for analysis of duration data and
"lost for good." The most commonly used models are dis- superior to other common methods, such as logistic and
crete Markov chains, so called because (1) time periods are least squares regressions, in terms of stability, face validity,
discrete and (2) the probability of entering a particular stateand predictive accuracy (Helsen and Schmittlein 1993). To
in the subsequent period depends only on the current state.distinguish our approach from the more common propor-
tional hazards model (Cox

time t 1972), we refer
+ 81 + 82. This customer's to
duration consists it as
of two
hazard model" (it may also
subspells: be termed
one of length a "censored
81 (which is right censored) that is
model"). associated with the reacquisition price and one of length 82
that is associated with the second price. This approach
Model Specification
enables us to capture parsimoniously the effect of price on
As we stated previously, we focus on a single spell that second tenure duration.3
starts when the customer is reacquired by the firm and ends We model the duration component for customer i during
when the customer terminates the subsequent relationship. subspell si (si = 1, ..., Si) as a conditional regression with the
Our split hazard specification comprises separate reacquisi- following observation equation:
tion and duration components. The reacquisition component
measures the probability of recapturing a lapsed customer,
and the duration component predicts the length of the sec- (3) y
ond tenure, given that the firm successfully recaptures the
customer. This approach to linking acquisition and retention where yrsi is the latent duration of the relationship and cisi is
explicitly incorporates left censoring (Thomas 2001). the censoring value, or length of time that a given price was
For customer i (i = 1, C), we specify the reacquisition offered. If the customer terminates the relationship before
component as a latent variable probit with observation the price changes, then yIsi (i.e., it is observed). Other-
equation wise, the duration of the subspell is right censored. The
observed duration may also be limited by the observation
(1) Z horizon, in which case it is also right censored. Note that the
censoring value is known; it is a necessary condition to esti-
mate this model (Amemiya 1985, p. 363). We specify the
We model the latent dependent variable zi* with the linear latent duration of subspell si of customer i as
model
(4) ln(yrsi) = Xisi i3 + cisc
(2) z
where x;sii3 is the deterministic component, Eisi is the sto-
where wiNy is the deterministic component, is the sto-
chastic component, xisi is the customer's vector of predictors
chastic component, wi is the customer's vector of predictors,
during subspell si, and 13 is the associated parameter vector.
and y is the associated parameter vector. Probit specifica-
The dependent variable is log-transformed to approximate
tions have been used previously to model customer acquisi-
more closely the normality assumption about the residuals.4
tion (Hansotia and Wang 1997; Thomas 2001).
Given the likely relationship between the customer's
Modeling of the second tenure is somewhat complicated
acquisition and retention behavior (Thomas 2001), it is
by the firm's propensity to change the offer price during the
important that these components be linked. We explicitly
relationship. Although this occurs relatively infrequently
model this linkage, along with customer heterogeneity, by
(see the data description in the following section), we must
specifying variance components. We specify errors of the
nevertheless allow for price changes to affect the customer's
reacquisition and duration components, respectively, as
probability of terminating the relationship. We note that follows:
Amemiya (1985, pp. 433-35) shows that a split hazard
model of the form that we specify is simply a generalization (5) rh
of a single spell of a continuous Markov model. Recogniz-
ing this relationship, we adopt the continuous Markov
(6) Eisi
process assumption that the probability of the customer ter-
minating the relationship at any point in time is independent
where N(04) and N(0,0).
of the current duration (i.e., it is stationary). Using this For the reacquisition and duration components, respec-
assumption of stationarity, we partition the second tenuretively,
so ti and ai represent customer-specific preferences, and
that each period during which a given price is offered to the
vi and Eisi are random errors. As the subscripts oft and a
customer is a "subspell." Thus, each customer's second indicate, customer-specific preferences for reacquisition and
tenure consists of one or more subspells that differ only duration
in are not fixed; they are distributed across house-
the offer price. Moreover, the stationarity assumption holds. In this way, we allow for heterogeneity across cus-
implies that the duration of a subspell does not depend on tomers. Moreover, we allow the distributions of customer-
the length of prior subspells.2 To illustrate, consider a hypo-
specific preferences to be correlated so that BVN(0,X0),
thetical customer who is reacquired by the firm at time t as
where
a result of a reacquisition price offer, is offered a second
price at time t + 81, and then terminates the second tenure at
2We tested whether there were systematic differences between customers

a
with a single subspell and customers with multiple subspells by estimating
a more general specification of our model with a dummy variable that cap-
tured single versus multiple subspell customers as a predictor of duration. 3We empiri
The CAIC and BIC for this specification are 1289 and 1282, respectively. tion. We est
For the specification without the dummy variable, CAIC and BIC are 1285 dummy var
and 1279, respectively. We conclude that there is no evidence of systematic in duration
differences between the durations of customers with single and multiple 4Details of
subspells. from the aut
In this way, we allowoffers from the firm at any time before ourprefer
a customer's observation
relationship with theperiod. firm
Thus, our application
to does be not include
correlatsequential
offers to
tomer's preference for lapsed customers,
the duration and we conjecture
of aboutthat
their
ship. Allowing for correlation
response to such a sequence of between
offers. cus
for the discrete and Observations
continuous for the duration component
component differ from
reacquisition observations
ard specification is similar in in two ways. First,
spirit tosome acus- se
tomers have
which single stochastic multiple subspells.
error terms More specifically,
for of the
th
416 customers that werethe
are correlated. In summary, reacquired,error
140 received multiple
vari
2 and 4 are price offers during the observation period (an average of
2.19 price offers for customers who received multiple
(7) Var(rii prices), for a total of 582 subspells. Recall that each subspell
represents a different price offer from the firm during a
given customer's second tenure. Thus:
(8) Var(Eis)
We follow Ainslie and Rossi (1998) in estimating our
variance components specification in a Bayesian framework
by using Markov chain Monte Carlo methods. The
(9) Si
The second difference
simulation-based methods of estimating is thatdistribu-
posterior the price in ea
characteristic
tions of parameters in censored of that data
and missing subspell and so may
problems
have only recently become reacquisition price. Thus,
available (Casella andalthough most
George
1992; Gelfand and Smith 1990; tomers
forpaid a single price
application to(the reacquisition
censored
out their relationship
regressions, see Chib 1993). (A complete discussion of with the firm,
ourmore
received multiple price offers.
estimation procedure is available as a technical report on
request.) All the customers examined in this analy
newspaper seven days a week. When cust
DATA AND MODEL VARIABLES
receive the newspaper, they commit to a w
Data remains fixed for a given period (roughly
were unable to determine from the data wheth
The data we used to estimate the model come from a
precommits to buy for several consecutive p
newspaper subscription database and comprise 566 lapsed
less, the firm does not engage in price disco
customers targeted for reacquisition (i.e., C = 566).5 Of the
chase commitments of more than one perio
customers, 416 were successfully reacquired. The reacquisi-
unlike some other contractual selling agreem
tion component of the data includes a single observation for
tions are nonbinding; therefore, customers
each lapsed customer that consists of the reacquisition price
continue or terminate the subscription at an
offered, the result of that offer (i.e., successful or failed reac-
period. The average length of the second
quisition), the number of periods elapsed since the cus-
days (see Table 2). Of the customers who re
tomer's most recent purchase (each period as defined by the
tionships with the firm, 66.8% terminated
company is roughly one month), the price of the last pur-
during the two-year observation horizon. T
chase, and the length of the first tenure. Note that the firm
customer durations are right censored. Desc
made only one reacquisition offer per customer during our
are provided in Table 2.
observation period (i.e., customers who did not respond did
Model Variables
not receive multiple offers). Moreover, each customer in the
data set lapsed only once, so none had received reacquisition The first and most obvious variable included in the esti-
mation is price, that is, the current price offered/paid in the
reinitiated
5We removed customers who lapsed or defected as a result of relocation relationship (recall that customers are offered a
or vacation from the data.
single price for reacquisition but may be offered multiple
Table 2
DESCRIPTIVE STATISTICS
Total Sample Successfully Reacquired Sample

Standard Standard
Mean Deviation Range Mean Deviation Range
Reacquisition price offer $2.22 $.44 $3.00-$1.75 $2.28 5.44 $3.00-$1.75
Average retention price offera $2.43 $.46 $3.00-$1.75
Last price paid in prior relationship $2.32 $.47 $3.00-$1.75 $2.38 $.47 $3.00-$1.75
Reacquisition price difference $.10 $.41 $1.15-($1.25) $.10 $.48 $1.15-($1.25)
Average retention price decreasea $.12 $.30 $1.15-$.00
Average retention price increasea $.17 $.32 $1.25-$.00
Duration of lapse (in periods) 10.51 11.05 34-1 5.03 4.62 30-1
Prior tenure (in days) 179.75 279.97 684-1 229.00 282.05 684-7
Observed restart relationship tenure (in days) 177.42 217.92 684-2
aBecause the price is time varying, we present the averages over the duration of the reinitiated relationship.
Table Tenure
prices over the duration of the second tenure). The 3 1
variable measures the total duration of the customer's rela- COVARIATE IMPACT
tionship with the firm before lapsing. Lapse duration meas-

ures the number of periods elapsed since the customer's last B
purchase. Variable (Posterior Probability)a Elasticity
A fourth variable measures the difference in price with Impact on Customer Reacquisition
respect to the last price paid before the relationship lapsed, Price -.900 -.9000
which we suggest is a logical reference for the customer. We (.97)

Lapse duration -.686 -.6860
considered two alternative approaches to modeling price dif-
(1)
ferences, and we specified the preferred approach on the Tenure 1 .784 .7843
basis of model selection tests. The first approach is to have (1)
a single variable, price difference, which is the difference Price difference -.526 -.0526
between the last price in the prior relationship and the cur- (.91)
rent price paid or offered (i.e., price difference = current Impact on Relationship Duration
Price 1.194 1.1936
price - reference price). A positive value for price difference
(1)
means that the current offer price is higher than the last price
Lapse duration -.047 -.0473
observed by the customer. The second approach to modeling (.78)
price differences enables us to assess asymmetric response Tenure 1 .454 .4539
to gains and losses by defining price decrease (i.e., a gain) (1)
Price decrease 1.556 .1802
and price increase (i.e., a loss) variables. Specifically, we
(1)
define price decrease as the dollar amount of price decrease Price increase -.312 -.0523
relative to the last price the customer was offered by the firm (.76)
before the relationship lapsed; price increase is the dollar
al3 is less than or greater than zero.
amount of the price increase relative to the last price the firm
offered the customer. In interpreting the results, note that
both variables are coded as positive values. The data show
for the reacquisition model as well as posterior probabilities
that 26.6% of the prices paid or offered in the restart rela-
(in parentheses) that the parameter is less than or greater
tionship were decreases relative to the last price paid in the
than zero, depending on the sign of the posterior mean.
prior relationship, and 34.7% of the prices paid or offered
In terms of relative impact, the elasticities indicate that
were increases relative to the last price paid in the prior
the offer price has the largest effect on reacquisition likeli-
relationship.
hood. Tenure 1 and lapse duration also have a material influ-
RESULTS ence on the reacquisition outcome. Notably, price difference
has a much smaller effect. Thus, the absolute effect of price
Covariate Effects on the Reacquisition Probability
is much more important than the effect of price relative to
Consistent with Hia and H2, the reacquisition model the last price paid in the prior relationship. This is an impor-
shows that the probability of a firm reacquiring a customer
tant insight for managers because it suggests that reacquisi-
is higher if the lapse duration is shorter and/or if the first
tion strategies that emphasize decreasing price relative to the
tenure is longer. Consistent with economic theory and prior
H4, relationship are not likely to be effective. A more fruit-
the results also show that customers are more likely toful beapproach to winning back lapsed customers is simply to
reacquired if the reacquisition price is lower. offer a low price, regardless of the price that the customer
As we noted previously, we estimated alternative specifi-
was accustomed to paying before the lapse. This also
cations, one with the price difference variable only and implies
one that customers who previously paid low prices
which separated gains and losses. Comparing the corrected should not be enticed with significantly lower prices.
Akaike information criterion (CAIC) and the Bayesian
information criterion (BIC) for the two specifications, Covariate
we Effects on Length of the Second Tenure
find that the model specification with price difference wasTable 3 also reports posterior means for parameters (and
preferred to the specification with separate price increase
associated posterior probabilities that parameters are less
and price decrease variables.6 Thus, we report parameterthan or greater than zero) for the duration equation. As in the
estimates from the former specification. The results support
reacquisition decision, the duration elasticity of price has a
H6: The likelihood of a customer being reacquired decreases
higher magnitude than other predictors. However, there are
with the difference between the reacquisition price and several
the notable differences between the factors that affect
last price offered in the prior relationship. This result implies
reacquisition and the length of the second tenure. An impor-
that pricing in the prior relationship anchors response totantthedifference is the sign of the price effect. Consistent with
reacquisition offer. More generally, this supports the asser-
Bolton and Lemon (1999) and contrary to economic theory,
tion that customers make decisions about reinitiating the we find that higher retention prices lead to longer relation-
relationship based on comparisons with the lapsed relation-
ship durations. Comparing this to the other relevant CRM
ship. Table 3 reports posterior means of parameter estimates
literature, we find that it is consistent with the assertion that
loyal customers are willing to pay higher prices (Reichheld
6For the model specification with price decrease and price increase vari-
1996) but contradicts Reinartz and Kumar's (2000)
finding
ables, CAIC and BIC are 1288 and 1281, respectively. For the specification
that long-life customers pay lower average prices
than do
with only price difference, CAIC and BIC are 1285 and 1279, respectively.
Thus, the second specification represents a better balance of fitshort-life
and customers (in a catalog retailing context). How-
parsimony. ever, Reinartz and Kumar (2000, p. 28) state, "we expect
these factors to have differential

variance components impa
approach, we allow for customer het-
industries." erogeneity and correlation between customers' intrinsic
Bolton and Lemon preference
(1999) explain
to be reacquired this
and for their intrinsic preferenceeff
payment equity. Anto alternative explanation
maintain the relationship after reacquisition. The esti-
in reservation prices.
mates revealHowever,
that customers have a negative our
bias towardmo
allows for different reference
reacquisition (posterior mean of i = -1.592,prices
95% posterior a
Because replicate we
Bolton
probability that -2.789 .5_andi 5_ -.094), Lemon's
but when they have (
the presence
of heterogeneous
been reacquired, they are positively inclined referen
to continue the
relationship
reject this alternative (posterior mean of a = 1.743, 95%
explanation. Itposterior
is als
probability that 2.309). This positive
contexts that price sensitivity decreases intercept w
product familiarity,parameter
knowledge
estimate for the duration about
model is consistenthowwith
efficiently, or bias in favor
our previous of
conjecture that the
a customer's status
price sensitivity q
repeat purchases. Bothmay be lowerpayment equity an
in the reinitiated relationship.
The posterior
sensitivity are plausible mean of the correlation between the
explanations in two ou
Although customers did
preferences not
(estimated using adistinguish
is is -.09, which suggests
increases and price that the likelihood of the customer
decreases in the being reacquired
reacqu is
they made this distinction after
inversely related to the likelihood the
of the customer relat
remaining
reinitiated, which is in thein support
relationship. of who
In other words, customers H7.may be Th
highlights the importance
more inclined to restart of the(i.e.,
a relationship prior
customers who rel
the reinitiated relationship.
are easiest to win back) may not Consistent
always be the best cus-
research (Bolton, Kannan,
tomers in terms of retention.and Bramlett
Lemon 1999), the positive
The descriptive data in Table impact
2 reveal some short first ofand g
decreases) on secondsecond tenure observations. To
tenure assess the robustness of ourhas a
duration
tude than does the negative
results, we estimated our model with impact
three reduced data sets: of
increases). Furthermore,
elimination of (1) we find
all first tenures that
less than ten days, (2) all the
decreases on the second tenure
second tenures less than ten days, is statistical
and (3) all first and sec-
the effect of price increases is Compared
ond tenures less than ten days. not. with This
our parame- asy
to gains and losses points to
ter estimates from using the a notable
full data beh
set, we found no sign
Specifically, the results suggest
changes in any parameter estimate in thethat
three reduced whe
data
(i.e., price increases sets.
or In addition,
price in none of the three reduced data sets did are
decreases)
the customer's decisionany parameterto estimatereestablish a re
that had been statistically significant
have a noticeable effect. However,
at a = .05 in the full the
data set become nonsignificant, and in cus
fected by deviations
only one that do
case did a parameter that not
had not been suppor
significant
reestablish the relationship.
become significant.8 Thus,This behavior
we suggest that our inferences
Festinger's (1957, 1964) assertion
are robust to the inclusion that
or exclusion of unusually short p
postdecision processing
tenures. that reinforces th
have made.
Although Festinger's (1957, 1964) theory about postdeci- FORECASTING SLTV
sion processing can explain the effect of price comparisons, The asymmetric impact of price increases and decreases in
it does not explain our results about the effect of lapse dura- conjunction with the current price effect suggests that pric-
tion on the second tenure. Specifically, the results do not ing decisions for restart customers is complex. Specifically,
support H3, suggesting that there is no relationship between the parameter estimates suggest that the last price in the prior
the length of the lapse and the customer's second tenure. It relationship affects a customer's price sensitivity and behav-
is possible that this null result is due to the relationship ior in both reacquisition and retention. In this section, we
between lapse duration and the incorporation of customers' explore how the prior price affects the profitability of restart
preference heterogeneity. If unmodeled individual charac- customers and the optimal pricing strategy of the firm.
teristics, such as a propensity toward variety seeking or iner- Ultimately, a firm's targeting decision and offer decision
tial behavior (Bawa 1990) in newspaper subscription, were should be based on the expected profitability of a reacquired
correlated with the lengths of both the lapse and the second customer. We used estimates from the reacquisition and
tenure, any relationship between the lapse and the second duration models to predict the likelihood of a firm reacquir-
tenure would be obscured by the specification of preference ing and retaining a customer with certain characteristics. By
heterogeneity.? In summary, although firms are less likely to assuming specific marketing costs, we determined the
reacquire customers who have had longer lapses, when they expected SLTV of a potentially reacquired customer.9
have been reacquired, the length of the second tenure
appears to be unaffected.
8The price difference parameter for the reacquisition equation, which
Link Between Reacquisition and Length of the Second
Tenure was significant at a = .10 in the full data set, was significant at a = .05
(actual p-value = .041) in the second reduced data set (i.e., eliminating all
Another issue in our investigation is the link between second tenures of less than ten days).
91t is important to note that we computed expected SLTV and not the
reacquisition and duration of the second tenure. Using a NPV of the profits generated after a customer is reacquired. The difference
is that the expected SLTV takes into account the reacquisition profit or loss
7We thank an anonymous reviewer for pointing out the possibility that and discounts the post-reacquisition LTV by the probability of reacquiring
our heterogeneity specification might mask this relationship. the customer.
Targeting Decision Note that whereas this strategy is theoretically "optimal,"

it is uncertain whether the anticipated customer response
We assumed that before reacquisition, the firm knew three
and
characteristics of lapsed customers: (1) the associated profit will
duration ofactually
theiroccur. The theoretically
optimalwith
lapse, (2) the length of their first tenure reacquisition price is $.45
the firm, and less than the last price in
the prior relationship,
(3) the last price they paid. Logically, these are variables that and the optimal retention price is $.80
greater than the last
the firm can use to distinguish customer reacquisition targets
price. Price changes of this magnitude
rarely occurred in
who are likely to be profitable from targets who are not. Fix-
the data, in which the average reacquisi-
tion price was $.10 less than the last price paid before lapse,
ing all other predictors at their median values except the
and if the retention price was increased, the average increase
three factors that are known before reacquisition, we pre-
was $.17.
dicted the expected SLTV for various customer types, as is
shown in Table 4. The analysis shows Pricing
that relative
the to expected
costs. A class of heuristic approaches
that the firm can use is cost based: pricing at cost ($1.00) or
SLTV of a customer whose profile reflects the tenth per-
centile (from the data) of each of the below
three costfactors
(we used $.50)
is for
$.01.reacquisition.
In At the extreme,
the firm may price low enough so that the probability of
theory, this is a profitable customer and one the firm should
reacquisition is nearly one. To generate a reacquisition prob-
attempt to reacquire. However, this customer's expected
ability that is virtually one, we fixed the covariates at their
profit is sensitive to reacquisition costs. If the reacquisition
median levels and found that the model suggests a reacqui-
costs increase even slightly, the customer becomes unprof-
sition price of $.30. As the firm lowers the reacquisition
itable for the firm to pursue. Firms that target this type of
price through these three levels, respectively, the likelihood
customer must carefully manage their reacquisition invest-
of reacquiring a customer increases from .890 to .985 to
ment. This is important because it suggests that customer
.998. Given that our estimation results reveal that the second
winback should be a selective process and that not all lapsed
tenure duration is negatively correlated with the likelihood
customers should be pursued.
Table 4 also shows the expected SLTV of reacquiring
of thea average
customer, theand logic of considering these
reacquisition pricing strategies might be questioned. How-
modal customers of the firm. For the average and modal cus-
ever, it is important to acknowledge that duration and prof-
tomers, we fixed all covariates at their averages and modes,
its are not always strongly, or even positively, correlated
respectively. The predictions reveal that, on average, the
(Reinartz and Kumar 2000). Our analysis shows that lower-
firm targets attractive prospects among its lapsed customers
ing the reacquisition price to $.30 and then increasing the
and implements a profitable reacquisition strategy.
price above the last price in the first tenure results in the
Offer Decision highest SLTV. Table 6 highlights some of the calculations at
a reacquisition price of $.30. Again, note that the second
To assess how the offer decision affects expected
tenure duration and marketSLTV,
share are not maximized with
we provide highlights of a numerical simulation in Table 5.
this pricing strategy. As with the pure optimization
The values reported in Table 5 are the expected SLTVs of
approach, we find that implementing a heuristic approach in
customers for different reacquisition prices and the average
which prices increase over time maximizes profits, even
retention prices. In our numerical analysis, we fixed all vari-
though second tenure duration is not maximized.
ables that characterize the prior relationship at their median
Pricing relative to the last price paid before lapse.
values and assumed that the costs are fixed over time.10 This
Another class of heuristics that can guide pricing strategies
analysis addresses several important issues about pricing
for restart customers is pricing above, below, or equal to the
strategies for the reacquisition and retention of lapsed cus-
last price paid before lapse. In our data, the median value of
tomers. When conducting this type of analysis, it is impor-
the last price paid before lapse is $2.20. Firms may be
tant to acknowledge that firms may not focus on price opti-
tempted to choose $2.20 as the reacquisition price, because
mization but rather employ heuristics to set their prices. In
the likelihood of reacquiring a customer is .56. Our analysis
this section, we assess optimal pricing strategies in terms of
shows that reinstating a customer at the same price as the
SLTV and evaluate heuristics that may be observed in
last price paid before lapse ($2.20) and maintaining this
practice.
price is suboptimal from a profitability perspective. Profits
Optimal pricing strategy. If the optimal strategy is
can be improved if the firm follows one of two pricing
explored within the range of prices that the firm typically
strategies.
offers ($1.75 to $3.00), the result is that the firm should offer
First, if the firm's tendency is to offer restart customers
a reacquisition price of $1.75 and then raise the retention
prices that are lower than the last price paid before lapse,
price to $3.00. This strategy results in an acquisition likeli-
the most profitable approach is to offer a low reacquisition
hood of approximately .677. Note that this strategy does not
price and a low retention price. If we limit the price
maximize the length of the second tenure and thus does not decrease to be within two standard deviations of the mean
maximize the firm's long-term market share. However, the
price decrease, the most profitable offer is a reacquisition
increased margin from the retention price compensates for
price and a retention price of $1.80. This price is two stan-
the reduced duration of the second tenure.
dard deviations below the mean of the last price before
lapse. Region A in Table 5 shows some of the possible
pricing combinations that are consistent with this strategy.
loRelaxation of the time-invariant costs assumption is a trivial exercise
because our demand functions are not a function of costs. If we had At a fixed price of $1.80, for a prospective target, the reac-
individual-level cost information, we could make demand a functionquisition
of likelihood is .66 and the expected SLTV is
costs, and the results might vary. $27.49.
1
.869
91
13
$.65 $.68
$1.75 $1.75 $1.0 $2.43 $1.0

Modal
$32.89 $28.59 $29.4

Customer
Reacquird
5 6
.74
$.9 $.05
178
29
$2.8 $2.38 $1.0 $2.43 $1.0
$51.7 $40.8 $41.07

Averag
Customer
Reacquird
1 .968
7
$.68
$2. $1.75 365 $1.0 $7.62 $2.43 213 $1.0
$61.83 $59.83 $67.45
Covarites
ValueofA Observal
90thPercnil
2
.847 5
192 15
$2. $1.75 $1.0 $6.7 $2.43 $.68 $1.0 $45.09 $38.19 $4.86
Covarites
ValueofA Observal
75thPercnil
6 .53 4
17 127
$2. $2.0 $1.0 $4.35 $2.43 $.23 $1.0 $36.9 $20.43 $24.79
Covarites
ValueofA Observal
50thPercnil
Table4
.146
3
17
50
103
$2. $2.75 $1.0 $1.5 $2.43 $(.32) $1.0 $29.7 $4.38 $5.3
Covarites
ValueofA Observal
TARGEINCUSOMPFL 25thPercnil
.01
1
5
35 38
$2. $2.90 $1.0 $.0 $2.43 $1.0 $1.07 $.01 $.01

$(.47)
Covarites
ValueofA Observal
10thPercnil
Reacquistonpr Lastpricenolh Lapsedurtion() Tenur1(days) Probailtyfecqusn Reacquiston Expectdraquisonmg Averagtniopcd Pricehang PredictTnu2(ays) Periodtncs Expectdrniomagvqus ExpectdRaquisonVl
PredictTnu2(pos) Expectdrniomag Notes:Fracquindpfl,mb.Whxv
Profiles Reacquiston Retnio
$3.15 51.62 48.2 38. 37.49 36.17 34.87 3.60 32.98 32.6 31.6 29. 28.6 27. 26.71 25.70 24.7 23.78 2.87
$51.48
$3.05 49.51 46.52 37.9 36.1 34.8 3.60 32.8 31.79 31.20 30.4 28.93 27.84 26.80 25.78 24.81 23.87 2.97 2.10
$49.35
$2.95 47.38 4.59 35.92 34.70 3.49 32.1 31.5 30.58 30.2 28.91 27.85 26.81 25.81 24.8 23.91 23.01 2.14 21.3
$47.19
$2.85 45.2 42.6 34. 3.28 32.1 31.0 29.0 29.36 28. 27. 26.75 25.76 24.81 23.89 23.0 2.14 21.3 20.51
$4.9
19.7
$2.75 43.0 40.6 32.94 31.84 30.75 29.68 28.63 28.1 27.61 26.1 25.6 24.71 23.80 2.9 2.07 21.5 20.47
$42.78
19.62 18.90
$2.65 40.82 38.67 31.42 30.9 29.36 28.35 27.36 26.87 26.39 25.4 24.53 23.64 2.78 21.95 21.4 20.37
$40.5
19.47 18.76 18.0

$2.5 38.61 36. 29.0 28.9 27.96 27.01 26.08 25.6 25.16 24.8 23.41 2.57 21.75 20.97 20.1
$38.0
19.8 19.27 18.57 17.90 17.26

$2.45 36.8 34.65 28.37 27.46 26.5 25.6 24.79 24.36 23.9 23.10 2.8 21.49 20.73
$36.04
RetnioPrc
19.6 19.0 18.32 17.6 17.04 16.4

$2.35 34.15 32.6 26.84 25.9 25.1 24.3 23.50 23.10 2.70 21.9 21.6 20.41
$3.79
Averag
Table5
RegionB
EXPCTDSLV
19.34 18.67 18.0 17.38 16.7 16.8 15.6

$2.5 31.92 30.62 25.31 24.5 23.74 2.97 2.1 21.84 21.47 20.74 20.3
$31.5
19.47 18.0 18.5 17.52 16.9 16.3 15.76 15.20

$2.0 30.81 29.6 24.5 23.79 23.04 2.30 21.57 21. 20.85 20.15
$30.41
19.84 19.6 18.49 17.85 17.23 16.2 16.04 15.48

$2.15 31.5 30.28 25.0 24.8 23.51 2.75 2.0 21.63 21.6 20.54
$31.5
19.83 19.4 18.47 17.8 17.9 16.58 15.9

$2.05 32.94 31.5 26.01 25.0 24.39 23.59 2.80 2.4 2.03 21.8 20.5
$32.56
19.72 19.02 18.35 17.69 17.06 16.4

$1.95 34.21 32.69 26.8 26.03 25.18 24.35 23.5 23.1 2.73 21.95 21.8 20.4
$3.84
RegionA
19.50 18. 18.3 17.48 16.8

$1.85 35.29 3.67 27.6 26.74 25.87 25.01 24.16 23.74 23. 2.5 21.73 20.97 20.
$34.9
19.8 19.7 18.4 17.8 17.

$1.75 36.15 34. 28.1 27.31 26.41 25.3 24.6 24.3 23.81 2.98 2.17 21.38 20.6
$35.81
.50
1.0 1.75 1.85 1.95

2.05 2.15 2.0 2.5 2.35 2.45 2.5 2.65 2.75 2.85 2.95 3.05 3.15
Reacquiston Price $.30
Table 6
EXAMPLE OF SLTV CALCULATION
Profiles Variations in Retention Price
Reacquisition
Reacquisition price $ .30 $ .30 $ .30 $ .30 $ .30 $ .30 $ .30
Last price in prior relationship $ 2.20 $ 2.20 $ 2.20 $ 2.20 $ 2.20 $ 2.20 $ 2.20
Lapse duration (periods) 6 6 6 6 6 6 6
Tenure 1 (days) 117 117 117 117 117 117 117
Probability of reacquisition .998 .998 .998 .998 .998 .998 .998
Reacquisition costs $ 1.00 $ 1.00 $ 1.00 $ 1.00 $ 1.00 $ 1.00 $ 1.00
Expected reacquisition margin $ .20 $ .20 $ .20 $ .20 $ .20 $ .20 $ .20
Retention
Average retention price per period $ 1.75 $ 2.00 $2.25 $ 2.50 $ 2.75 $ 3.00 $ 3.25
Absolute price change $ (.45) $ (.20) $ .05 $ .30 $ .55 $ .80 $ 1.05
Predicted Tenure 2 (days) 178 142 118 123 128 131 134
Retention costs $ 1.00 $ 1.00 $1.00 $ 1.00 $ 1.00 $ 1.00 $ 1.00
Expected retention margin given reacquisition $35.68 $33.09 $31.39 $37.05 $42.66 $48.16 $53.48
Expected retention margin $35.61 $33.02 $31.33 $36.97 $42.58 $48.07 $53.37
Expected Reacquisition Value $35.81 $33.22 $31.53 $37.17 $42.78 $48.27 $53.57
There are two key factors that increase the expected Second, to hedge against the risk and potential costs of a
SLTV with the low-price strategy. The first is the signifi- restart customer lapsing again, firms may opt to increase
cantly higher reacquisition rate, relative to a price of $2.20. prices above their prior levels. Our analysis shows that the
Table 7 shows how the reacquisition rates vary with the firm can moderately reduce the reacquisition rate (see Table
reacquisition prices. The second factor is the impact of the 7) and still increase expected profits.11 To increase profits by
price decrease on length of the second tenure. Specifically, using this strategy, the firm must increase both the reacqui-
the length of Tenure 2 increases when the reacquisition price sition and the retention prices above the last price before
is less than $2.20. The increased duration compensates for lapse. Region B of Table 5 shows some of the price combi-
the lower retention margin that is received when the reten- nations for which expected SLTV exceeds the SLTV obtain-
tion price is less than $2.20. able when reacquisition and retention prices are both $2.20.
The key to a firm's successful implementation of a price
increase strategy is managing the trade-off between a lower
Table 7 acquisition rate (due to higher reacquisition prices) and
CUSTOMER REACQUISITION RATES increased margins (due to higher retention prices). Because
of this trade-off, there are limits on how high the firm can set
the reacquisition price relative to the retention price and
Reacquisition Price Reacquisition Probability
generate profits beyond those of a $2.20 fixed price. A reac-
$1.75 .677
1.80 .663
quisition price that is greater than $2.20 and significantly
1.85 .649 higher than the retention price will result in an expected
1.90 .635 SLTV below the expected SLTV attainable at a fixed price
1.95 .622 of $2.20.
2.00 .608
2.05 .595
Summary of Numerical Simulation
2.10 .582
2.15 .570 From these results, it might be concluded that the key to
2.20 .557
managing customer winback most profitably is successful
2.25 .545
2.30 .533
customer reacquisition. However, our targeting discussion
2.35 .521 has revealed that the firm may not want to price so as to
2.40 .510 reacquire all lapsed customers and instead may want to
2.45 .498
focus on customers with attractive profiles. Note that we
2.50 .487
2.55 .476
performed our simulation analysis by assuming a moder-
2.60 .466 ately attractive profile (i.e., at the median value of the prior
2.65 .455 relationship characteristics). When attractive customers
2.70 .445 have been recaptured, their behavior is such that the firm can
2.75 .435
recoup the losses from reacquisition by charging higher
2.80 .425
2.85 .416
prices. This approach maximizes margins but not long-term
2.90 .407
2.95 .398
3.00 .389
11 It is notable that when we computed expected profits for price increase
Notes: We calculated all probabilities at the median values of the prior strategies, the effect of price increase relative to prior price on relationship
tenure, lapse duration, and last prior price variables. duration was small (the effect of price decrease was irrelevant).
market share. To maximize share, firms must

prior history be concerned
and/or current response. In this context,
with second tenure duration. A betterdemandstrategy
becomes a functionforof marketing
maximiz- expenditures, and
thus actual profit-maximizing
ing share is to implement the heuristic of lowering pricing prices
strategies may vary.
relative to the last price before theAlso
lapse.
related to individual-level remarketing efforts, it would
be useful to address price endogeneity (Villas-Boas and
DISCUSSION AND LIMITATIONS Winer 1999), particularly for reacquisition pricing. Many
firms make reacquisition offers conditional on customers'
Early CRM advocates touted the benefits of having
responses to prior offers; however, in our application, only a
mature or long-life customers and consequently emphasized
single reacquisition offer is made. Price endogeneity might
retention. However, a 100% retention rate is seldom feasi-
be explored using a game theoretic analysis. In addition,
ble. Our analysis provides the additional insight that 100%
given a data set in which the firm makes many price changes
retention (or other high retention rates) is not always desir-
over time, dynamics of the baseline hazard could be studied.
able or profitable, particularly when it requires setting a low
Such an investigation would improve our understanding of
retention price. It is not worth it for firms to try to reestab-
the dynamics of customer winback and foster insights into
lish relationships with customers who are likely to lapse or
dynamic pricing strategies.
to defect rapidly. Furthermore, our research suggests that
lapsed customers who are more likely to be reacquired have REFERENCES
a shorter second tenure with the firm after they have been
Ainslie, Andrew and Peter E. Rossi (1998), "Similarities in Choi
reacquired. Behavior Across Product Categories," Marketing Science, 1
For the relationships that are attractive to reestablish, (2), 91-106.
firms need to employ profitable winback strategies that Amemiya, Takeshi (1985), Advanced Econometrics. Cambridg
maximize customers' SLTV. Our analysis shows that reac- MA: Harvard University Press.
quisition is the critical phase in the winback initiative. Bawa, Kapil (1990), "Modeling Inertia and Variety Seeking Ten
Firms' significant lowering of reacquisition prices to dencies in Brand Choice Behavior," Marketing Science, 9 (3
263-78.
increase the likelihood of reacquisition is an optimal strat-
Berger, Paul D. and Nada I. Nasr-Bechwati (2001), "The Alloca-
egy. To maximize profits, firms should increase prices when
tion of Promotion Budget to Maximize Customer Equity,"
the relationship has been reestablished. If market share is the
OMEGA: The International Journal of Management Science, 29
metric of interest, the firm should focus on maximizing
(1), 49-61.
duration of the second tenure. To do this, our analysis shows Blattberg, Robert C. and John Deighton (1996), "Manage Market-
that both reacquisition and retention prices should be low. ing by the Customer Equity Test," Harvard Business Review, 74
In addition, this research shows that lapsed consumers' (July-August), 136-44.
response to price varies at different phases of the relation- , Gary Getz, and Jacquelyn S. Thomas (2001), Customer
ship. Throughout the relationship, the customer is sensitive to Equity: Building and Managing Relationships as Valuable
Assets. Boston: Harvard Business School Press.
the absolute price. At the point of reacquisition, customers
respond negatively to price. However, customers who are Bolton, Ruth N. (1998), "A Dynamic Model of the Duration of the
Customer's Relationship with a Continuous Service Provider:
reacquired at higher prices have longer second tenures. Con-
The Role of Satisfaction," Marketing Science, 17 (1), 45-67.
sistent with this outcome, our research shows that there is a
, P.K. Kannan, and Matthew D. Bramlett (2000), "Implica-
negative correlation between a customer's preference for tions of Loyalty Program Membership and Service Experiences
reacquisition and their intrinsic "retainability." for Customer Retention and Value," Journal of the Academy of
Customers also demonstrate a dynamic response to rela- Marketing Science, 28 (1), 95-108.
tive prices. At the time of reacquisition, customers respond and Katherine N. Lemon (1999), "A Dynamic Model of
to deviations in price relative to the last price paid before Customers' Usage of Services: Usage as an Antecedent and Con-
lapse. However, after the relationship is reestablished, cus- sequence of Satisfaction," Journal of Marketing Research, 37
tomers respond in a way that reinforces their decision to (May), 171-86.
reenter the relationship. In other words, price increases have Boulding, William, Ajay Kalra, and Richard Staelin (1999), "The
Quality Double Whammy," Marketing Science, 18 (4), 463-84.
no affect on second tenure duration, and price decreases rel-
Casella, George and Edward I. George (1992), "Explaining the
ative to the prior price result in longer second tenures. This
Gibbs Sampler," The American Statistician, 46 (August),
is an important insight for firms because it shows that the 167-74.
last price paid during the first tenure plays a role throughout Chib, Siddhartha (1993), "Bayes Inference in the Tobit Censored
the winback process. Regression Model," Journal of Econometrics, 51 (1-2), 79-99.
Although this research highlights the importance of Cohen, Arthur R. (1960), "Attitudinal Consequences of Induced
dynamic pricing and the customer's prior history with the Discrepancies Between Cognitions and Behavior," Public Opin-
firm to a winback strategy, there are some limitations and ion Quarterly, 24 (2), 297-318.
areas that need further exploration. In particular, individual- Cox, David R. (1972), "Regression Models and Life Tables," Jour-
nal of the Royal Statistical Society, Series B (34), 187-200.
level remarketing efforts and marketing cost data might
Dwyer, F. Robert (1989), "Customer Lifetime Valuation to Support
affect our conclusions. A paucity of cost data prevents us
Marketing Decision Making," Journal of Direct Marketing, 11
from exploring this; however, our findings apply to firms (4), 6-13.
that have relatively standard costs across customers in their Eagly, Alice H. and Shelly Chaiken (1993), The Psychology of Atti-
winback programs. A firm that sends fixed-value coupons to tudes. Orlando, FL: Harcourt Brace Jovanovich.
all lapsed customers, such as HoneyBaked (as we mentioned Einhorn, Michael A. (1994), "Optimal Nonuniform Pricing with
in the first section), is an example of this. Ideally, winback Generalized Consumer Choice," Journal of Industrial Econom-
spending levels should vary on the basis of the customer's ics, 42 (March), 105-112.
- and Robin M. Hogarth (1981),

Operations Research "Behaviora
and Management Science, Jehoshua
ory: Process of Eliashberg and and
Judgment G.L. Lilien, Choice,"
eds. Amsterdam: North-Holland,
Annu
chology, 32,53-88. 349-408.
Festinger, L. (1957), AMarple, Mark and Michael
Theory Zimmerman
of (1999), "A Customer Di
Cognitive
IL: Row, Peterson. Retention Strategy," Mortgage Banking, 59 (11), 45-49.
- (1964), Conflict, Decision, and
Pfeifer, Phillip E. and Robert Dissonanc
L. Carraway (2000), "Modeling Cus-
Stanford University Press. tomer Relationships as Markov Chains," Journal of Interactive
Ganesh, Jaishankar, Marketing, Mark 14 (2), J.
43-55. Arnold, and K
(2000), "Understanding the Customer Base of Service Reichheld, Frederick F. (1996), The Loyalty Effect. Boston: Har-
Providers: An Examination of the Differences Between Switch- vard Business School Press.
ers and Stayers," Journal of Marketing, 64 (July), 65-87. and W. Earl Sasser Jr. (1990), "Zero Defections: Quality
Gelfand, Alan E. and Adrian F.M. Smith (1990), "Sampling-Based Comes to Services," Harvard Business Review, 68
Approaches to Calculating Marginal Densities," Journal of the (September-October), 105-111.
American Statistical Association, 85 (410), 398-409. Reinartz, Werner and V. Kumar (2000), "On the Profitability o
Griffin, Jill and Michael W. Lowenstein (2001), Customer Win- Long-Life Customers in a Noncontractual Setting: An Empirical
back: How to Recapture Lost Customers-And Keep Them Investigation and Implications for Marketing," Journal of Mar-
Loyal. San Francisco: Jossey-Bass. keting, 64 (October), 17-35.
Hansotia, Behram J. and Paul Wang (1997), "Analytical Chal- Rust, Roland T., Katherine N. Lemon, and Valarie A. Zeithaml
lenges in Customer Acquisition," Journal of Direct Marketing, (2001), "Driving Customer Equity: Linking Customer Lifetim
11 (2), 7-19. Value to Strategic Marketing Decisions," working paper, Depart
Heckman, James (1979), "Sample Selection Bias as a Specification ment of Marketing, University of Maryland.
Error," Econometrica, 47 (January), 153-61. , Valarie A. Zeithaml, and Katherine N. Lemon (2000), Dri-
Helsen, Kristiaan and David Schmittlein (1993), "Analyzing Dura- ving Customer Equity: How Customer Lifetime Value Is Reshap-
tion Times in Marketing: Evidence for the Effectiveness of Haz- ing Corporate Strategy. New York: The Free Press.
ard Rate Models," Marketing Science, 11 (4), 395-414. Samuelson, William and Richard Zeckhauser (1988), "Status Quo
Hughes, Arthur M. (1996), The Complete Database Marketer, Rev. Bias in Decision Making," Journal of Risk and Uncertainty, 1
ed. Chicago: Irwin Professional Publishing. (1), 7-59.
Jain, Dipak and Siddhartha S. Singh (2002), "Customer Lifetime Schmid, Jack (1998), "The 8 Commandments ... of Design," Cat-
Value Research in Marketing: A Review and Future Directions," alog Age, 15 (1), 57-64.
Journal of Interactive Marketing, 16 (2), 34-46. Schmittlein, David C., Donald G. Morrison, and Richard
Kahneman, Daniel and Amos Tversky (1979), "Prospect Theory: Colombo (1987), "Counting Your Customers: Who Are They
An Analysis of Decision Under Risk," Econometrica, 47 and What Will They Do Next?" Management Science, 33 (1),
(March), 263-91. 1-24.
Kalyanam, Kirthi (1996), "Pricing Decisions Under Demand - and Robert Peterson (1994), "Customer Base Analysis: An
Uncertainty: A Bayesian Mixture Model Approach," Marketing Industrial Purchase Process Application," Marketing Science, 13
Science, 15 (3), 207-221. (Winter), 41-67.
Kalyanaram, Gurumurthy and Russell S. Winer (1995), "Empirical Stauss, Bernd and Christian Friege (1999), "Regaining Service
Generalizations from Reference Price Research," Marketing Sci- Customers," Journal of Service Research, 1 (4), 347-61.
ence, 14 (3), G161-G169. Thaler, Richard (1985), "Mental Accounting and Consumer
Keaveney, Susan M. (1995), "Customer Switching Behavior in Choice," Marketing Science, 4 (Summer), 199-214.
Service Industries: An Exploratory Study," Journal of Market- Thomas, Jacquelyn S. (2001), "A Methodology for Linking Cus-
ing, 59 (2), 71-82. tomer Acquisition to Customer Retention," Journal of Marketing
Kiley, Kathleen (1996), "Back to Life," Catalog Age, 13 (2), Research, 38 (May), 262-68.
37-38. Varki, Sajeev and Mark Colgate (2001), "The Role of Price Per-
Krishnamurthi, Lakshman, Tridib Mazumdar, and S.P. Raj (1992), ceptions in an Integrated Model of Behavioral Intentions," Jour-
"Asymmetric Response to Price in Consumer Brand Choice and nal of Service Research, 3 (3), 232-40.
Purchase Quantity Decisions," Journal of Consumer Research,Villas-Boas, J. Miguel and Russell S. Winer (1999), "Endogeneity
19 (3), 387-400. in Brand Choice Models," Management Science, 45 (October),
Mahajan, Vijay, Eitan Muller, and Frank M. Bass (1993), "New 1324-38.
Product Diffusion Models," in Marketing, Vol. 5, Handbooks in

Recapturing Lost Customers

Diunggah oleh

Informasi Dokumen

Hak Cipta

Format Tersedia

Bagikan dokumen Ini

Bagikan atau Tanam Dokumen

Opsi Berbagi

Apakah menurut Anda dokumen ini bermanfaat?

Apakah konten ini tidak pantas?

Hak Cipta:

Format Tersedia

Recapturing Lost Customers

Diunggah oleh

Hak Cipta:

Format Tersedia

Recapturing Lost Customers

Author(s): Jacquelyn S. Thomas, Robert C. Blattberg and Edward J. Fox

American Marketing Association is collaborating with JSTOR to digitize, preserve and

For both academics and practitioners, the dominant focus of customer

Recapturing Lost Customers

Loyalty and retention have been the dominant themes Table 1

Journal of Marketing Research

(1999) define SLTV as the future value

Building on the framework introduced

having experienced another firm'sofservice exhibit

ship Dwyer (1989) outlines alternative

tional hazards model (Cox

2We tested whether there were systematic differences between customers

Total Sample Successfully Reacquired Sample

tionship with the firm before lapsing. Lapse duration meas-

which we suggest is a logical reference for the customer. We (.97)

these factors to have differential

Targeting Decision Note that whereas this strategy is theoretically "optimal,"

$1.75 $1.75 $1.0 $2.43 $1.0

$32.89 $28.59 $29.4

$2.8 $2.38 $1.0 $2.43 $1.0

$51.7 $40.8 $41.07

$2. $1.75 365 $1.0 $7.62 $2.43 213 $1.0

$61.83 $59.83 $67.45

$2. $2.90 $1.0 $.0 $2.43 $1.0 $1.07 $.01 $.01

Profiles Reacquiston Retnio

19.47 18.76 18.0

19.8 19.27 18.57 17.90 17.26

19.6 19.0 18.32 17.6 17.04 16.4

19.34 18.67 18.0 17.38 16.7 16.8 15.6

19.47 18.0 18.5 17.52 16.9 16.3 15.76 15.20

19.84 19.6 18.49 17.85 17.23 16.2 16.04 15.48

19.83 19.4 18.47 17.8 17.9 16.58 15.9

19.72 19.02 18.35 17.69 17.06 16.4

19.50 18. 18.3 17.48 16.8

19.8 19.7 18.4 17.8 17.

1.0 1.75 1.85 1.95

Reacquiston Price $.30

Profiles Variations in Retention Price

market share. To maximize share, firms must

- and Robin M. Hogarth (1981),

Anda mungkin juga menyukai