ArticlePDF Available

Developing Donor Relationships: The Role of the Breadth of Giving

Authors:

Abstract and Figures

This research proposes a mechanism to develop long-term donor relationships, a major challenge in the nonprofit industry. The authors propose a metric, donation variety, which captures a donor's breadth of donations with a given nonprofit organization, controlling for the distribution of donations to different initiatives. Using donation data spanning 20 years from a major U.S. public university, the authors find that improvements in donation variety increase the likelihood that the donor will make a subsequent donation, increase the donation amount, and reduce the sensitivity of donations to negative macroeconomic shocks. In the acquisition phase, most donors give to a single initiative, and these decisions are more influenced by a donor's intrinsic motivations. In contrast, as the donor-nonprofit organization relationship develops over time, nonprofit marketing efforts have a more significant influence on a donor's decision to give to multiple initiatives. Finally, the authors conduct a field study that validates the econometric analysis and provides causal evidence that marketing efforts by nonprofit organizations can encourage donors to spread donations across multiple initiatives.
Content may be subject to copyright.
Farnoosh Khodakarami, J. Andrew Petersen, & Rajkumar Venkatesan
Developing Donor Relationships:
The Role of the Breadth of Giving
This research proposes a mechanism to develop long-term donor relationships, a major challenge in the nonprofit
industry. The authors propose a metric, donation variety, which captures a donor’s breadth of donations with a
given nonprofit organization, controlling for the distribution of donations to different initiatives. Using donation data
spanning 20 years from a major U.S. public university, the authors find that improvements in donation variety
increase the likelihood that the donor will make a subsequent donation, increase the donation amount, and reduce
the sensitivity of donations to negative macroeconomic shocks. In the acquisition phase, most donors give to a
single initiative, and these decisions are more influenced by a donor’s intrinsic motivations. In contrast, as the
donor–nonprofit organization relationship develops over time, nonprofit marketing efforts have a more significant
influence on a donor’s decision to give to multiple initiatives. Finally, the authors conduct a field study that validates
the econometric analysis and provides causal evidence that marketing efforts by nonprofit organizations can
encourage donors to spread donations across multiple initiatives.
Keywords: donation variety, field study, cross-buying, donor relationship management
Online Supplement: http://dx.doi.org/10.1509/jm.14.0351
Farnoosh Khodakarami is a doctoral student, Kenan-Flagler Business
School, University of North Carolina at Chapel Hill (e-mail: Farnoosh_
khodakarami@kenan-flagler.unc.edu). J. Andrew Petersen is Associate
Professor of Marketing, Smeal College of Business Administration, Penn-
sylvania State University (e-mail: jap57@psu.edu). Rajkumar Venkatesan is
Bank of America Research Professor of Business Administration, Darden
School of Business, University of Virginia (e-mail: Venkatesanr@darden.
virginia.edu). The authors thank a major public university foundation for
providing the data used in the study and facilitating the field study. They
also thank Erin Mitchell for copy editing a previous version of the manu-
script and participants at the 2014 AMA Winter Educators’ Conference for
providing useful feedback on a previous version. This article is part of the
first author’s dissertation. Dhruv Grewal served as guest editor and
Praveen Kopalle served as area editor for this article.
© 2015, American Marketing Association
ISSN: 0022-2429 (print), 1547-7185 (electronic)
Journal of Marketing
Vol. 79 (July 2015), 77 –93
77
From 1999 to 2009, there has been a 59% growth in
the number of public charities and a 54% growth in
the number of foundations in the United States. How-
ever, the growth of donations has been relatively slow.
Adjusting for inflation, private charitable giving remained
steady at $290.89 billion between 2000 and 2010.1With
decreasing government support and the slow growth of
donations, competition for scarce resources among non-
profit organizations has become intense (Foster and Mein-
hard 2002; Sargeant and Woodliffe 2007; Thornton 2006).
Nonprofits must spend substantial resources on donor
acquisition activities, and approximately half the newly
acquired donors are lost after the first donation (Magson
1999; Masters 2000; Sargeant and Woodliffe 2007). For
example, in the higher education sector, the retention rate
for first-time donors was below 30% in 2012 and 2013
(Blackbaud Inc. 2014). Thus, building long-term relation-
ships with donors becomes critical. To address this issue,
many nonprofits keep donor-level information for donor
selection and optimal resource allocation (Kumar and
Petersen 2005; Lemon, White, and Winer 2002).
This raises an interesting question about whether strate-
gies to manage customer relationships in the for-profit sec-
tor can be easily translated as strategies to manage donors
in the nonprofit sector. Although the data collection process
is similar, donors’ motivations to give vary significantly
from customers’ motivations to purchase (Ariely and Nor-
ton 2009). However, there is scant research on donors’
repeat giving behavior to nonprofit organizations. Thus, it is
important to understand how acquisition and retention
strategies affect sustained giving.
An approach observed in the nonprofit industry to moti-
vate sustained giving is to give donors control over how
their gift is utilized. Indeed, many nonprofit organizations
now offer multiple causes or initiatives to which donations
can be directed. The American Red Cross and UNESCO
have been using this strategy for some time, and many orga-
nizations now run cause-related marketing campaigns with
multiple causes from which donors can choose. Providing
donors with the opportunity to direct their gifts toward spe-
cific causes or fundraising projects is driven by manage-
ment beliefs that such options can help increase donation
intentions and donor retention. Anecdotal evidence corrob-
orates this belief. Survey studies have shown that con-
sumers have greater interest in participating in cause-related
campaigns that allow donors to choose a charity to
1National Center for Charitable Statistics data (see http://nccs.
urban. org/statistics/index.cfm).
support.2Lab experiments have also shown that targeting
donations has a positive effect on gift amount (Li et al.
2013; Robinson, Irmak, and Jayachandran 2012). However,
there is no field-based empirical research exploring the dri-
vers and consequences of a donor’s decision to support
multiple initiatives at a nonprofit organization.
In our research, we empirically test whether donors who
support multiple initiatives are more valuable over time. We
measure a donor’s breadth of giving, controlling for the
distribution of donations to different initiatives, using an
individual- and time-varying metric we call “donation vari-
ety.” Donation variety is a single metric that captures the
relationship strength between the donor and the nonprofit
organization. Donation variety is defined as the weighted
sum of the share of each initiative a donor supports, in
which the weight for each initiative is the logarithm of the
share (similar to the measurement of entropy).
Our research explores three main research questions.
For the first two research questions, we want to better
understand what motivates donors to increase donation
variety. We expect that the drivers of donation variety are
likely to change over the course of the donor’s relationship
with the nonprofit organization (from acquisition to reten-
tion). Thus, our first research question is as follows:
RQ1: What factors motivate donors to increase the breadth
(donation variety) of their support of a nonprofit at dif-
ferent stages of their relationship life cycle?
In general, we find that a donor’s ties to the nonprofit
are more influential in driving donation variety at the acqui-
sition stage. However, as the donor–nonprofit relationship
develops (during the retention phase), marketing efforts
from the nonprofit organization become more influential at
driving donation variety. In the second part of the first study,
we aim to shed light on the impact of donation variety on a
donor’s future giving behavior. Specifically, our goal is to
empirically test how the distribution of past gifts among
multiple causes (donation variety) affects a donor’s future
giving. Thus, our second research question is as follows:
RQ2: What is the effect of the breadth of past gifts (donation
variety) to a nonprofit on future donor giving behavior?
We find that after we control for marketing efforts,
donor’s ties to the nonprofit, and amount of past gifts, the
higher the donation variety of the past donations, the more
likely the donor is to give again in the next fiscal year and,
conditional on a gift occurring, the larger the expected
value of the gift. Furthermore, we find empirical evidence
that donation variety also lowers a donor’s responsiveness
to declines in the macroeconomic condition.
In the second study, we explore the causal relationship
between marketing actions and donation variety through a
field study with the focal nonprofit organization. The goal
is to understand how changes in the nonprofit organiza-
78 / Journal of Marketing, July 2015
tion’s marketing communications can induce different lev-
els of donation variety. Thus, our third research question is
as follows:
RQ3: Can nonprofit organizations use targeted marketing
efforts to encourage donors to increase their donation
variety?
We find through our field study that marketing commu-
nications that encourage donating to an additional initiative
were able to significantly increase the probability of a
donor giving again in the future and, conditional on the gift,
giving more in total. Furthermore, we find that when donors
who already give to multiple initiatives (n = x) were
encouraged to give to another initiative (n = x + 1), they
were just as likely to give again as donors in the control
group, but the total amount of giving was higher.
Contribution
We believe that the results of our study provide several key
contributions to the nonprofit and marketing literature and
practice. Targeted (or directed) giving (i.e., allowing donors
to target their gift to specific causes) has recently received
attention in the literature. However, more research in this
area is needed. A review of studies (see Web Appendix A)
shows that prior research on the effectiveness of targeted
giving has focused on whether giving donors the choice to
direct gifts affects giving behavior. These studies were con-
ducted mostly in an experimental setting or as a small-scale
field study. In such settings, choice of causes is limited and
the long-term effect of targeted giving has not been investi-
gated. To the best of our knowledge, this is the first study
that uses donation data at the individual-donor level to
investigate giving behavior across initiatives over the donor
life cycle. We consider what motivates donors when they
can direct their gifts to multiple causes and how this deci-
sion to spread the gift across causes affects the donor’s
future giving to that nonprofit. We also take advantage of a
field study to provide causal evidence that nonprofit organi-
zations’ marketing efforts can encourage donors to give to
multiple causes and that the resulting donation variety
improves the donor–nonprofit relationship.
Study 1: Drivers and
Consequences of Donation Variety
From an exchange process perspective, the ongoing rela-
tionship process between the nonprofit organization and the
donor follows steps similar to the firm–customer exchange
process (Gupta and Zeithaml 2006). First, the nonprofit
organization communicates with potential donors to acquire
them using various marketing efforts. Second, the potential
donors make a set of simultaneous decisions on (1) whether
to donate; (2) conditional on donating, how much to donate;
and (3) how to allocate the donation to different initiatives.
Donors’ decisions on whether and how much to give and to
which donation options to allocate the gift are affected by
their ties to the nonprofit, their characteristics, and the non-
profit-initiated marketing efforts. After the nonprofit receives
2For example, 73% of respondents surveyed reported that they
would be more likely to participate in a cause-marketing program
if they were allowed to choose which charity is selected in-store
(Cause Marketing Forum 2010).
the donations in the given time period, the nonprofit– donor
exchange process then repeats dynamically over time as the
relationship between the two continues to develop.
Nonprofits that offer multiple donation options hope
that this strategy increases donation intentions and donor
retention. Giving donors choices increases their perception
of having a personal role in helping a nonprofit organiza-
tion. This facilitates role identity development by creating a
sense of “self-determination” and “ownership,” and enables
donors to contribute in personally meaningful ways (Grant
2012, p. 605). In experimental settings, giving donors con-
trol to choose between multiple initiatives offered by a
charity increased both donation amount and purchase inten-
tion for the associated products (Null 2011; Robinson,
Irmak, and Jayachandran 2012). Despite the prevalence of
multiple-cause donations, there is little empirical research
on whether giving to multiple causes increases repeat giv-
ing behavior and what factors may motivate a donor to dis-
tribute gifts across multiple causes (Bennett 2012; Ly and
Mason 2012).
Literature on variety in consumption of goods and ser-
vices has suggested that consumers may seek variety
because of satiation with current options or need for novelty
(Kahn 1995; McAlister and Pessemier 1982). It could also
be the case that as consumers’ financial well-being
increases, the options available to them also increase. Either
way, when donors make gifts, they sacrifice their own
physical consumption to get satisfaction from trading off
positive physical consumption for positive conceptual con-
sumption (Ariely and Norton 2009). This suggests that the
process and drivers of giving to charities are conceptually
different than those of consuming goods or services. We
draw from the literature on tie strength (Granovetter 1985;
Uzzi 1999) and social capital theory (Coleman 1988; Put-
nam 1995) to hypothesize the factors that explain drivers
and consequences of giving to multiple causes.
Hypothesis Development
Drivers of donation variety. Similar to customer–firm
relationships, both internal and external motivators may
affect a donor–nonprofit relationship. People are intrinsi-
cally motivated when they get inherent satisfaction and
enjoyment from their acts. Intrinsic motivators are thus “an
endogenous part of a person’s engagement in the activity”
(Amabile 1993, p. 189). Extrinsic motivations, in contrast,
come from an outside source and encourage people to
obtain a desired outcome (Amabile 1993; Ryan and Deci
2000). A donor’s engagement with a nonprofit may inter-
nally drive his or her decision to support multiple causes of
the nonprofit. For example, a donor who has personal inter-
est and experience with multiple (few) causes may feel
more (less) broadly tied to the nonprofit. In addition, non-
profit-initiated marketing efforts are an external motivator
for a donor to support multiple causes. Given that the con-
text of charitable giving is highly relationship based, we
investigate how these two key factors (intrinsic vs. extrinsic
motivation) affect donation variety during a donor’s life
cycle (acquisition vs. retention).
Developing Donor Relationships / 79
Intrinsic motivators. Donors often prefer to give to
charities that they can inherently relate to. Personal experi-
ence with a charity, whether a donor has benefited from a
cause in the past or believes that (s)he will benefit from it in
the future, motivates giving to that charity (Ariely and Nor-
ton 2009; Bekkers and Wiepking 2007; Bennett 2012; Null
2011; Robinson, Irmak, and Jayachandran 2012). Research
on philanthropy has also shown that many people prefer to
support those who are similar to themselves and help chari-
ties that are congruent with their identification (Bennett
2012; Sargeant and Woodliffe 2007). Identification in this
context refers to the extent to which donors feel connected
with a specific cause and how those causes align with their
internal fit (i.e., causes that are closer to their heart) (Aaker
and Akutsu 2009; Sirgy 1982). Thus, a donor who has per-
sonal experiences or identifies with multiple causes of the
nonprofit is more likely to support multiple causes without
external influences.
When donors make their first gift to a nonprofit organi-
zation, they may have limited knowledge about the various
donation options the organization offers. In addition, at the
early stage of the relationship, people have less confidence
in their evaluation of an organization’s offerings and might
feel uncertain about the way the nonprofit provides value to
the recipients of various causes (Bolton 1998; Swann and
Gill 1997; Verhoef, Franses, and Hoekstra 2002). Thus, at
the initial stage of relationship (acquisition), donors are
more likely to make donation choices on the basis of their
personal experiences with specific causes and degree of
identification with a cause (internal motivators). However,
as the relationship with the nonprofit organization evolves,
it is likely that donors will learn more about new initiatives
that are worth supporting that may be less related to their
intrinsic motivators. Thus, we hypothesize the following:
H1: The positive effect of a donor’s intrinsic motivators on
donation variety is stronger in the acquisition phase than
the retention phase.
Extrinsic motivators. People may engage in an act
because of an external source that motivates them to obtain
a desired outcome. Donors may be driven to donate by
external motivators—in this case, ongoing marketing com-
munications between the foundation and the donor. Like
customer loyalty, donor loyalty requires appropriate com-
munication and a relationship-building strategy. If donors
are “neglected and not asked for a second gift” (Sargeant
2001, p. 65), their contributions might decrease or even stop
after a first donation (Andreoni 2006; Bekkers and Wiep-
king 2007; Sargeant 2001). Nonprofits that give feedback to
donors by expressing appreciation and/or by responding to
donor concerns can influence donors’ attitudes toward the
organization and their willingness to engage in repeat giving
(Bekkers and Wiepking 2007; Kottasz 2004; Sargeant 2001).
Marketing efforts such as loyalty programs and direct
mailings have a positive effect on cross-buying of addi-
tional products and services (George, Kumar, and Grewal
2013; Kumar, George, and Pancras 2008; Li, Sun, and
Montgomery 2011; Verhoef, Franses, and Hoekstra 2001).
Likewise, nonprofits can leverage targeted marketing tech-
niques to develop relationships with donors, provide infor-
mation about the organization’s various causes and pro-
grams, and introduce donors to new donation opportunities.
Furthermore, as the donor–nonprofit relationship grows
over time, donors develop more trust toward the nonprofit
and its ability to provide value to the recipient of donation.
This trust influences them to give to more causes they are
aware of through marketing communications, even if they
do not have prior ties with the causes. As a result, we expect
that over time, the donor becomes more receptive to the
nonprofit’s solicitation requests to support additional causes
(Celsi and Olson 1988). Thus, we hypothesize the following:
H2: The positive effect of extrinsic motivators on donation
variety is stronger in the retention phase than the acquisi-
tion phase.
Consequences of Donation Variety
Main effect of donation variety. Charitable organiza-
tions can often enhance donation intentions by allowing
donors to choose which causes they want to support.
Donors who support multiple initiatives build a more
extended network with the nonprofit organization. In a
commercial context, social capital, measured as the strength
of the buyer–seller tie, has a significant positive impact on
purchase behavior (Frenzen and Davis 1990). Social capital
theory posits that a donor who is connected to multiple
causes of a nonprofit has stronger social ties to the non-
profit through his or her multiple connections (Putnam
1995). Stronger ties to the nonprofit through involvement
with its multiple causes can reinforce a donor’s contribution
to the nonprofit (Apinunmahakul and Devlin 2008; Brooks
2005; Brown and Ferris 2007; McAdam and Paulsen 1993).
Furthermore, people with more diverse and extended social
networks are more exposed to donation and volunteering
solicitations and may have a lower cost of giving (Brown
and Ferris 2007; Uzzi 1999). These donors with a more
extended network with the nonprofit may gain higher per-
ceived utility from their gift to the nonprofit and, in turn,
are expected to make more donations in future.
In addition, giving repeatedly to a single initiative might
lead to a decrease in marginal warm-glow utility derived
from the act of giving, which in turn decreases a donor’s
willingness to give in the future (Andreoni 1990). In an
experimental study, Null (2011, p. 455) shows that warm-
glow utility of giving can lead to “a love of variety” among
charities. In her experiment, most participants gave simulta-
neously to multiple charities even when charities were simi-
lar in mission and even when the benefits of the gift to the
recipient were set at different levels by varying matching
rate. We expect that giving to a variety of causes increases
the marginal utility and total satisfaction a donor experi-
ences from giving to a nonprofit organization. Satisfaction
and positive evaluation of an experience leads to repeated
engagement with that experience (Bennett 2012; Grant
2012). Thus, we hypothesize the following:
H3: All else being equal, people who have greater donation
variety give more in the future than people with a lesser
donation variety.
80 / Journal of Marketing, July 2015
Moderating effect of donation variety on economic
shocks. Economic shocks have a significant impact on a
donor’s ability and desire to make a donation. As the eco-
nomic condition declines and purchasing power decreases
across the population, people begin to cut down on unneces-
sary costs. For example, during the recent economic down-
turn in the United States, the percentage of consumers
involved in a nonprofit cause dropped from 60% to 53%
within two years.3The uncertainties of economic shocks have
a significant impact on a nonprofit’s ability to predict future
donor value. During an economic downturn, the cost of giving
increases, and all donors face a declining budget, requiring
some costs to be cut. In such situations, donors with weak ties
to the nonprofit are likely more willing to cease their support
to such a charity. In contrast, donors with strong ties to the
nonprofit have internalized a donor role into their identity
and likely feel more committed to sustain their support even
in an economic downturn (Brown and Ferris 2007; Sargeant
and Woodliffe 2007). Thus, we hypothesize the following:
H4: As the macroeconomic climate declines, people with
greater donation variety respond less negatively than
donors with lesser donation variety.
Data
Context. We chose a university foundation as the con-
text in which to empirically test our hypotheses. Donations
to educational organizations are of great importance. In the
United States, education organizations receive the second-
largest share of all charitable contributions.4Higher educa-
tion organizations are greatly dependent on contributions of
alumni donors. Furthermore, the ability to acquire and retain
alumni donors is a major challenge for higher education
organizations. Despite the economic growth after the Great
Recession, the declining acquisition rate of new alumni
donors is a threat to the survival of such organizations.
Another major challenge for educational institutions is a low
retention rate, especially because the majority of donors are
lost after their first gift (Blackbaud Inc. 2014). Thus, under-
standing a donor’s behavior and motivations for sustained
support is of great importance for educational organizations.
Furthermore, educational foundations allow donors to
give to multiple units. Within a college or university, donors
can choose to either donate to a general unrestricted fund to
be used at the foundation’s discretion or make a donation
targeted to specific departments, associations, scholarships,
memorials, and so on. To motivate our econometric model,
we run an exploratory analysis on the donation data of a
major public university foundation to determine whether
there is a difference between the ongoing giving behaviors
of donors who give to one initiative versus those who give
to multiple initiatives.
The focal university in this analysis has 44 specific
departments and associations to which donors can direct
3See http://www.causemarketingforum.com.
4In 2014, 16% of charitable contributions were directed at educa-
tional organizations, making this category the second-largest after
religious organizations (31%) (see http://www.givingusareports.
org/).
their donations—including, but not limited to, funds for
specific colleges, schools, groups, scholarships, and memo-
rials—as well as a general unrestricted fund. We focus only
on donors from the annual giving program and exclude
donors involved with or targeted for major planned and
capital gifts. Planned gifts and capital gifts are one-time
large gifts often given at or near the end of the relationship
(e.g., bequests) or dedicated to special projects (e.g., build-
ings). These major gifts are only made by a small segment
of donors. In contrast, the majority of donors participate in
the annual giving program. Annual gifts are smaller gifts
that require yearly decision making by donors. We believe
these recurring gifts offer a good representation of the
ongoing relationship between most donors and nonprofit
organizations.
Sampling. We use a stratified random sample of 500
donors in the annual giving program who made their first
gift in each of the years between fiscal year (FY)51993 and
FY 2003 and record each donor’s characteristics and dona-
tion behavior, aggregated at the annual level through the
end of FY 2012. This gives us a sample of 5,500 donors
(500 for each year), with an average of 15 years of data for
each donor. In our data set, each donation is, on average,
approximately $381, the total donations per donor over the
observation window are approximately $2,391, and the total
number of gifts in the observation window is approximately
11.2 per donor. Ninety-six percent of donors made their ini-
tial donation to a single initiative, leaving only approxi-
mately 4% who donated to more than one initiative in the
first year. By the end of FY 2012, 67% of donors gave to
multiple initiatives. This means after the initial gift, a major-
ity of donors gave to new initiatives in a subsequent year.
Variable Operationalization
Donation variety. The key variable we use in our model
is donation variety. We propose that, in addition to the num-
ber of different initiatives a donor supports, it is important
to measure the strength of the ties a donor builds with a
nonprofit by supporting multiple initiatives. Therefore, we
introduce the variable donation variety to differentiate
donation patterns.Here, we define the donation variety for
a given donor as the weighted sum of the share of each ini-
tiative a donor supports, in which the weight for each initia-
tive is the logarithm of its share. Thus, donation variety can
be represented as
At a given time t for donor i, Sijt is the share of the total
donation donor i made to initiative j until time t relative to the
total donation donor i made to the nonprofit organization
until that time. For a donor who gives exclusively to a single
initiative, donation variety is zero. The donation variety
increases as a donor gives to more initiatives and gives evenly
=
=
(1) do nati on v a r i et y S ln S .
i, t ijt ijt
j1
m
Developing Donor Relationships / 81
across many different initiatives. Donation variety is a cumu-
lative measure of giving behavior that takes into account the
number of initiatives a person selects for donation, control-
ling for the distribution of donations to different initiatives.
The donation variety index is similar to the entropy
measure that is used to measure the level of diversity in a
company’s business portfolio as well as in a personal
investment portfolio (e.g., Chatterjee and Blocher 1992;
Hoskisson et al. 1993; Palepu 1985; Palich, Cardinal, and
Miller 2000; Woerheide and Persson 1993). Marketing
scholars have also applied the entropy measure to model
customer brand preferences using the market share of vari-
ous brands (Herniter 1976; Kapur, Bector, and Kumar
1984). Similarly, Simonson and Winer (1992) apply a vari-
ety score on the basis of the overall share of items pur-
chased by a household to account for choice variety. They
also use the sum of the squares of the brands’ shares to mea-
sure “taste concentration” (homogeneity) in a household’s
purchase portfolio. Furthermore, Kahn (1995) recommends
using the entropy measure to account for variety in a con-
sumer’s purchase portfolio. Kahn (1995, p. 145) argues that
“even if the number of items included in the choice set is
constant, there is more variety in the choice history if the
choice shares of the items included are equal (maximum
entropy) than if one alternative dominates (low entropy).”
Thus, we apply a similar measure to gauge the breadth of a
donor’s giving to multiple donation options, controlling for
the distribution of donations to different initiatives.
Intrinsic motivators. To test H1, we need a variable (or set
of variables) that represents the extent to which a donor has
an intrinsic motivation to donate to multiple initiatives. We
expect that a donor’s strength of ties to various initiatives of
a nonprofit can act as a good indicator of broad connected-
ness to the university. Research has shown that shared
demographic characteristics are a good proxy for measuring
the strength of a tie between two people (Reagans 2005).
Thus, we expect that variables describing cases in which
alumni likely have multiple connections with different
initiatives/departments of a university are good indicators of
broader intrinsic ties to the university. We use two variables
as indicators of the breadth of ties across initiatives/
departments: (1) the number of degrees the alumnus/ alumna
(“alum” hereinafter) earned from the university and (2) hav-
ing a spouse who also graduated from the university. We
expect that these two variables are likely indicators of
alumni who have had or shared broader experiences across
the university. For example, donors with multiple degrees
have often had different experiences across programs (e.g.,
bachelor’s degree, master’s degree, doctorate) or across
schools (e.g., Arts and Sciences, Business, Medicine). We
find that the alumni of the university have, on average,
approximately 1.2 degrees from the university, and approxi-
mately 30% are married to other alumni from the university.
Extrinsic motivators. To test H2, we need a variable (or
set of variables) that represents the extent to which a donor
has been exposed to extrinsic motivators to donate to multi-
ple initiatives. Much of the external motivation for alumni
to make donations comes from the marketing efforts initi-
5The fiscal year for this nonprofit organization begins on July 1
and ends on June 30.
ated by the university foundation. These marketing efforts
include personal visits, phonathon calls, invitations to
events, and direct mail/e-mail solicitations. We expect that
different types of marketing efforts are likely to have vary-
ing effects on a donor’s decision to make a gift. For reasons
of parsimony, we choose to group phonathon calls and
direct mail/e-mail as impersonal marketing efforts because
the message content used for these marketing efforts is
homogeneous across the donor population and less interac-
tive. Furthermore, we group personal visits and invitations to
events as personal marketing efforts because their content is
richer and donor specific (Venkatesan and Kumar 2004).
In our sample, on average, alumni receive approximately
five times as many impersonal (.49) versus personal (.09)
marketing communications from the university per year. This
is common, given the much higher cost of personal market-
ing communications. We note that the focal educational
foundation follows a similar process when initiating any type
of marketing communications to alumni. All alumni typically
receive a communication from both the university and the
general alumni association beginning just before graduation.
All communications, including informational newsletters,
contain an appeal letter asking for a gift. These communica-
tions usually continue for several years postgraduation,
regardless of whether a gift is given. When an initial gift is
given, the alum can choose whether the gift is given to a spe-
cific initiative, to multiple initiatives, or to an unrestricted
fund.6After a gift is given, the initiative(s) supported often
communicates regularly with the donor to ask for subse-
quent gifts to the same initiative(s) previously supported.
Gift-giving behavior. To test H3, we need a variable that
represents the outcome of the gift-giving process. In this
case, because we aggregate the data on an annual basis by
fiscal year, we define gift giving as the total amount of
donations by a given donor in a given year. We find that,
conditional on giving, the average gift amount is approxi-
mately $381. We also control for the effect of previous giv-
ing by including lag of donation amount in our analysis.
Macroeconomic condition. To test H4, we need a variable
that represents the macroeconomic conditions that the
alumni are facing. Similar to many other studies in market-
ing, we measure the overall macroeconomic condition as
the cyclical component of gross domestic product data after
we apply a Hodrick–Prescott (1997) filter to remove the
long-term trend component of gross domestic product.
Thus, the average macroeconomic condition in the sample
is zero, and any deviation above (below) zero suggests a
positive (negative) macroeconomic climate.
Control variables. In addition, we include several control
variables in our model. We include variables meant to capture
the financial strength or capacity of giving of a given alum.
First, we include the average household income and aver-
age charitable contribution at the zip code level based on
the donor’s residence. Second, we include time since gradu-
ation because the longer it has been since an alum gradu-
82 / Journal of Marketing, July 2015
ated, the greater the likelihood that his or her earning power
and assets are higher. Finally, we control for the amount of
the previous gift. We also include some demographic
variables to account for observed heterogeneity, including
gender and location (in-state vs. out-of-state). We provide a
list of the variables, descriptive statistics, and description of
the operationalization of each variable in Table 1.
Model Development
Model-free evidence. For the first step of our model
development, we run an exploratory analysis to provide
model-free evidence of how donation variety might be
related to donor value. We use the same sample of 5,500
donors from the focal university of this study. We split the
sample into two groups: (1) donors who began at some
point to give to multiple initiatives (n = 3,720) and (2)
donors who only gave to a single initiative during the entire
observation window (n = 1,780) through FY 2012. For the
first group (labeled “Multiple Initiatives”), we split the data
into the time period when the donors only gave to a single
initiative and the time period after they began giving to
multiple initiatives. We then determine the average gift
amount for donors before ($149.34) and after ($500.16)
giving to a second initiative. For the second group (labeled
“Single Initiative”), we split the data into two time periods
as well. Here, we treat the first five years as early gifts (this
is the average time that a donor in the first group waits
before donating to multiple initiatives) and any time after
five years as later gifts. We then determine the average gift
amount for donors in the early ($114.03) and later
($194.20) time periods (see Figure 1).
First, we observe that in both cases, the average gifts in
the early time period are lower than the average gifts in the
later time periods, suggesting that, over time, there is an
increase in average giving for all donors. However, we find
that the increase in average gift amount for donors who give
to multiple initiatives is significantly larger than the
increase for the single-initiative donors ($270.65; p< .01).
This provides some evidence that increases in donation
variety lead to increases in giving amounts. However, it is
also important to quantify the benefits of donation variety
by controlling for as many other factors as possible. To do
so, we build an econometric model that can identify the
antecedents and consequences of donation variety.
Endogeneity of marketing. To empirically test our
hypotheses, we must first control for the endogeneity of
marketing efforts because nonprofit organizations usually
do not send solicitations at random. They commonly focus
their fund-raising efforts on donors and prospects that have
a higher likelihood of donation. To address the issue of
endogeneity with marketing efforts, we use instrumental
variable models for both personal and impersonal market-
ing efforts. Then, we use a control function approach
(Petrin and Train 2010) to include marketing efforts (both
personal and impersonal) in the next step of analysis, along
with the computed error from the instrumental variable
equations. Web Appendix B presents a detailed discussion
on instrumental variable models and estimations.
6We note that an unrestricted gift is also an initiative a donor
can support, similar to a gift to a specific department.
Methodology. To test H1and H2, we need a model that
can help us understand what factors motivate a donor to
increase donation variety over his or her life cycle. To
accommodate both the acquisition and retention stages of
donor relationship in the model, we use a binary variable
(Firsti, t) to distinguish initial and subsequent gifts, where
Firsti, t = 1 when it is the initial gift of donor i at time t and
Firsti, t = 0 when it is a subsequent donation by a given
donor i. For Firsti, t = 1, the interaction of Firsti, t with donor
intrinsic motivators and nonprofit-initiated marketing
efforts help us identify whether the impact of donors’ intrin-
sic and extrinsic motivators are strengthened or weakened
across acquisition and retention stages.
The focal variable of interest in this model is donation
variety, as measured in Equation 1. Donation variety is cen-
sored at zero for donors who exclusively support a single
Developing Donor Relationships / 83
TABLE 1
Variable Operationalizations and Descriptive Statistics
Variable M SD Operationalization
Single Versus Multiple Giving Behavior
Multiple giving 4.40% N.A. Percentage of the sample who gave to multiple initiatives in
their first gift
67.60% N.A. Percentage of the sample who gave to multiple initiatives by
the end of FY 2012
Time until multiple giving 4.65 3.71 For a donor who begins by giving to single initiative (95.6%),
the average time (in years) from first donation until the donor
begins giving to a second initiative
Donation variety .42 .40 The average donation variety of a given donor at the end of
observation period (FY 2012) as measured by Equation 1
Intrinsic Motivators
Spouse at the university 32.10% N.A. Percentage of donors with a spouse who also went to the
same university
Number of degrees 1.20 .50 Average number of degrees a donor received from the
university
Extrinsic Motivatorsa
Personal marketing .09 .35 Average number of times each donor was invited to fund-
raising events and received personal visits from the
university foundation each year
Impersonal marketing .49 .36 Average number of mailings and phone calls each donor
received each year
Control Variables
Gender 59.30% N.A. Percentage of female donors
In-state resident 60.80% N.A. Percentage of donors residing in the same state as the focal
university
Time since graduation 17.65 4.44 Average number of years since graduation at the end of
observation period (FY 2012)
Average AGI $73,076 $48,243 Average AGI at the zip code level
Average charitable contribution 3.20% 2.60% Average percentage of the AGI donated at the zip code level
Exchange Variables
Average donation amount $381 $4,132 Average amount of donation per donor per year ($)
Total donation amount $2,391 $17,822 Average total amount of donation ($) made by each donor
over the observation period
Total number of gifts 11.2 17.2 Average total number of gifts made by each donor over the
observation period
aWe use actual personal and impersonal marketing costs in the model rather than just the number of touches. However, we cannot provide the
descriptive statistics (mean and standard deviation) of marketing costs at the request of the university foundation. We can note that the cost
of each personal marketing touch is significantly larger (>10 times) than that of each impersonal marketing touch.
Notes: N.A. = not applicable; AGI = adjusted gross income.
FIGURE 1
Model-Free Evidence: Before and After Giving to
Multiple Initiatives
Early Gift
$600
$500
$400
$300
$200
$100
$0
Multiple initiatives
Single initiative
Later Gift
$149.34
$114.03
$500.16
$194.20
initiative. Furthermore, there is likely high inertia in the
measure of donation variety because it is measured as a
cumulative index. To handle any potential serial correlation,
we estimate a dynamic panel model with unobserved
heterogeneity, with donation variety as the dependent
variable and the lag of donation variety as an independent
variable. As noted, we do not observe a positive donation
variety for all donors. Indeed, most donors do not give to
multiple initiatives in their initial gift (>95% give to one
initiative on their first gift), and some donors (approxi-
mately 33%) never give to multiple initiatives. Therefore,
our model must handle the partial observability of donation
variety (i.e., it is censored at zero). To do so, we estimate
the following panel data model:
where Varietyi, t is the donation variety for donor i up to
time t as computed from Equation 1 (for t = 1, 2, …, T); Xi,t
contains individual and time-varying explanatory variables
such as intrinsic and extrinsic motivators, lagged donation
amount, lagged donation variety, first-year dummy, and the
interaction between first-year dummy and both intrinsic and
extrinsic motivators; and ciand mi,t are the individual-specific
unobserved effect and normally distributed idiosyncratic
error term, respectively. The limited dependent variable
model is generally fitted using Tobit specification. One of the
key limitations of Tobit specification is that the underlying
process driving the probability of observing a positive value
[P(Varietyi, t > 0
|
Xi, t )] and the actual value [E(Varietyi, t)
|
Varietyi,t > 0, Xi, t )] are both driven by the same underlying
process. We adopt a general class of model specification
proposed by Cragg (1971) that integrates the probit and
truncated normal models:
Unlike the Tobit model specification, the Cragg model
permits different explanatory variables for each decision
(i.e., Z ≠ X), and even when Z = X, the underlying process
driving the two decisions could be different. Note that the
Tobit model is a special form of Cragg model in which Z =
X and g= b/d. Thus, the Cragg model is a more flexible
alternative to the Tobit model, and it also has the benefit of
enhanced efficiency due to the simultaneous estimation of
both stages. We estimate the Cragg model specification
using a maximum likelihood–based Craggit procedure in
{}
()
() ()
()
=−Φ γ Φ γπσ
Φβ
σ
=>
=
()
()
()
=
−−β
σ
=
(3) f d ,Variety Z , X
1Z Z2 e
X,
where d
1ifVariety 0
0ifVariety 0.
i, t i, t i, t i, t
i, t
1d 0
i, t
1
21
Variety X
2
i, t
1d 1
i, t
i, t
i, t
i, t
i, t i, t
2
2
i, t
++µ
=>
=
(2) Variety X c
Variety
Variety if Variety 0
0ifVariety0
,
i, t
*i, t 1 i i, t
i, t
i, t
*i, t
*
i, t
*
84 / Journal of Marketing, July 2015
STATA 13.1 (Burke 2009). We use clustered standard errors
to account for panel-specific heteroskedasticity.
Similarly, to test H3and H4, we need a model that can
accommodate partial observability of the dependent variable,
gift amount (i.e., we only observe a value for gift amount
when a donation occurred). The model takes the following
format:
where ln(Gifti,t) is the log of the gift amount given by donor
i at time t (for t = 1, 2, …, T); Xi, t contains individual and
time-varying explanatory variables such as lagged donation
amount, lagged donation variety, marketing efforts, and
donor’s intrinsic motivators; and ciand mi,t are the individual-
specific unobserved effect and normally distributed idio-
syncratic error term, respectively. We use the same maxi-
mum likelihood–based Craggit procedure (Burke 2009) to
estimate the model:
Although we could use a different set of X variables for the
selection and conditional regression models for Equations 3
and 5 given the flexibility of the Cragg model, we chose to
use the same X variables across the selection and condi-
tional regression models for each equation.
Results
Table 2 presents the results of estimation for Equations 3
and 5. Both models have good fit, and the majority of coef-
ficients are statistically significant.7However, interpreta-
tion of coefficients in the Cragg model can be precarious
because the effect of an independent variable can vary in
magnitude as well as direction across the probit model and
{}
()
() ()
()
()
()
=−Φ γ Φ γπσ
Φβ
σ
=>
=
()
()
()
=
−−β
σ
=
(5) f(d , Ln Gift Z , X )
1Z Z2 e
X,
where d
1ifLnGift 0
0ifLnGift 0
.
i, t i, t i, t i, t
i,t
1d0
i,t
1
21
Ln Gift X
2
i,t
1d1
i, t
i, t
i, t
i, t
i,t i, t
2
2
i,t
()
() () ()
()
++µ
=>
=
(4) Ln Gift X c
Ln Gift
Ln Gift if Ln Gift 0
0if Ln Gift 0
,
i, t
*
i, t 1 i i, t
i, t
i, t
*
i, t
*
i, t
*
7We also estimated the two models in Equations 3 and 5 using a
Type I Tobit specification with the same X variables we used
when estimating Equations 3 and 5. We find in both cases that the
coefficients are similar in direction and magnitude. However, we
find that the in-sample and out-of-sample predictions are better for
the model estimated using the Cragg model. Here, we conducted
the out-of-sample prediction by using the data from the main study
for calibration and predicting the outcomes in the holdout time
period: FY 2013. We believe this is the case because the Cragg
model allows for the flexibility that the coefficients of the selec-
tion and conditional regression models can be different, even
when the X variables across the two models are the same.
the truncated normal regression. The problem of interpreta-
tion of coefficients associated with interaction effects in
nonlinear models is even more complicated (Ai and Norton
2003). Thus, to empirically test the hypotheses, we calcu-
late the unconditional expected value of the dependent
variables at each time t. We can then compare the mean of
the predicted values across different groups.
To interpret the effect of intrinsic and extrinsic motiva-
tors on driving donation variety, we estimate the expected
donation variety at different levels of intrinsic and extrinsic
variables for both acquisition (first gift) and retention
(repeat gift) stages of the donor–nonprofit relationship. We
present the estimations in Figure 2, Panels A–D. Panels A
and B present the interaction effect of intrinsic motivators
and first donation indicator. These figures show that a
donor’s multiple ties to the nonprofit are more influential
on driving donation variety at the initial stage of the donor–
nonprofit relationship than the retention stage. For the first
gift, donors who have multiple degrees from the university
(Panel A) and donors whose spouses are also university
alumni (Panel B) are more likely to have greater donation
Developing Donor Relationships / 85
variety than donors who have fewer ties to the university.
However, we also note that in the retention phase, the
donor’s intrinsic motivators have no additional effect on
donation variety, confirming H1.
In Figure 2, Panels C and D, we present the interaction
effect of marketing efforts and first donation indicator. These
figures show that nonprofit marketing efforts are more
influential in driving donation variety at the retention stage
than the acquisition stage. For the first gift, marketing seems
to play a small role in encouraging donors to give to increase
donation variety. However, for repeat gifts, donors who
receive more marketing efforts from the university foundation
have greater expected donation variety, on average, than
donors who receive fewer marketing efforts, confirming H2.
To estimate the direct and moderating effect of donation
variety on gift amount, we predict the expected gift amount
given different levels of donation variety. For donors who
give to only one initiative, donation variety is zero (no vari-
ety). For donors who give to multiple initiatives, we use
median split to group them as donors with either a low (0 <
donation variety < .64) or a high (donation variety .64)
TABLE 2
Results for the Drivers and Consequences of Donation Variety
DV = Varietyi, t DV = ln(Gifti, t)
Selection Main Selection Main
Variables Coefficient (SE) Coefficient (SE) Coefficient (SE) Coefficient (SE)
Intercept –2.223 (.027)** .177 (.005)** –.032 (.025) 2.412 (.046)**
ln(Gifti, t – 1) .102 (.003)** .012 (.0004)** .204 (.003)** .252 (.005)**
Varietyi, t – 1 2.870 (.021)** .802 (.003)** .071 (.021)** .157 (.032)**
Firsti, t
.159 (.078)* .150 (.037)** — —
Extrinsic
ImpMktgCosti, t .012 (.003)** .011 (.001)** .044 (.009)** .130 (.020)**
PerMktgCosti, t .096 (.005)** .002 (.0002)** .006 (.002)** .032 (.007)**
Econt
6.793 (.397)** .118 (.054)* 1.524 (.377)** 4.753 (.503)**
Intrinsic
Degreesi, t .092 (.013)** .010 (.002)** .065 (.013)** .087 (.025)**
Spousei .256 (.013)** .008 (.002)** .093 (.014)** .005 (.027)
Interaction Effects
Varietyi, t – 1 ¥ImpMktgCosti,t
.043 (.004)** .003 (.001)**
Varietyi, t – 1 ¥PerMktgCosti,t .001 (.0004)* .002 (.0005)**
Varietyi, t – 1 ¥Econt –4.143 (.774)** –.606 (.101)**
Firsti, t ¥ImpMktgCosti,t
–.015 (.003)** –.001 (.0004)* — —
Firsti, t ¥PerMktgCosti,t –.002 (.001)* –.003 (.0008)** — —
Firsti, t ¥Degreesit .106 (.053)* .006 (.003)* — —
Firsti, t ¥Spousei .065 (.013)** .001 (.0002)** — —
Control Variables
Genderi
.044 (.012)** .016 (.002)** .039 (.013)** .439 (.026)**
Locationi
.036 (.012)** .015 (.002)** .054 (.014)** .054 (.025)*
Time Since Graduationi, t
.089 (.022)** .002 (.0003)** –.071 (.002)** .033 (.003)**
AvgAGIi, t
.001 (.0001)** .0003 (.0002) .0002 (.0001)* .003 (.0003)**
AvgGivingi, t .655 (.228)** .039 (.038) .064 (.024)** .643 (.259)*
Control Function Variables
ImpMktgErrori,t –.012 (.012) .012 (.001)** –.058 (.009)** –.146 (.020)**
PerMktgErrori,t –.093 (.005)** .002 (.0002)** .009 (.002)** –.028 (.007)**
Model Fit
Log-pseudo-likelihood –13,723.53 –107,989.78
*p< .05.
**p< .01.
Notes: AGI = adjusted gross income.
level of donation variety. To test the direct effect of dona-
tion variety on gift amount, we compare the difference in
average expected gift amount across donors with low and
high levels of donation variety. The average expected gift
size is $115.62 for donors who give to only one initiative.
The average expected gift amount for donors with low and
high levels of donation variety is $394.76 and $795.80,
respectively. This indicates a positive effect of donation
variety on gift amount, confirming H3.
Next, we test the interaction effect of donation variety
and macroeconomic condition. Economic conditions below
and above the trend line derived from the Hodrick–Prescott
filter are classified as negative and positive economic condi-
tions, respectively. As Figure 3, Panel A, shows, as economic
condition declines, donors decrease their financial support
for the university foundation.8However, the decrease in
financial support is much greater for donors who give to a
single initiative than donors who give to multiple initia-
tives, confirming H4.
Although we do not formally hypothesize the moderat-
ing effect of donation variety on marketing efforts, we still
include the interaction between the two marketing variables
86 / Journal of Marketing, July 2015
and donation variety to control for the potential impact of
the university foundation’s marketing efforts. We want to
determine whether there is some evidence that donors with
different levels of donation variety tend to respond differ-
ently to both impersonal and personal marketing efforts.
Figure 3, Panel B, shows the interaction of donation variety
and impersonal marketing costs. It illustrates that donors
who receive more impersonal marketing efforts make larger
gifts compared with donors who receive fewer impersonal
marketing efforts. The difference in average gift size for the
two groups becomes larger as donors give to a more varied
portfolio of initiatives. Figure 3, Panel C, shows the interac-
tion effect of donation variety and personal marketing
effort. Given that the majority of donors receive no personal
marketing effort, we split them into two groups: donors
who receive no personal marketing and donors who receive
some personal marketing. Figure 3, Panel C, shows that
personal marketing effort is mostly targeted at high-value
donors. However, the difference in average gift size for
donors who receive some marketing and donors who
receive no marketing becomes much larger for donors with
higher levels of donation variety.
Comparing Donation Variety Cross-Donation
We want to determine whether our proposed measure of
donation variety (see Equation 1) outperforms a measure
commonly used to represent the breadth of a customer or
FIGURE 2
Moderating Impact of Intrinsic and Extrinsic Motivators on Donation Variety
A: Degrees and Donor Life Cycle B: Spouse and Donor Life Cycle
C: Impersonal Marketing and Donor Life Cycle D: Personal Marketing and Donor Life Cycle
First Gift Repeat Gift
.35
.30
.25
.20
.15
.10
.05
0.009
.061
.311 .319
One degree
Multiple degrees
First Gift Repeat Gift
.35
.30
.25
.20
.15
.10
.05
0.015
.069
.315 .319
Non-alum spouse
Alum spouse
First Gift Repeat Gift
.35
.30
.25
.20
.15
.10
.05
0.031 .035
.331
.294
Low impersonal marketing
High impersonal marketing
First Gift Repeat Gift
.60
.50
.40
.30
.20
.10
0.030 .033
.293
.515
No personal marketing
Some personal marketing
8We note that the focal university does not alter its marketing
budget in accordance with changes in macroeconomic condition
(i.e., the budget does not increase [decrease] in good [bad] macro-
economic conditions).
donor’s relationship with an organization (e.g., cross-buy).
In this case, we measure cross-donation as the total number
of initiatives supported by donor i up to time t – 1. First, we
measured the correlation of the donation variety with cross-
donation. The correlation of donation variety with number
of initiatives supported (cross-donation) is .762.
This finding suggests that donation variety and cross-
donation seem to represent the same construct measuring
the breadth in relationship a donor has with a nonprofit
organization. However, when we only compare the cases in
which the donation variety is positive (i.e., donors giving to
Developing Donor Relationships / 87
multiple initiatives), the correlation with cross-donation is
.604. This seems to suggest that, as the distribution of sup-
port to each initiative varies, the measure of donation vari-
ety we use in this study begins to capture subtle differences
from the measure of cross-donation.
We also want to investigate the economic significance
of the difference in the results when we use the measure
donation variety versus cross-donation. To do so, we esti-
mate the models from Equations 3 and 5 after substituting
cross-donation for donation variety (see the results of the
estimation in the Appendix). In general, we find cross-
donation is significant in the probit model and main model
(the same as donation variety). To test the in-sample model
fit differences, we compare the mean absolute deviation
(MAD) and mean absolute percentage error (MAPE) of the
expected donation value for all donors in all time periods in
the observation window. We find that the in-sample MAD
(MAPE) for the donation variety model is $60.11 (17.27%)
and for the cross-donation model is $68.81 (18.01%), which
suggests that the model with a single measure of donation
variety captures more of the variation in expected donation
amounts than the model with cross-donation. Next, we test
our model’s out-of-sample fit. To do so, we use the coeffi-
cients from the models to select the most valuable donors
on the basis of the predicted expected donation in FY 2013,
given that the original observation window of the data extends
through FY 2012. We then determine whether the foundation
is better off selecting the top percentiles of donors (10%,
15%, and 25%) on the basis of their expected donation from
the donation variety or cross-donation (see Table 3).
Table 3 shows that regardless of whether the foundation
selects the top 10%, 15%, or 25% of donors on the basis of
the expected giving amount predicted by the two different
models, the model using donation variety helps the univer-
sity foundation select donors with higher giving amounts in
FY 2013 by 6.4%, 3.9%, and 2.7%, respectively. These
results suggest that using the proposed measure of donation
variety is more valuable than cross-donation because dona-
tion variety can also control for the distribution of dona-
tions to different initiatives.
Study 2: Field Study
The evidence from Study 1 shows that there is a strong cor-
relation between donation variety and expected future giv-
ing. To test whether there is a causal relationship between
donation variety and expected future giving, we propose a
FIGURE 3
Moderating Impact of Donation Variety on
Expected Giving
A: Economic Condition and Donation Variety
B: Impersonal Marketing and Donation Variety
No Variety Low Variety High Variety
$900
$800
$700
$600
$500
$400
$300
$200
$100
$0
$310.42
$466.24
$808.85
$38.99
$315.54
$793.95
Negative economic condition
Positive economic condition
C: Personal Marketing and Donation Variety
No Variety Low Variety High Variety
$1,000
$900
$800
$700
$600
$500
$400
$300
$200
$100
$0
$99.38
$405.88
$912.98
$78.20
$266.94
$502.01
Low impersonal marketing
High impersonal marketing
No Variety Low Variety High Variety
$12,000
$10,000
$8,000
$6,000
$4,000
$2,000
$0
$9,962.69
$5,436.22
$3,520.65
$106.66 $228.31 $478.34
No personal marketing
Some personal marketing
TABLE 3
Donor Selection Using Donation Variety and
Cross-Donation/Total Donation
Cross-Donation
Percent of Donors Donation (Breadth) and Total
Selected in FY 2013 Variety Donation (Depth)
Top 10% of donors $94,900 $89,200
Top 15% of donors $104,200 $100,300
Top 25% of donors $115,900 $112,800
Notes: Total values are all rescaled by the same constant at the
request of the focal university foundation.
field study to determine the extent to which the university
foundation’s efforts can be targeted at donors to encourage
giving across multiple initiatives. The main benefit of run-
ning this field study is twofold. First, a field study offers
strong causal evidence that nonprofit firms can actually
motivate single-initiative donors to increase donation vari-
ety. Second, it can provide a general framework for other
nonprofit foundations to motivate donors to increase dona-
tion variety. In our field study, we manipulate the content of
direct mail and e-mail appeals from the focal university
foundation to motivate single-initiative donors to give to
multiple initiatives and to motivate donors of multiple initia-
tives to further increase the number of initiatives supported.
Objective and Setup
E-mail and direct mail solicitations at this university foun-
dation often ask donors to repeat their donation to the initia-
tive(s) they have supported previously. We want to test
whether changing the content of these solicitations can
affect donation variety. We test whether highlighting other
initiatives in the university with which the donor is likely to
identify can encourage donation variety. We expect that
increasing donation variety should lead to increases in dona-
tion probability and total giving amounts by donors who
respond favorably to the treated marketing communications.
Method
The field study includes a stratified random sample of 1,200
alumni from the focal university who all graduated with a
degree from the business school (undergraduate, master’s of
business administration, and/or doctorate), made at least
one donation in FY 2012, and, as of January 2013, had yet
to donate in FY 2013. We took a stratified random sample
of donors who fell into one of four groups: donors who had
given (1) only to the business school in the past (no donation
variety; n = 200); (2) to multiple initiatives, including the
business school, in the past (positive donation variety; n =
200); (3) only to one other initiative in the past, not including
the business school (no donation variety; n = 400); and (4)
to multiple initiatives in the past, not including the business
school (positive donation variety; n = 400). We use these four
groups so that we can differentiate the effect of encouraging
alumni to increase donation variety and give to schools other
than the one from which they graduated and the effect of
encouraging alumni to donate to the school from which they
graduated. We try to control for the latter by choosing a set of
donors who all graduated with a degree from the same school.
Before we ran the study, we wanted to confirm whether our
88 / Journal of Marketing, July 2015
stratified random samples were not significantly different
from one another. To do so, we provide some descriptive
statistics for each of the four groups (see Table 4).
We compare the differences in means of multiple
variables for Groups 1 and 3 (started with no donation vari-
ety) and Groups 2 and 4 (started with positive donation
variety). We find that there is no statistically significant dif-
ference between Groups 1 and 3 or Groups 2 and 4 on four
dimensions: (1) average FY 2012 gift size, (2) average total
past giving, (3) average number of years since their first
gift, and (4) average time since graduation. We find that
there is no significant difference between the samples that
included the business school and the others. This suggests
that donors who chose to give to a single initiative or multi-
ple initiatives, regardless of whether they included the busi-
ness school in the past, are not different in their donation
behaviors over time.
For the field study, we treated each of the 200 donors
from the two groups (1 and 2) who had given to the busi-
ness school in the past as one set of controls by only target-
ing them with appeal letters asking them to consider giving
again to the same initiatives they had supported in the past,
including the business school (i.e., the usual appeal all
alumni from the focal university receive). We then ran-
domly split the two groups (3 and 4) that had not yet given
to the business school into two groups of 200 each for each
group. One-half of each group (n = 200) received the con-
trol message asking them to consider giving again to the
same initiatives they had supported in the past (not includ-
ing the business school). Again, this is the usual appeal that
each donor who gives the previous FY receives. The other
half of each group received the treatment message asking
them to consider giving to the same initiatives they had
considered in the past and also to consider giving to the
school from which they had received their degree (i.e., the
business school). Thus, the only difference between the
control and treated message was the addition of language
asking the donor to consider giving money to the business
school as well as the other initiative(s) that the donor had
supported in the past.
The study was run during the second half of FY 2013
(January 1 to June 30). Initial e-mails and direct mailings
were sent to participants in January, and each donor
received one e-mail and one direct mail. A follow-up e-mail
and direct mail were sent at the beginning of May as a
reminder to those donors who had yet to give in FY 2013.
Donations were collected and recorded until the end of the
FY 2013. Given the similarities in donor histories across
TABLE 4
Descriptive Statistics for Field Study
Total Past Years Since Time Since
Group 2012 Gift Size Giving First Gift Graduation
1. No variety—only business school $447 ($270) $2,208 ($5,743) 12.7 (6.7) 21.5 (15.2)
2. Variety—including business school $801 ($410) $8,989 ($15,899) 13.4 (6.1) 25.0 (11.1)
3. No variety—no business school $455 ($255) $2,402 ($6,192) 12.9 (6.3) 21.4 (15.4)
4. Variety—no business school $790 ($428) $8,690 ($15,024) 13.2 (5.9) 24.4 (11.4)
Notes: This table presents means with standard deviations in parentheses.
samples, the difference in giving in FY 2013 cannot be
attributed to differences in a donor’s intrinsic traits. Rather,
the differences should be attributed to the treatment effect
(i.e., change in content of the appeal).
Results
To determine whether the field study was successful, we
must first determine whether donors who have already
given to the business school previously are similar in their
giving patterns to donors who have yet to give to the busi-
ness school. To do so, we compare the results for the con-
trol Groups 1 and 3 (no donation variety) and the treated
Groups 2 and 4 (positive donation variety). For Groups 1
and 3, we find that of the 200 donors contacted for each
group with the control message, 86 and 88 donors, respec-
tively, responded positively with a donation (for detailed
results, see Table 5). Furthermore, the average gift amount
for Group 1 was $398 and for Group 3 was $381, which are
not statistically significantly different (t = .65, p= .51). For
Groups 2 and 4, we find that of the 200 donors contacted for
each group with the control message, 140 and 141 donors,
respectively, responded positively with a donation. We also
find that the average gift amount for Group 2 was $642 and
for Group 4 was $625, which are not statistically significantly
different (t = .35, p= .73). This suggests that any difference
between the treatment and control groups for Groups 3 and
4 should be a result of the change in marketing content and
not attributable to donors’ different giving histories.
Next, we want to determine whether encouraging
donors to increase their donation variety by asking them to
consider adding the business school as another recipient of
their donation was successful. To do so, we compare the
results from Groups 3 and 4 across the treatment and con-
trol groups. We find for Group 3 (no previous donation
variety) that asking the donor to consider the business
school as an initiative to support led to an increase from 88
to 119 gifts, a 35% increase in repeat giving, and an
increase in the average gift size from $381 to $489—a 28%
increase in donation size (t = 4.57, p< .01). Furthermore,
donation variety increased from 0 to .395. We find for
Group 4 (with positive donation variety) that asking the
donor to consider the business school as an initiative to sup-
port led to an increase from 141 to 144 gifts, a 2% increase
Developing Donor Relationships / 89
in repeat giving, and an increase in the average gift size
from $625 to $835—a 33% increase in donation size (t =
4.13, p< .01). Furthermore, we observe an increase in
donation variety from .573 to .661.
Discussion
The results of this field study suggest that encouraging
donors to support additional initiatives (especially when the
additional initiative has a high degree of fit) can success-
fully increase both the number of donors who repeat and the
donation amount. In the case of donors who have never
spread gifts across multiple initiatives (Group 3), we found
that there was a significant increase in both the percentage
of donors who were retained from the previous year and the
average value of each donor’s total gift. However, in the
case of donors who have already spread gifts across multi-
ple initiatives (Group 4), we found that there was only a
significant increase in the average value of each donor’s
total gift. This suggests that when donors have only given to
one initiative, successfully aligning them with an initiative
that has a high degree of fit (in this case, being an alum of
the focal school) can affect both the total size of the gift and
donors’ decisions to give in the first place. When donors are
already spreading donations across multiple initiatives,
encouraging them to give to another initiative with a high
degree of fit does not seem to affect their decision to make
a gift to the university. However, it does seem to signifi-
cantly increase the total amount of the gift.
Implications for Marketing Theory
and Practice
The results of this study have several implications for mar-
keting theory and practice. First, we find that the amount and
distribution of gifts across multiple initiatives (donation vari-
ety) is a good predictor of future donation behavior. We show
that donors with a higher level of donation variety of their
past gifts are more likely to give in the future and, conditional
on the gift being made, are likely to give more than donors
with a lower degree of donation variety. This result offers
some empirical validation to the literature on social impact
theory and donor–nonprofit relationship management.
TABLE 5
Results from the Field Study
Message Content
Group Control Treatment
1. No variety—only business school Made a gift: 86 (43%) N.A.
Average gift (SD): $398 ($261)
2. Variety—including business school Made a gift: 140 (70%) N.A.
Average gift (SD): $642 ($494)
3. No variety—no business school Made a gift: 88 (44%) Made a gift: 119 (59.5%)
Average gift (SD): $381 ($260) Average gift (SD): $489 ($210)
4. Variety—no business school Made a gift: 141 (70.5%) Made a gift: 144 (72%)
Average gift (SD): $625 ($487) Average gift (SD): $835 ($530)
Notes: N.A. = not applicable.
Furthermore, we present empirical evidence that dona-
tion variety can also change a donor’s responsiveness to the
changes in macroeconomic environment. We find that
donors with a higher degree of donation variety are less
responsive to negative changes in the macroeconomic con-
dition, making them less risky because their donation pat-
terns have lower volatility with external shocks. In addition,
we tested the interaction effect between marketing efforts
and donation variety. We find that donors with greater dona-
tion variety are more responsive to marketing efforts of the
foundation. More testing is necessary to validate this con-
clusion, as it might be the case that donors who ultimately
give to multiple initiatives might be more responsive to
marketing efforts even before they begin giving to multiple
initiatives. However, this result does offer some evidence
that nonprofits can increase their return on marketing
investments by focusing their marketing efforts on donors
with a higher level of donation variety.
Second, we find that our measure of donation variety is
better able to capture the variation in donation amount and
more accurately predicts the expected donation amount
(both in-sample and out-of-sample) than the measure of
cross-donation. This finding contributes to the literature on
measurements of the breadth of customer relationships. Fur-
thermore, it suggests that measures such as cross-buy (i.e.,
number of product categories purchase), commonly used in
the customer relationship management literature to measure
the breadth of a relationship, could be enhanced by also
capturing the distribution of purchases.
Third, this study empirically shows that factors that
drive donation variety systematically change over the
course of the relationship between the donor and the non-
profit organization. We find that donor intrinsic motivations
(marketing efforts) are more (less) effective at driving
donors to give to multiple initiatives at the acquisition stage
rather than the retention stage. This is an important contri-
bution to the literature on donation behavior, as studies to
this point have not distinguished the changes in a donor’s
responsiveness to marketing efforts over the course of the
donor–nonprofit organization relationship.
Fourth, for many nonprofit organizations that silo their
initiatives into separate subfoundation departments, the results
suggest that the overall goal should be to share donors with
other departments rather than get donors to give only to a
single initiative. Encouraging donors to give across multiple
departments leads to an increase in the likelihood of donation
and amount of money a donor will provide the university on
the whole in the future and acts as a buffer against the
potential loss of donations due to an economic downturn.
Finally, our study provides causal evidence of the value of
donation variety through a field study applied in marketing
practice. We show that by altering the content of the market-
ing message with an appropriate appeal to increase dona-
tion variety, we were able to increase the likelihood and
amount of giving by donors who had only given to single
initiatives to that point and increase the amount of giving by
donors who had already donated to multiple initiatives. This
contribution is important because it offers external valida-
90 / Journal of Marketing, July 2015
tion to academic research, making it more likely that other
marketing research studies will be implemented in practice.
Limitations and Opportunities for
Further Research
We acknowledge that this study was completed with a
single nonprofit organization. Although this potentially lim-
its the generalizability of the study, it requires significant
effort to run a field study with a single firm (in this case, an
educational foundation), which is a significant contribution
of this study. Moreover, while gifts to foundations make up
a significant amount of the total giving to nonprofit organi-
zations, further research in the nonprofit context can
address several issues, such as the following: how the
breadth of supporting multiple causes and fund-raising pro-
jects offered by a nonprofit might affect a donor’s total
value to the nonprofit across different types of nonprofit
organizations; how financial well-being might affect a
donor’s decision to donate to a nonprofit and to multiple
initiatives; and how measuring breadth as a single measure,
controlling for the distribution of donations across initia-
tives, has applications in other contexts (e.g., cross-buying).
Appendix: Comparing Donation
Variety and Number of Initiatives
Supported
We estimate the same model shown in the article body with
one exception. We substitute donation variety with cross-
donation to represent the number of initiatives supported.
We also include the log of the cumulative donation amount
(ln(Gift_Cum)) to represent the depth of gift amounts
across all the initiatives and the interaction between the two
new variables to represent the breadth and breadth of giv-
ing. Table A1 presents the results.
The results of the model suggest that estimates for
cross-donation, log of the cumulative donation amount
(ln(Gift_Cum)), and the interaction effects are all significant
(like the donation variety and interaction effects in the pro-
posed model). This suggests that both sets of measures cap-
ture a similar effect (i.e., the key insights are the same).
However, we note that the model fit (log-pseudo-likelihood)
is better for the donation variety model.
It is also important to determine which model is able to
better predict the expected donation amount both in-sample
and out-of-sample. To test the in-sample model fit differ-
ences, we compare the MAD and MAPE of the expected
donation value for all donors in all time periods in the
observation window. We find that the in-sample MAD
(MAPE) for the donation variety model is $60.11 (17.27%)
and for the cross-donation and total donation model is
$68.81 (18.01%), which suggests that the model with a
single measure of donation variety captures more of the var-
iation in expected donation amounts than the model with
separate variables representing the breadth and depth of
donations (cross-donation and total donation) as well as the
interaction between cross-donation and total donation.
Next, we test the out-of-sample fit of the model. To do so,
we used the coefficients from the models to select the
“best” donors on the basis of the predicted expected dona-
tion in FY 2013, given that the original observation window
of the data extends through FY 2012. We then determine
whether the university foundation is better off selecting the
top percentiles of donors (10%, 15%, and 25%) on the basis
of their expected donation from the donation variety or
cross-donation and total donation models (see Table 3).
Table 3 shows that regardless of whether the foundation
selects the top 10%, 15%, or 25% of donors on the basis of
Developing Donor Relationships / 91
the expected giving amount predicted by the two different
models, the model using donation variety helps the univer-
sity foundation select donors with higher giving amounts in
FY 2013 by 6.4%, 3.9%, and 2.7%, respectively. These
results suggest that using the proposed measure of donation
variety is more valuable than the traditional (and separate)
measures of breadth (cross-donation) and depth (total dona-
tion) because donation variety is a single measure (rather
than two independent measures) that captures both the
breadth and distribution of depth of donations made to a
nonprofit organization.
TABLE A1
Results for the Donation Variety and Cross-Donation Models
Cross-Donation Model (CDi, t) Donation Variety Model (Varietyi, t)
Selection Main Selection Main
Variables Coefficient (SE) Coefficient (SE) Coefficient (SE) Coefficient (SE)
Intercept –.575 (.438) 2.819 (.020)** –.032 (.025) 2.412 (.046)**
ln(Gifti, t – 1) .436 (.004)** .108 (.003)** .204 (.003)** .252 (.005)**
Varietyi,t – 1 — — .071 (.021)** .157 (.032)**
CDi, t – 1 1.641 (.295)** .076 (.007)**
ln(Gift_Cumi, t – 1) .629 (.008)** .298 (.004)**
Extrinsic
ImpMktgCosti, t .117 (.007)** .039 (.004)** .044 (.009)** .130 (.020)**
PerMktgCosti, t .015 (.001)** .022 (.005)** .006 (.002)** .032 (.007)**
Econt 1.341 (.654)* .969 (.277)** 1.524 (.377)** 4.753 (.503)**
Intrinsic
Degreesi, t .032 (.007)** .084 (.011)** .065 (.013)** .087 (.025)**
Spousei .148 (.017)** .032 (.011)** .093 (.014)** .005 (.027)
Interaction Effects
Varietyi,t – 1 ¥ImpMktgCosti,t
.043 (.004)** .003 (.001)**
Varietyi,t – 1 ¥PerMktgCosti,t .001 (.0004)* .002 (.0005)**
Varietyi,t – 1 ¥Econt –4.143 (.774)** –.606 (.101)**
CDi, t – 1 ¥ln(Gift_Cumi, t – 1) .289 (.005)** .149 (.002)**
CDi, t – 1 ¥ImpMktgCosti, t .012 (.002)** .003 (.001)**
CDi, t – 1 ¥PerMktgCosti, t
.002 (.0004)** .0011 (.0001)**
CDi, t – 1 ¥Econt –1.354 (.279)** –.529 (.133)**
ln(Gift_Cumi, t – 1) ¥ImpMktgCosti, t
.018 (.002)** .001 (.0003)**
ln(Gift_Cumi, t – 1) ¥PerMktgCosti, t .001 (.0002)** .0002 (.00005)**
ln(Gift_Cumi, t – 1) ¥Econt
–.007 (.002)** –.182 (.020)**
Control Variables
Genderi .054 (.017)** .282 (.010)** .039 (.013)** .439 (.026)**
Locationi
.038 (.018)* .037 (.010)** .054 (.014)** .054 (.025)*
Time Since Graduationi, t
–.032 (.003)** .041 (.003)** –.071 (.002)** .033 (.003)**
AvgAGIi, t
.0005 (.0001)** .002 (.0003)** .0002 (.0001)* .003 (.0003)**
AvgGivingi, t
.033 (.016)* .541 (.388) .064 (.024)** .643 (.259)*
Control Function Variables
ImpMktgErrori, t –.035 (.007)** –.096 (.016)** –.058 (.009)** –.146 (.020)**
PerMktgErrori, t .005 (.002)* –.019 (.004)** .009 (.002)** –.028 (.007)**
Model Fit
Log-pseudo-likelihood –108,121.62 –107,989.78
*p< .05.
**p< .01.
Notes: AGI = adjusted gross income.
Ai, Chunrong and Edward C. Norton (2003), “Interaction Terms
in Logit and Probit Models,” Economics Letters, 80 (1),
123–29.
REFERENCES
Aaker, Jennifer L. and Satoshi Akutsu (2009), “Why Do People
Give? The Role of Identity in Giving,” Journal of Consumer
Psychology, 19 (3), 267–70.
92 / Journal of Marketing, July 2015
Amabile, Teresa M. (1993), “Motivational Synergy: Toward New
Conceptualizations of Intrinsic and Extrinsic Motivation in the
Workplace,” Human Resource Management Review, 3 (3),
185–201.
Andreoni, James (1990), “Impure Altruism and Donations to Pub-
lic Goods: A Theory of Warm-Glow Giving,” Economic Jour-
nal, 100 (401), 464–77.
(2006), “Philanthropy,” in Handbook of the Economics of
Giving, Altruism and Reciprocity, Vol. 2, Serge-Christophe
Kolm and Jean Mercier Ythier, eds. Amsterdam: North-Holland,
1201–1269.
Apinunmahakul, Amornrat and Rose A. Devlin (2008), “Social
Networks and Private Philanthropy,” Journal of Public Eco-
nomics, 92 (1), 309–328.
Ariely, Dan and Michael I. Norton (2009), “Conceptual Consump-
tion,” Annual Review of Psychology, 60, 475–99.
Bekkers, René and Pamala Wiepking (2007), “Understanding
Philanthropy: A Review of 50 Years of Theories and
Research,” paper presented at the 35th Annual Conference of
the Association for Research on Nonprofit and Voluntary
Action, Chicago (November 16–18).
Bennett, Roger (2012), “What Else Should I Support? An Empiri-
cal Study of Multiple Cause Donation Behavior,” Journal of
Nonprofit & Public Sector Marketing, 24 (1), 1–25.
Blackbaud Inc. (2014), “2013 donorCentrics™ Annual Report on
Higher Education Alumni Giving,” research report, (accessed
March 30, 2015), [available at https://www.blackbaud.com/
files/resources/downloads/2014/04-14_TA_ donorCentrics_
HEAnnualReport_UPDATED.pdf].
Bolton, Ruth N. (1998), “A Dynamic Model of the Duration of the
Customer’s Relationship with a Continuous Service Provider:
The Role of Satisfaction,” Marketing Science, 17 (1), 45–65.
Brooks, Arthur C. (2005), “Does Social Capital Make You Gener-
ous?” Social Science Quarterly, 86 (1), 1–15.
Brown, Eleanor and James M. Ferris (2007), “Social Capital and
Philanthropy: An Analysis of the Impact of Social Capital on
Individual Giving and Volunteering,” Nonprofit and Voluntary
Sector Quarterly, 36 (1), 85–99.
Burke, William J. (2009), “Fitting and Interpreting Cragg’s Tobit
Alternative Using Stata,” Stata Journal, 9 (4), 584–92.
Cause Marketing Forum (2010), “Do Well Do Good Cause Mar-
keting Survey,” (accessed April 2, 2015), [available at http://
www.causemarketingforum.com/site/apps/nlnet/content2.aspx?
c= bkLUKcOTLkK4E&b=6412289&ct=8973411].
Celsi, Richard L. and Jerry C. Olson (1988), “The Role of
Involvement in Attention and Comprehension Processes,
Journal of Consumer Research, 15 (2), 210–24.
Chatterjee, Sayan and James D. Blocher (1992), “Measurement of
Firm Diversification: Is It Robust?” Academy of Management
Journal, 35 (4), 874–88.
Coleman, James S. (1988), “Social Capital in the Creation of
Human Capital,” American Journal of Sociology, 94 (Supple-
ment), S95–S120.
Cragg, John G. (1971), “Some Statistical Models for Limited
Dependent Variables with Application to the Demand for
Durable Goods,” Econometrica, 39 (5), 829–44.
Foster, Mary K. and Agnes G. Meinhard (2002), “A Regression
Model Explaining Predisposition to Collaborate,” Nonprofit
and Voluntary Sector Quarterly, 31 (4), 549–64.
Frenzen, Jonathan K. and Harry L. Davis (1990), “Purchasing
Behavior in Embedded Markets,” Journal of Consumer
Research, 17 (1), 1–12.
George, Morris, V. Kumar, and Dhruv Grewal (2013), “Maximiz-
ing Profits for a Multi-Category Catalog Retailer,” Journal of
Retailing, 89 (4), 374–96.
Granovetter, Mark (1985), “Economic Action and Social Struc-
ture: The Problem of Embeddedness,” American Journal of
Sociology, 91(3), 481–510.
Grant, Adam M. (2012), “Giving Time, Time After Time: Work
Design and Sustained Employee Participation in Corporate
Volunteering,” Academy of Management Review, 37 (4), 589–
615.
Gupta, Sunil and Valarie Zeithaml (2006), “Customer Metrics and
Their Impact on Financial Performance,” Marketing Science,
25 (6), 718–39.
Herniter, Jerome D. (1976), An Entropy Model of Brand Purchase
Behavior. Berlin/Heidelberg: Springer.
Hodrick, Robert J. and Edward C. Prescott (1997), “Postwar U.S.
Business Cycles: An Empirical Investigation,” Journal of
Money, Credit, and Banking, 29 (1), 1–16.
Hoskisson, Robert E., Michael A. Hitt, Richard A. Johnson, and
Douglas D. Moesel (1993), “Construct Validity of an Objective
(Entropy) Categorical Measure of Diversification Strategy,”
Strategic Management Journal, 14 (3), 215–35.
Kahn, Barbara E. (1995), “Consumer Variety-Seeking Among
Goods and Services: An Integrative Review,” Journal of
Retailing and Consumer Services, 2 (3), 139–48.
Kapur, J.N., C.R. Bector, and Uma Kumar (1984), “A Generaliza-
tion of the Entropy Model for Brand Purchase Behavior,”
Naval Research Logistics Quarterly, 31 (2), 183–98.
Kottasz, Rita (2004), “How Should Charitable Organizations
Motivate Young Professionals to Give Philanthropically?”
International Journal of Nonprofit and Voluntary Sector Mar-
keting, 9 (1), 9–27.
Kumar, V., Morris George, and Joseph Pancras (2008), “Crossbuy-
ing in Retailing: Drivers and Consequences,” Journal of
Retailing, 84 (1), 15–27.
and J. Andrew Petersen (2005), “Using a Customer-Level
Marketing Strategy to Enhance Firm Performance: A Review
of Theoretical and Empirical Evidence,” Journal of the Acad-
emy of Marketing Science, 33 (4), 504–519.
Lemon, Katherine N., Tiffany Barnett White, and Russell S. Winer
(2002), “Dynamic Customer Relationship Management: Incor-
porating Future Considerations into the Service Retention
Decision,” Journal of Marketing, 66 (January), 1–14.
Li, Sherry Xin, Catherine C. Eckel, Philip J. Grossman, and Tara
Larson (2013), “Who’s in Charge? Donor Targeting Enhances
Voluntary Giving to Government,” (accessed March 30, 2015),
[available at http://papers.ssrn.com/sol3/papers.cfm?abstract_
id= 2293407].
Li, Shibo, Baohong Sun, and Alan L. Montgomery (2011), “Cross-
Selling the Right Product to the Right Customer at the Right
Time,” Journal of Marketing Research, 48 (August), 683–700.
Ly, Pierre and Geri Mason (2012), “Individual Preferences over
Development Projects: Evidence from Microlending on Kiva,”
Voluntas: International Journal of Voluntary and Nonprofit
Organizations, 23 (4), 1036–55.
Magson, Nigel (1999), “Donors: How Much Do They Give in a
Lifetime?” International Journal of Nonprofit and Voluntary
Sector Marketing, 4 (1), 11–25.
Masters, Tony (2000), “Deciding to Recruit Only Donors with
High Lifetime Values,” International Journal of Nonprofit and
Voluntary Sector Marketing, 5 (3), 241–47.
McAdam, Doug and Ronnelle Paulsen (1993), “Specifying the
Relationship Between Social Ties and Activism,” American
Journal of Sociology, 99 (3), 640–67.
McAlister, Leigh and Edgar Pessemier (1982), “Variety Seeking
Behavior: An Interdisciplinary Review,” Journal of Consumer
Research, 9 (3), 311–22.
Null, Clair (2011), “Warm Glow, Information, and Inefficient
Charitable Giving,” Journal of Public Economics, 95 (5), 455–
65.
Palepu, Krishna (1985), “Diversification Strategy, Profit Perfor-
mance and the Entropy Measure,” Strategic Management Jour-
nal, 6 (3), 239–55.
Palich, Leslie E., Laura B. Cardinal, and C.C. Miller (2000),
“Curvilinearity in the Diversification? Performance Linkage:
An Examination of Over Three Decades of Research,” Strate-
gic Management Journal, 21 (2), 155–74.
Petrin, Amil and Kenneth Train (2010), “A Control Function
Approach to Endogeneity in Consumer Choice Models,” Jour-
nal of Marketing Research, 47 (February), 3–13.
Putnam, Robert D. (1995), “Bowling Alone: America’s Declining
Social Capital,” Journal of Democracy, 6 (1), 65–78.
Reagans, Ray (2005), “Preferences, Identity, and Competition:
Predicting Tie Strength from Demographic Data,” Manage-
ment Science, 51 (9), 1374–83.
Robinson, Stefanie Rosen, Caglar Irmak, and Satish Jayachandran
(2012), “Choice of Cause in Cause-Related Marketing,” Jour-
nal of Marketing, 76 (July), 126–39.
Ryan, Richard M. and Edward L. Deci (2000), “Intrinsic and
Extrinsic Motivations: Classic Definitions and New Direc-
tions,” Contemporary Educational Psychology, 25 (1), 54–67.
Sargeant, Adrian (2001), “Managing Donor Defection: Why
Should Donors Stop Giving?” New Directions for Philan-
thropic Fundraising, 2001 (32), 59–74.
— and Lucy Woodliffe (2007), “Gift Giving: An Interdiscipli-
nary Review,” International Journal of Nonprofit and Volun-
tary Sector Marketing, 12 (4), 275–307.
Simonson, Itamar and Russell S. Winer (1992), “The Influence of
Purchase Quantity and Display Format on Consumer Prefer-
ence for Variety,” Journal of Consumer Research, 19 (1), 133–
38.
Developing Donor Relationships / 93
Sirgy, M. Joseph (1982), “Self-Concept in Consumer Behavior: A
Critical Review,” Journal of Consumer Research, 9 (3), 287–
300.
Swann, William B., Jr., and Michael J. Gill (1997), “Confidence
and Accuracy in Person Perception: Do We Know What We
Think We Know About Our Relationship Partners?” Journal of
Personality and Social Psychology, 73 (4), 747–57.
Thornton, Jeremy (2006), “Nonprofit Fund-Raising in Competi-
tive Donor Markets,” Nonprofit and Voluntary Sector Quar-
terly, 35 (2), 204–224.
Uzzi, Brian (1999), “Embeddedness in the Making of Financial
Capital: How Social Relations and Networks Benefit Firms
Seeking Financing,” American Sociological Review, 64 (4),
481–505.
Venkatesan, Rajkumar and V. Kumar (2004), “A Customer Life-
time Value Framework for Customer Selection and Resource
Allocation Strategy,” Journal of Marketing, 68 (October), 106–
125.
Verhoef, Peter C., Philip H. Franses, and Janny C. Hoekstra
(2001), “The Impact of Satisfaction and Payment Equity on
Cross-Buying: a Dynamic Model for a Multi-Service
Provider,” Journal of Retailing, 77 (3), 359–78.
—, ——, and (2002), “The Effect of Relational Con-
structs on Customer Referrals and Number of Services Pur-
chased from a Multiservice Provider: Does Age of Relation-
ship Matter?” Journal of the Academy of Marketing Science,
30 (3), 202–216.
Woerheide, Walt and Don Persson (1993), “An Index of Portfolio
Diversification,” Financial Services Review, 2 (2), 73–85.
... 1 A major challenge that many NPOs face is the volatility of individual giving: the literature reports that approximately half of newly acquired donors churn after they give once (Sargeant and Woodlie 2007;Khodakarami et al. 2015). In response to such instability, NPOs strive to increase repeat giving by individuals as well as to identify and retain givers who are more committed. ...
... Despite the prevalence of multiple forms of giving in practice, and the recognition that benets can have an eect on subsequent giving, previous marketing studies typically treat giving via membership and donation as a composite (Netzer et al. 2008;Van Diepen et al. 2009;Khodakarami et al. 2015;Kumar et al. 2015). Such aggregation across forms of giving masks potentially large dierences in individuals' motivations and giving behaviors. ...
... We compare the full model estimated using Equation 5, which incorporates both random eects (unobserved heterogeneity) and GP (dynamics), which we denote by M1, with more restrictive benchmark models, M2-M5. As discussed, a distinctive feature of our framework is that we allow multiple forms of giving, whereas previous research combines dierent forms of nancial giving as a composite amount (Netzer et al. 2008;Van Diepen et al. 2009;Khodakarami et al. 2015;Kumar et al. 2015 Posterior means and standard deviations related to the GP and theR statistic are shown in Table 2. As explained in the modeling section, these parameter estimates summarize the characteristics of the GP structures. ...
Article
Full-text available
Nonprofit organizations (NPOs) play a central role in many economies in the form of private entities discharging a public purpose. Strengthening the fundraising capabilities of NPOs can have a large impact on their survival and effective functioning. NPOs typically elicit financial contributions through multiple forms of giving, such as donation and membership. These options enable individuals to express their altruism by giving in one or multiple forms. The authors develop a utility-based multiple discrete continuous model that provides insights into potentially large differences in individuals' giving behaviors. The model also incorporates via Bayesian Gaussian processes changes in givers' preferences for forms of giving. The authors apply their model to five years of individual giving data. They find that the effects of lifetime, recency, seasonality, and appeals on donation and membership options change non-monotonically over time, and in distinctive ways. The authors demonstrate that the model estimates help to predict who will give in more than one form in the future, and to build appeal targeting strategies. Additional marketing insights that the model provides are that fundraising attempts should emphasize participation rather than amount, and that long-lapsed members are still worth pursuing for renewal, while long-lapsed donors are less productive for repeat giving.
... The first research stream comprises studies that examine how policies and/or different marketing initiatives of an organization influence consumers' behavior and/or attitudes (e.g., Bolton, Bhattacharjee, and Reed 2015;Dority, McGarvey, and Kennedy 2010;Moore et al. 2002;Wang, Lewis, and Singh 2016;Webb and Mohr 1998). The second research stream comprises studies that evaluate how marketing efforts can be utilized for fundraising and managing donor behavior and donor relationships (e.g., Anik, Norton, and Ariely 2014;Arnett, German, and Hunt 2003;Hung and Wyer 2009;Kara, Spillan, and DeShields 2004;Khodakarami, Petersen, and Venkatesan 2015;Savary, Goldsmith, and Dhar 2015;Shang, Reed, and Croson 2008;Townsend 2017;Van Diepen, Donkers, and Franses 2009;Zhou, Kim, and Wang 2018). The third research stream comprises studies that have proposed methodological advancements in the context of NPs, with the underlying objective of improving prediction and/or understanding donor behavior (e.g., Aravindakshan, Rubel, and Rutz 2015;Dubé, Luo, and Fang 2017;Gopalakrishnan, Bradlow, and Fader 2017;Schweidel and Knox 2013;Netzer, Lattin, and Srinivasan 2008). ...
Article
Full-text available
The critical role of marketing in driving nonprofit performance has been recognized for decades. However, in practice, there has been a disturbingly weak acknowledgment and/or implementation of marketing practices across nonprofits to date. Marketing is often perceived as an avoidable and costly overhead. The issue is complicated by the fact that nonprofit performance is relatively difficult to measure and may often comprise multiple tangible and intangible outcomes with different (linear and nonlinear) functional forms. Furthermore, nonprofit performance outcomes often depend on behavioral and attitudinal changes of the target segment. The authors address these challenges by presenting a methodology to link marketing efforts to nonprofits' mission-based performance outcome(s). The authors apply their approach with data from a large nonprofit and find empirical support for the notion that marketing can play a pivotal and significant role in improving nonprofits' mission-based performance outcomes. The findings help present a strong case for nonprofit leaders and policy makers to fund and treat marketing as a critical investment to drive nonprofit entities' performance.
Article
Despite the popularity of not-for-profit organizations (NFPs) as a research topic, existing comprehensive reviews are rather limited, resulting in a large body of literature that lacks integration. In this review, we summarize prior research on NFPs, focusing on four key differences between for-profit organizations and NFPs: organizational goals, financial resources, human resources, and leadership and governance. Our aims are to delineate the major inroads that have been made—not only in theory testing, but also in theory building—and to identify fruitful avenues for continued inquiry.
Article
Full-text available
The introduction of digital technologies in agricultural production is one of the most important elements of strategic development in the agricultural sector and rural areas in Ukraine. In agriculture, these new technologies can modernize the industry, promoting innovation in agribusiness and creating new opportunities for rural development. The introduction of digital technologies in agriculture ensures the accuracy of measurements, speed data collection and processing. Digitization in rural areas is an inevitable process that brings a number of economic, social and environmental benefits. The immediate aim of this paper is to assess the state of implementation of digital technologies in agriculture and to examine opportunities for rural development in Ukraine. Research methods: monographic, descriptive, analysis, synthesis, induction. The results indicate that only large agricultural enterprises in Ukraine are able to implement and use digital technologies. Thus, it is proposed to create an integrated digital portal for agricultural needs, combining solutions that optimize activities of agricultural enterprises: land bank management, production, crop monitoring, warehouse, procurement and supply, equipment and repairs, logistics, inventory and finished products. The article identifies technological and human barriers to introduction of digital technologies in rural areas of Ukraine. In addition it proposes strategies for development of digital literacy and skills among rural residents in Ukraine. The results of the research can have a significant impact on the development of agriculture in Ukraine, promoting digital technologies among other agricultural enterprises and ensuring the development in rural areas, attracting additional agricultural market participants and infrastructure that provide relevant information and digital services to rural residents.
Chapter
Nowadays, integrating new technology into all management processes leads to significant evolution, particularly in the logistics business, which is one of the fourth industrial of Industry 4.0. Blockchain, an emerging idea, provides for the decentralized and unchangeable storing of verified data as integrated technology. The shipping industry has had to look for innovative ways to keep the accelerated growth of the planet on track, in the face of development threats and vulnerabilities from the mild development of foreign markets, expanding protectionism, correcting natural guidance, the current episode of COVID-19 pandemics. One of the successful innovations in blockchain technologies was to promote a computerized market change. It is also making the utilization of blockchain innovation in the maritime industry will empower quicker, more secure and more productive businesses. The objective of this study is to review sustainable in supply chain management through blockchain technology in the maritime industry. This work was done by discussion of literature by classifying the application according to the operation in the shipping process.
Article
Relationship marketing has been key to developing stronger relationships and promoting donor continuity in nonprofit organizations. However, it is not clear whether these efforts are effective for all individuals who choose to financially support an organization. Taking a relationship orientation approach, this study investigates the effect of consenting to direct marketing on occasional donors’ decisions to become members (i.e., contributing through membership programs), and analyzes the moderating role of past donation behavior (i.e., frequency and amount). The results reveal that occasional donors who consent to direct marketing are more likely to develop a long-term relationship as members. This effect is reinforced when donors have contributed greater amounts in the past. The findings also demonstrate that, through their membership, donors consenting to direct marketing provide greater financial support on an ongoing basis. This study offers novel insights into the relationship marketing literature and provides contributions to fundraising managers.
Article
Purpose The purpose of this study is to investigate whether a U-shaped relationship exists between the length of time a donor has been a regular member in a nonprofit organization and the amount donated over time. In addition, this research analyzes whether this relationship is moderated by donation frequency. Design/methodology/approach Using a database of 6,137 members from a collaborating nonprofit organization, a longitudinal study is conducted over an eight-year period (2013–2020). A set of ordinary least square (OLS) regression analyses are carried out to empirically test the proposed hypotheses. Findings This study finds a nonlinear, U-shaped relationship between donation amount and relationship length. This effect can be explained through the dynamic evolution of two dimensions of commitment: affective (decreasing over time) and normative (increasing over time). The results also reveal that these effects, however, become flatter for members who engage in more frequent donations. Originality/value The results provide novel insights revealing the nonlinear nature of the relationship between the length of time a donor has been a member of a nonprofit organization and the amount donated, and underscores the moderating role of donation frequency, which makes the U-shaped relationship flatter, thus increasing the amounts donated. Despite their relevance in the service ecosystem, nonprofits have been under-represented in prior work. This study offers important practical insights into the effective management of the regular donor portfolio.
Article
With the popularity of user-generated content (UGC), an increasing number of studies have investigated its impact on business performance. However, prior studies were limited to a single platform and showed the effects of UGC of a platform, for example, customer textual comments or customer numeral ratings, on sales/reservation of the same platform. In practice, users often refer to a UGC, for example, Instagram, and purchase it on other platforms. To incorporate the spillover effect, we considered the restaurant industry because it has active participation across various channels. Using topic modelling, we first identified from Instagram four topics of users’ interest regarding a restaurant, such as location, nightlife, food and celebration. From fixed effects models’ estimation, we found that (a) recommendation and mention of Instagram have positive effects, and (b) comments of location and food also have positive significant fixed effects, but (c) the impact of Instagram volume is curvilinear and positive significant effect the sales. Since the curvilinear effects may come from reverse causality, that is, higher reservation, might bring more customers and comments on social networking service (SNS) (echo verse effect in our paper). Therefore, we further analysed two-way Granger causality and panel vector autoregression to identify the endogeneity, and the results showed the existing Granger causality loop between OpenTable review and Instagram post volumes.
Article
Research suggests that children develop best in families. However, millions of children live in residential care centres worldwide. Many residential centres desire to transition their programs from a residential to a family-based model of care, but face barriers surrounding funding and donor support. Little research exists on how organisations address these concerns. The current article investigated the financial impact of transitioning to a family-based model of care, donor engagement practices used in this process, and changes in donor support resulting from the transition. Twenty-six organisations that had fully or partially transitioned their model completed a brief survey. Data revealed an initial increase in cost per child, but long-term the cost of services decreased. Further, findings suggested that involving donors early and using multiple methods and types of communication led to better long-term donor support. Based on these findings, five recommendations were made for organisations planning to transition their care model.
Article
The critical role of marketing in driving nonprofit performance has been recognized for decades. However, in practice, there has been a disturbingly weak acknowledgment and/or implementation of marketing practices across nonprofits to date. Marketing is often perceived as an avoidable and costly overhead. The issue is complicated by the fact that nonprofit performance is relatively difficult to measure and may often comprise of multiple tangible and intangible outcomes with different (linear and nonlinear) functional forms. Furthermore, nonprofit performance outcomes often depend on behavioral and attitudinal changes of the target segment. The authors address these challenges by presenting a methodology to link marketing efforts to nonprofits’ mission-based performance outcome(s). The authors apply their approach with data from a large nonprofit and find empirical support for the fact that marketing can play a pivotal and significant role in improving nonprofits’ mission-based performance outcomes. The findings help present a strong case for nonprofit leaders and policymakers to fund and treat marketing as a necessary investment to drive nonprofit entities’ performance. Implications for conceptual, practical, and substantive contributions and future research directions are discussed.
Article
Full-text available
This paper investigates the potential effect of allowing donors to target voluntary giving to specific government organizations on government’s ability to generate additional revenues. We design a ‘real donation’ lab experiment to compare giving to general revenue with donor-targeted giving to specific causes. Results show that granting donors control over the use of their funds more than doubles giving to government. A comparison with private organizations that serve similar functions shows that targeting significantly shrinks the disadvantageous gap in voluntary contributions to government. We argue that creating new mechanisms to facilitate voluntary giving has the potential to enhance revenue during fiscal crisis, especially when donations can be targeted for specific programs.
Article
Full-text available
Spurred by the consumer demand for companies to be socially responsible, cause-related marketing (CM), in which fund raising for a cause is tied to purchase of a firm's products, has become popular in recent years. The authors demonstrate the conditions in which CM campaigns that allow consumers to choose the cause that receives the donation lead to greater consumer support than those in which the company determines the cause. They show that choice in this context is helpful as long as it increases consumers' perception of personal role in helping the cause. Specifically, allowing consumers to select the cause in a CM campaign is more likely to enhance perceived personal role and, thus, purchase intentions (1) for those consumers who are high (vs. low) in collectivism and (2) when the company and causes have low (vs. high) perceptual fit. Finally, the authors show that under certain conditions, choice may have a negative impact on perceived personal role and consumer support of CM campaigns.
Article
Reflecting the importance of variety-seeking in consumer choice, there has been an explosion of research in the marketing literature on this topic in the past decade. The goal of this paper is to provide an integrative review of the key findings. In particular, a conceptual, integrating framework for understanding the reasons why consumers seek variety is presented. Within this context, the implications of this research for retail and service management are discussed as well as a review of the measurement tools and predictive models of variety-seeking that have been proposed in the last decade.
Article
Why do people give to others? One principal driver involves one's identity: who one is and how they view themselves. The degree to which identities are malleable, involve a readiness to act, and help make sense of the world have significant implications determining whether and how much people give. Drawing on the Identity-Based Motivation model (IBM; Oyserman, D. (2009). Identity-based motivation: Implications for action-readiness, procedural-readiness, and consumer behavior. Journal of Consumer Psychology, 19.), we provide a tripartite framework to help advance the research on the psychology of giving.
Article
Intrinsic and extrinsic types of motivation have been widely studied, and the distinction between them has shed important light on both developmental and educational practices. In this review we revisit the classic definitions of intrinsic and extrinsic motivation in light of contemporary research and theory. Intrinsic motivation remains an important construct, reflecting the natural human propensity to learn and assimilate. However, extrinsic motivation is argued to vary considerably in its relative autonomy and thus can either reflect external control or true self-regulation. The relations of both classes of motives to basic human needs for autonomy, competence and relatedness are discussed.
Article
Corporate volunteering programs are important channels for expressing care and compassion, but little research has examined when and why employees sustain involvement. Integrating work design and volunteering theories. I introduce a model that explains how depleted task, social, and knowledge characteristics of jobs trigger compensatory motives during initial volunteering episodes. When these motives are fulfilled by volunteering projects, employees repeat participation, internalizing volunteer identities-contingent on pressure. matching incentives, recognition, managerial support, and targeted causes.
Article
Journal of Democracy 6.1 (1995) 65-78 As featured on National Public Radio, The New York Times, and in other major media, we offer this sold-out, much-discussed Journal of Democracy article by Robert Putnam, "Bowling Alone." You can also find information at DemocracyNet about the Journal of Democracy and its sponsor, the National Endowment for Democracy. Many students of the new democracies that have emerged over the past decade and a half have emphasized the importance of a strong and active civil society to the consolidation of democracy. Especially with regard to the postcommunist countries, scholars and democratic activists alike have lamented the absence or obliteration of traditions of independent civic engagement and a widespread tendency toward passive reliance on the state. To those concerned with the weakness of civil societies in the developing or postcommunist world, the advanced Western democracies and above all the United States have typically been taken as models to be emulated. There is striking evidence, however, that the vibrancy of American civil society has notably declined over the past several decades. Ever since the publication of Alexis de Tocqueville's Democracy in America, the United States has played a central role in systematic studies of the links between democracy and civil society. Although this is in part because trends in American life are often regarded as harbingers of social modernization, it is also because America has traditionally been considered unusually "civic" (a reputation that, as we shall later see, has not been entirely unjustified). When Tocqueville visited the United States in the 1830s, it was the Americans' propensity for civic association that most impressed him as the key to their unprecedented ability to make democracy work. "Americans of all ages, all stations in life, and all types of disposition," he observed, "are forever forming associations. There are not only commercial and industrial associations in which all take part, but others of a thousand different types -- religious, moral, serious, futile, very general and very limited, immensely large and very minute. . . . Nothing, in my view, deserves more attention than the intellectual and moral associations in America." Recently, American social scientists of a neo-Tocquevillean bent have unearthed a wide range of empirical evidence that the quality of public life and the performance of social institutions (and not only in America) are indeed powerfully influenced by norms and networks of civic engagement. Researchers in such fields as education, urban poverty, unemployment, the control of crime and drug abuse, and even health have discovered that successful outcomes are more likely in civically engaged communities. Similarly, research on the varying economic attainments of different ethnic groups in the United States has demonstrated the importance of social bonds within each group. These results are consistent with research in a wide range of settings that demonstrates the vital importance of social networks for job placement and many other economic outcomes. Meanwhile, a seemingly unrelated body of research on the sociology of economic development has also focused attention on the role of social networks. Some of this work is situated in the developing countries, and some of it elucidates the peculiarly successful "network capitalism" of East Asia. Even in less exotic Western economies, however, researchers have discovered highly efficient, highly flexible "industrial districts" based on networks of collaboration among workers and small entrepreneurs. Far from being paleoindustrial anachronisms, these dense interpersonal and interorganizational networks undergird ultramodern industries, from the high tech of Silicon Valley to the high fashion of Benetton. The norms and networks of civic engagement also powerfully affect the performance of representative government. That, at least, was the central conclusion of my own 20-year, quasi-experimental study of subnational governments in different regions of Italy. Although all these regional governments seemed identical on paper, their levels of effectiveness varied dramatically. Systematic inquiry showed that the quality of governance was determined by longstanding traditions of civic engagement (or its absence). Voter turnout, newspaper readership, membership in choral societies and football clubs -- these were the hallmarks of a successful region. In fact, historical analysis suggested that these networks of organized reciprocity and civic solidarity...