Exploring options to leverage partnership information for designing donation calls

Purpose – This study aims to help understand individual donors ’ preferences over different designs of humanitarian – business partnerships in managing humanitarian operations and to help understand if donors ’ preferences align with their actual donation behavior. Design/methodology/approach – Choice-based conjoint analysis was used to understand donation preferences for partnership designs, and a donation experiment was performed using real money to understand the alignment of donors ’ preferences with actual donation behavior. Findings – The results show that partnering with the business sector can be a valuable asset for humanitarian organizations in attracting individual donors if these partnerships are managed well in terms of partnership strategy, partnership history and partnership report and disclosure. In particular, the study ﬁ nds that the donation of services and products from businesses corporations to humanitarian organizations are preferable to individual donors, rather than cash. Furthermore, donors ’ preferences are not necessarily aligned with actual donation behavior. Practical implications – The results highlight the importance of presenting objective data on projects to individual donors. The results also show that donors value the provision of services and products by business corporations to humanitarian operations. Originality/value – Partnerships between humanitarian organizations and business corporations are important for the success of humanitarian operations. However, little is known about which partnership designs are most preferable to individual donors and have the biggest chance of being supported ﬁ nancially.


Introduction
These days, donors demand more and more from humanitarian organizations (HOs) in terms of impact on beneficiaries compared to the past (Tomasini and Van Wassenhove, 2009;Scholten et al., 2010;Müller-Stewens, 2019).This is reinforced by the fact that donors are more aware than ever of how HOs manage their operations and use donations received from their donors (Caviola et al., 2014;Michaelidou et al., 2015;Mejia et al., 2019).As information transparency in the non-profit sector is increasing (Saxton et al., 2012;Haski-Leventhal and Foot, 2016), donors will search for more information before they decide to donate.
In the recent decade, the amount donated by the private sector to the humanitarian sector has grown significantly (Development Initiatives, 2018).Private donations currently make up about one-third of all financial contributions collected by the humanitarian sector (Development Initiatives, 2018).However, the significant growth of the contribution of private donors to the humanitarian sector is also accompanied by a growth in the number of HOs, which leads to increased competition among HOs to preserve existing donors and to attract new donors (Aldashev and Verdier, 2010;Caviola et al., 2014).For this reason, it is more and more essential for HOs to be aware of the factors affecting donors' decision to donate and to understand donors' preferences among donation options.
HOs have been struggling to enhance their operational performance (Van Wassenhove, 2006;Oloruntoba and Gray, 2006;Rueede and Kreutzer, 2015).It has been widely acknowledged that partnerships with the business sector may The current issue and full text archive of this journal is available on Emerald Insight at: https://www.emerald.com/insight/2042-6747.htm help HOs in achieving their objectives (Thomas andFritz, 2006: Nurmala et al., 2017).The initiation of partnerships between HOs and business corporations (BCs) to increase the operational capacity of HOs has been suggested in several studies (Van Wassenhove, 2006;Thomas and Fritz, 2006;Rueede and Kreutzer, 2015).Despite their potential impact, studies show that partnerships in managing humanitarian operations have not been optimized in terms of frequencies and activities (Nurmala et al., 2018).One of the causes of the reluctance of the humanitarian sector to invite the participation of the business sector in their operation is that the humanitarian sector worries to receive negative publications from the media, which in the end can reduce their potential donors (Nurmala et al., 2018).These negative publications come from the expectations that the humanitarian sector needs to be seen as independent (Thomas and Fritz, 2006;Rueede and Kreutzer, 2015) and because they worry that the negative image of BCs will also affect them (Thomas and Fritz, 2006;Kov acs and Spens, 2009).
Individual donors play a crucial role in supporting the operations of HOs (Michaelidou et al., 2015;Nurmala et al., 2018;Mejia et al., 2019).The motivations of individual donors for donating to humanitarian operations encompass various factors, including demographic characteristics (Ranganathan and Henley, 2008;Notarantonio and Quiqley, 2009), psychological aspects (Hysenbelli et al., 2013;Morin et al., 2015) and motivational reasons (Paramita et al., 2020).Studies on online donations indicate that the design of donation options influences donors' preferences (Peter and Bhaskar, 2012;Ahn et al., 2018).However, although previous studies have highlighted the significant role of donation option design in donors' intention to donate, the current literature lacks an explanation of how collaborative partnerships between HOs and BCs can affect individual donors' contributions.In today's context, HOs often collaborate with multiple stakeholders (Thomas and Fritz, 2006;Kaneberg, 2018;Müller-Stewens, 2019).Therefore, understanding how partnerships formed by HOs with specific partners can influence the donation intentions of various stakeholders, and determining the most effective donation options to offer them is crucial for the sustainability of humanitarian operations.
Regarding the aforementioned issues, the research problem we aim to address in this study is what the preferences are of individual donors toward different designs of partnerships between HOs and BCs in humanitarian operations.By understanding donors' preferences for specific partnership designs with the business sector, HOs can establish better architecture for humanitarian-business partnerships and optimize them to attract more potential donors.Therefore, the aims of this study are (1) to understand individual donors' preferences over different designs of humanitarian-business partnerships in managing humanitarian operations and (2) to understand if donors' preferences align with their actual donation behavior.To achieve our first research objective, we applied a choice-based conjoint (CBC) analysis to help us understand how partnership designs may affect donor preferences.CBC analysis has been used widely as a tool to explore how people value different attributes of a new product or service designs (Gustafsson et al., 1999;Pullman et al., 2002;Oliveira et al., 2015).It is suited to identify variants in such preferences or to segment markets (Hauser and Rao, 2004).As such, the results of a conjoint analysis study can serve as essential input for shaping calls for donations as they can be used to identify what information to include in such calls and how.CBC analysis first received its popularity in marketing (Kulshreshtha et al., 2018) and then received huge attention in many other disciplines, including operational research (Karniouchina et al., 2009;Maldonado et al., 2015), logistics and supply chain management (Nguyen et al., 2019), product adoption (Braun et al., 2016;Priessner and Hampl, 2020) and donation (Lee and Won, 2011;Donaldson et al., 2020), to help researchers understand respondents' preferences.The second aim of this study is to investigate whether donor preferences align with their actual donations.Studies have shown the occurrence of an intention-behavior gap (for example, see Chandon et al., 2005;Carrington et al., 2010).As such, to judge the value of our findings in the CBC study described above, we need to identify this intention-behavior gap as well.
To complete this research objective, we conducted an experiment using real cash (see Moser and Raffaelli, 2014) and focused on comparing intended donations with actual donations.The results of this study will contribute to the ongoing discussion on identifying the optimal architectures for humanitarian-business partnerships in humanitarian operations and how the partnership decisions between HOs and BCs affect individual donors' preferences to donate their money.
The remainder of this paper is structured as follows.Section 2 provides a literature review.Section 3 discusses the methodology of the study.Section 4 presents the results.Section 5 presents discussion and implications.Section 6 presents conclusion, limitation and future avenues.

Partnerships in humanitarian operations
As argued by several authors, partnerships between the business sector and the humanitarian sector may help improve humanitarian response.In many situations, HOs serve as intermediaries, connecting the resources provided by donors, including BCs, to the beneficiaries in need (Burkart et al., 2016;Thomas and Fritz, 2006).They are responsible for implementing programs, coordinating relief efforts and ensuring the efficient and effective delivery of aids.BCs can be engaged in the humanitarian sector in two ways: as sole donors or as partners (Thomas and Fritz, 2006;Nurmala et al., 2018;Guo and Xu, 2021).As donors, BCs often engage in corporate social responsibility initiatives and philanthropic endeavors, including donations to humanitarian causes (Lichtenstein et al., 2004;Hiller, 2013).Beyond their role as donors, BCs can also be considered external stakeholders that can partner with HOs and offer significant contributions in the humanitarian sector (Thomas and Fritz, 2006;Nurmala et al., 2018;Guo and Xu, 2021).
Distinct from the business sector, the objective of the humanitarian sector is to save lives, to meet human basic needs and to minimize human suffering (Van Wassenhove, 2006;Beamon and Balcik, 2008;Kov acs and Spens, 2009).Humanitarian operations serve both fast-onset disasters and slow-onset disasters and both natural disasters and man-made disasters (Van Wassenhove, 2006;Cozzolino and Rossi, 2012).The major actors in humanitarian operations are humanitarian agencies (Oloruntoba and Gray, 2009).Other significant players are beneficiaries, donors, governments, militaries and the media (Oloruntoba and Gray, 2009;Balcik et al., 2010).Humanitarian operations consist of three phases: a preparation phase, a response phase and a rehabilitation and reconstruction phase (Van Wassenhove, 2006;Beamon and Balcik, 2008).The preparedness phase is at the heart of humanitarian operations (Kov acs and Spens, 2009) as disaster impacts can be significantly reduced with appropriate disaster preparedness.
In the humanitarian sector, partnerships with the business sector are expected to facilitate the transfer of not only cash and products but also of knowledge, skills and expertise from the business sector to the humanitarian sector (Van Wassenhove, 2006;Beamon and Balcik, 2008;Oloruntoba and Gray, 2009;Nurmala et al., 2017).Several forms of partnerships between HOs and BCs have been suggested.Thomas and Fritz (2006) define four types of relationships between the humanitarian and business sector: single-company-philanthropic-partnerships, multi-company-philanthropic-partnerships, single-companyintegrative-partnerships and multi-company integrativepartnerships. Haigh and Sutton (2012) categorize relationships between the humanitarian sector and the business sector into four generic categories: philanthropic, strategic, business and political.Samii (2008) categorizes relationships between business and humanitarian actors into four different types: localized partnerships, strategic partnerships, brokered partnerships and cross-cutting partnerships.Binder and Witte (2007) categorize the engagements of the business sector in the humanitarian sector into three categories: single-company engagements, partnerships and meta-initiatives.Balcik et al. (2010) define that the relationships between HOs and BCs can involve single/multiple HOs or single/multiple BCs, and the commitment between the two sectors can be either philanthropic or commercial.
Appropriate logistics is key in timely and efficient response activities in the humanitarian sector.Partnerships on this matter between HOs and BCs can support response activities.Nurmala et al. (2018) define partnerships between the humanitarian and business sector in humanitarian logistics as a cross-sector partnership between an HO and a BC, mostly focusing on managing humanitarian logistics and with the mutual objective of expanding the performance of humanitarian operations for the interest of the beneficiaries.These partnerships need to meet at least three criteria: the existence of a mutual perspective to expand the performance of humanitarian operations for the interest of beneficiaries while paying respect to the individual goals of participating organizations, the existence of equal joint-decision-planning and decision-making among participants and the existence of resource sharing among participants.The resources exchanged in humanitarian logistics partnerships may include tangible sources (cash, people, products, etc.) and intangible sources (knowledge, skills, etc.).Nurmala et al. (2018) empirically investigated key dimensions for humanitarian-business partnerships in humanitarian logistics.Among these aspects, resources contributed between parties have been one of the most exposed in humanitarian calls (Nurmala et al., 2018), which underlines the importance of this research.

Donation preferences to humanitarian operations
Cash donations, time donations and in-kind donations are common types of donations (Liu and Aaker, 2008;Michel and Rieunier, 2012;Ülkü et al., 2015).Liu and Aaker (2008) and Michel and Rieunier (2012) argue that time donation involves more emotional decisions compared to cash donation.More human and social capital are involved in time donation compared to that in cash donation (Liu and Aaker, 2008;Michel and Rieunier, 2012).Therefore, cash donations are considered as low involvement (Liu and Aaker, 2008).The study by Nurmala et al. (2018) shows that cash is the most flexible resource transferred to HOs.
Studies have focused on understanding factors that explain donation behavior.Some studies have focused on demographic factors, such as gender, age, religiosity, education, ethnicity and income, as predictors of donation behavior (for example, see Auten and Rudney, 1990;Lee and Chang, 2007;Ranganathan and Henley, 2008;Notarantonio and Quiqley, 2009) However, they have faced criticism regarding the method used, the demographic-related predictor approach.For example, this approach does not explain the motive why people choose to donate (Smith and McSweeney, 2007), and a clear construct and theoretical framework are missing (Greenslade and White, 2005;Knowles et al., 2012).Studies on donation behavior have also revealed the importance of individual psychology on consumer decision-making (see Hysenbelli et al., 2013;Paramita et al., 2020).Examples of psychological factors that affect donation behavior are the perception of coldness (vs warmth), perception to community/celebrity anchor and happiness, which are argued to influence the donation behavior (Hysenbelli et al., 2013;Wymer and Drollinger, 2014;Morin et al., 2015).
Previous studies have identified factors that contribute to donor preferences with regard to humanitarian-business partnerships in managing humanitarian operations 2.2.1 Brand (brand awareness and brand image) of the collaborating business corporation In the business sector, studies have widely discussed the significant impact of brand awareness and brand image and other brand-related contexts as determinants of behavioral intentions such as the intention to purchase (Bottomley and Doyle, 1996;Keller and Lehmann, 2006).For customers, the brand promises a certain level of quality, risk reduction and trust (Keller and Lehmann, 2006).Research from Gupta and Pirsch (2006) shows that a corporate brand affects purchasing decisions on cause-related marketing.As such, brand awareness and the image of the collaborating business partner are expected to play a role in the intention to donate in support of a partnership.

Brand (brand awareness and brand image) of humanitarian organization
HOs need to focus on developing their brand to enhance their overall value (Sargeant et al., 2007;Paço et al., 2015).In fact, several studies have demonstrated that brand management holds the same significance in the humanitarian sector as in the business sector.The value of an HO's brand encompasses its identity, awareness, image and reputation (Paço et al., 2015).Several studies indicate that brand awareness and brand image have a positive impact on donation intentions (Michel and Rieunier, 2012;Paço et al., 2015;Huang and Ku, 2016).Additionally, while reputation is not synonymous with a brand, it also influences donors' interest in contributing (Ettenson and Knowles, 2008).Just as a strong brand can help convey the message of an HO, a solid reputation also helps in effectively communicating its mission (Cretu and Brodie, 2007;Ettenson and Knowles, 2008;Dubey et al., 2019).Brand awareness and the image of the HO are therefore expected to affect donation behavior.

Partnership strategy
Donors care about the impact of their donations (Ribar and Wilhelm, 2002;Caviola et al., 2014;Michaelidou et al., 2015).The concept of altruism could explain why individual donors will select upcoming humanitarian operations based on the opportunity of success.Donors want their donations to have significant impacts (Ribar and Wilhelm, 2002).In the humanitarian sector, the success of humanitarian operations depends on many aspects, including the ability to cope with the very agile situation (Van Wassenhove, 2006;Chandes and Pach e, 2010), the knowledge of the disaster and local situation (Kusumasari and Alam, 2012), high adaptation to the victims' needs (Chandes and Pach e, 2010) and the ability to monitor and evaluate within the context of emergencies (Beamon and Balcik, 2008).When it comes to partnerships between an HO and a BC, the success of humanitarian operations will also depend on how the two manage their partnership strategy to make the contribution of each party more effective (Nurmala et al., 2018).One consideration in these kinds of partnerships is the type of resources that can be contributed by HOs and BCs (Nurmala et al., 2018).As such, the type of partnership strategy is expected to play a role in donation behavior.

History and experience
Donors tend to donate to HOs that have a successful history of helping the needy (Green and Webb, 1997;Laufer et al., 2010).A longer partnership can lead to better relationship management between two parties (Balcik et al., 2010;Rueede and Kreutzer, 2015;Bealt et al., 2016).A longer partnership also can lead to better swift trust between organizations (Lu et al., 2018;Dubey et al., 2019) and better performance of a partnership (Rueede and Kreutzer, 2015).The likelihood of success of upcoming partnerships between an HO and a BC is thus affected by the history of partnerships between two parties, and we contend that this is an important aspect influencing donation behavior.

Report and disclosure
Information and accountability are relevant factors to build donors' trust and loyalty (Connolly and Hyndman, 2013;Haski-Leventhal and Foot, 2016).Donors perceive organizations that disclose information as more professional and as possessing better management capabilities; thus, donors will appreciate the disclosure and report provided by HOs.Report and disclosure will send a signal of accountability and transparency (Connolly and Hyndman, 2013;Haski-Leventhal and Foot, 2016).Both will reduce information asymmetry and will increase the expected efficacy of a donation (Zhuang et al., 2014;Haski-Leventhal and Foot, 2016).As such, reporting and disclosing information on partnerships will have an effect on donation behavior.
To summarize, this section discusses how partnerships between HOs and BCs are expected to improve the performance of humanitarian logistics and to explore factors that affect donor preferences toward humanitarian-business partnerships in managing humanitarian operations.In the next section, we will discuss the methodology used in this study.

Choice-based conjoint analysis
In this study, we used an exploratory research approach to investigate what the information provided in a call for donations could best look like.Conjoint analysis has been used more often in exploratory studies that aim to explore and validate attributes of goods and services (Kucukusta and Denizci Guillet, 2014;Rhee et al., 2016;Kim, 2018).We used CBC analysis, which is a specific type of conjoint analysis (Orme, 2000), which is a statistical market research technique used to understand customers' choices and determine their preferences for different product attributes (Orme, 2000).In CBC, respondents are presented with choice sets where each set includes multiple alternatives, each defined by specific attribute levels.Respondents are asked to select their preferred alternative from each choice set.By analyzing the choices made by respondents, CBC analysis estimates the relative importance of different attributes and their levels in driving consumer preferences (Orme, 2000;Miller et al., 2011).In our study, we adopted CBC analysis in the humanitarian context to understand how individual donors make choices and determine their preferences for different attributes of humanitarian operations before they allocate their donations.
There are several advantages of using CBC for a study aiming to understand respondents' choices and preferences for different attributes of products.First, CBC analysis helps researchers observe realistic decision-making.CBC replicates real-world decision-making scenarios by presenting respondents with a series of choice sets that resemble actual purchasing situations.This method captures the trade-offs and decision-making processes that occur in the real world (Halme and Somervuori, 2013;Braun et al., 2016).Furthermore, CBC analysis helps identify the relative importance of different attributes and levels that influence respondents' choices (Gustafsson et al., 1999;Braun et al., 2016).By understanding which attributes have the most significant impact on decisionmaking, organizations can prioritize their product development, marketing and pricing strategies accordingly.In addition, CBC analysis is a prominent method widely used in the design process for both goods and services (Gustafsson et al., 1999;Pullman et al., 2002;Kim, 2018).Although there are discussions on the artificial nature of the choice task in conjoint analysis (Jaeger et al., 2001), Miller et al. (2011) and Braun et al. (2016) argue that utility values revealed by conjoint analysis accurately reflect participants' preferences.
We adopted the steps for using conjoint analysis in the design process proposed by Gustafsson et al. (1999) and Pullman et al. (2002).We used hierarchical Bayes (HB) estimation to analyze the results of the conjoint analysis.HB estimation is among the most widely adopted techniques for CBC analysis (for example, see Voss, 2013;Maehle et al., 2015;Braun et al., 2016).Distinct from the multinomial logit model techniques, HB is based on the premise that respondents are not homogenous in behavior (Burda et al., 2008;Orme, 2000).CBC-HB allows for heterogeneity at the respondent level by specifying different utilities for each respondent and reduces the independence of irrelevant alternatives (Johnson and Orme, 2003;Solgaard and Hansen, 2003;Burda et al., 2008).The model comprises two levels: the first level corresponds to the responses of all respondents, and the second level corresponds to the individual specific behavior of respondents (Burda et al., 2008;Orme, 2000).This enables us to come up with more accurate estimates in the simulation phase as a simulation using aggregate or clustered models is likely to lead to biased results (Johnson and Orme, 2003).
In our model, y i is a discrete-choice variable for individual i and m is an alternative in the total set of alternatives M shown for a given choice task c i .Each individual i associates an alternative k with a latent utility U ik .Individual i will choose alternative k only if the utility of alternative k exceeds the individual utilities of all other available alternatives in a given choice task c i : U ik ¼ max (U i1 , U i2 , U iM ).Then, the probability of individual i choosing alternative k in a given choice task of c i is (1): Where x k is a row vector for the attribute levels of an alternative k and bi represents the column vector of an unknown individual-level parameter (part-worth utility).
The relative importance of an attribute measures the impact of an attribute on the overall utility of the chosen alternative.The relative importance is represented as a percentage.As an example, assume that n is an attribute in a subset of attributes of N, n [ {1,2,. .., N} with a number of levels p.The relative importance of RI in of attribute n [ {1,2,. .., N} as perceived by respondent i is (2):

Experiment design
As discussed above, we conducted two different experiments.
In the first experiment, we examined donors' preferences toward different designs of partnerships between an HO and a BC in managing a humanitarian operation.In the second experiment, we examined the extent to which donors' preferences align with their actual behaviors.Appendix details both experiments.

First experiment: Donors' preference over humanitarianbusiness partnership
The first experiment described an urgent call for individual donations to a humanitarian operation supported by an HO and a BC.The scenario focused on relief efforts after the 2015 earthquake in Nepal.We chose the event of the Nepal earthquake because the natural disaster was relatively recent to the time of experiment and generated a large amount of attention, which maximized the chance that our respondents were familiar with the event.We included information about the date of the disaster, the scale of the disaster and the number of victims in the scenario.We set up our scenario in line with the humanitarian-business partnership framework proposed by Nurmala et al. (2018).The partnership between the HO and BC was assumed as a dyadic collaboration as this type of collaboration is most common among humanitarian-business partnerships (Nurmala et al., 2018).The type of disaster assumed is a natural disaster since partnerships between HOs and the business sector are more common for natural disasters (Nurmala et al., 2018).Furthermore, most donors prefer to donate to natural disasters rather than to man-made disasters, because they think victims of natural disasters can be blamed less for the cause of the disaster (Zagefka and James, 2015).In the scenario, we stated that the organizations worked in partnership during the response phase and that the HO collected donations from worldwide for use in the earthquake in Nepal.This experiment included three parts and was performed using CBC package Sawtooth version 8.1.In the first part, respondents were asked to complete demographic questions.Then, in the second part, respondents were asked to rate their awareness and their perception toward the reputation of selected HOs and BCs.In the third part, respondents were given a set of CBC choice tasks with alternatives.We used Sawtooth 8.1 to create a randomized experimental design with minimal overlap, level balance and orthogonality (Chrzan and Orme, 2000).Each respondent received 12 choice tasks with each choice task containing four alternatives, which was a set that was in line with recommendations from (Johnson and Orme, 2003).We did not prohibit any combination of attributes to maximize the compounding utility estimation (Chrzan and Orme, 2000;Johnson and Orme, 2003).

Second experiment: Association between donors' preference and behavior
The second experiment described a donation calling for an ongoing joint humanitarian operation between an HO and a BC for the victims of the eruption of Mount Agung on the island of Bali (2017).To examine the association between donors' preferences over partnership designs and actual donation behavior, we divided our respondents into three groups.All groups were provided with the same brand of HO, brand of BC, partnership history and presence of disclosure and report, but different partnership strategies.For this experiment, we selected a well-known HO (International Federation of Red Cross and Red Crescent Societies, IFRC) and a well-known BC (Unilever), used a partnership history of one year and assumed the presence and disclosure of a report.For the partnership design, we provided three different partnership strategies: BC is donating money, BC is donating service (staff and facilities) and BC is donating products.
In this experiment, we gave every respondent e5 and asked each respondent whether they wanted to keep it or donate (a part of) that amount to a given partnership scenario (including not donating any money).We used such a real-money experiment as this method has been suggested as more accurate in predicting behavior compared to play-money experiments (Moser and Raffaelli, 2014).Respondents were entitled to keep for themselves the amount they decided not to donate.A record was kept of actual donations made, and amounts not donated were transferred to individual bank accounts of respondents.The individual's bank account information was requested after the respondent decided to donate and everyone received the amount they wished to keep in their bank account.The amount donated by the respondents was transferred to HOs in Indonesia.

Respondents
We collected data from a sample of 97 bachelor students studying business administration at a Dutch university.These students participated in both experiments consecutively in exchange for course credit.All participants fell within the age range of 18-30 years.Among the participants, 48% were female and 52% were male.The experiments were conducted in an isolated computer laboratory room using an online survey.Utilizing students as respondents offered several advantages for this type of experiment.First, they are frequently exposed to real donation advertisements.Second, they are less likely to be influenced by their own donation behaviors and experiences in the past.Third, students are familiar with online experiments and are able to allocate their time and presence in an isolated computer laboratory, which is a requirement for this type of experiment.

Attributes operationalization
In the first experiment, we used CBC analysis to analyze the preferences of respondents toward the different designs of partnerships between HO and BC in managing humanitarian operations.We included five attributes in our conjoint analysis (see Table 1).

Humanitarian organization brand
We used the HOs' names as the representation of their brand.We included one HO under the United Nations umbrella (World Food Program), one international non-governmental organization (Oxfam, because this is a well-known organization in The Netherlands and our respondents are located in The Netherlands) and the IFRC.

Business corporation brand
We used the BCs' names as the representation of their brands.We selected three internationally oriented BCs in different industries that have histories in forming partnerships with HOs.First, we used Unilever because it is a well-known international company in The Netherlands that has supported some international humanitarian efforts (Nurmala et al., 2018).Second, we selected DHL as a multinational logistics service provider well-known for and active in supporting multiple humanitarian efforts (Balcik et al., 2010;Nurmala et al., 2018).Third, we used American Airlines as a multinational airline well-known for their involvement in providing air transport for several humanitarian efforts (Nurmala et al., 2018).

Partnership strategy
We defined three types of partnership strategies: providing cash, providing services (staff and facilities) and providing products.For services, we further operationalized this as transportation service as this type of service is the most often delivered by the business sector during the response phase of humanitarian operations (Nurmala et al., 2018).

Partnership history
We defined three levels for the attribute of the history of partnership between HO and BC.The first was a partnership with more than five years of partnership history.The second was a partnership with only one year of partnership history.The third level was a partnership without a partnership history.

Presence of disclosure and reporting
We set up two levels for the presence of disclosure and report.The first was without disclosure and report, and the second was with disclosure and report to donors at the end of the year.

Control variables
At the beginning of the experiment, we asked respondents if they were aware of the 2015 earthquake of Nepal.If they said "yes", they could continue to the next steps.Next, we asked respondents for their awareness and perception of the reputation of the HOs and BCs.For the awareness, we examined if respondents recognized the brand name of the respective HOs and BCs.To this end, we first asked a categorical question (yes/no) whether respondents recognized the brand of HOs and BCs.Next, we used two questions using six-point Likert scale to identify the degree to which the respondents are aware of the HOs and BCs and to check the reputation of the HOs and BCs in the eyes of respondents (1 for very poor and 6 for very good).We used a six-point Likert scale as it will encourage respondents to choose between options (Cassia et al., 2017;Van Helvoort et al., 2017).The result of one-way ANOVA shows differences in the respondents' awareness of BCs [F(2, 261) ¼ 35.067, p < 0.001].The post hoc test shows that the level of donors' awareness of American Airlines is significantly lower than those of Unilever and DHL.The result of a one-way ANOVA test identifies differences in the reputation level of BCs [F(2, 261) ¼ 18.450, p < 0.001].The post hoc test indicates that the reputation level of American Airlines is significantly lower than those of Unilever and DHL.In addition, for BCs, the Pearson correlation coefficients imply that the familiarity level and the reputation level of the organizations correlate significantly.
Furthermore, it can be observed that in general, respondents are more aware of BCs compared to HOs, which is shown by higher percentages of respondents that recognize the name of BCs compared to those of HOs.This result supports Pope et al. (2009), Paço et al. (2015) and Huang and Ku (2016), who show that there is still a lack of brand recognition in the nonprofit sector compared to that in the for-profit sector.

The choice-based conjoint analysis 4.2.1 The part-worth utilities
This section reports on the first experiment.For each of the 97 respondents, we calculated the relative individual-level partworth for five attributes.Figure 1 shows the individual-level part-worth in gray lines.We highlight the aggregate-level partworth in bold-black lines.The aggregate-level part-worth can also be seen in Table 3.We obtain a mean root likelihood of 0.627, which is approximately 2.5 times better than the predictability of the responses using uninformative utilities (See Orme and Howell, 2009;Braun et al., 2016).
The CBC-HB output calculates the part-worth utilities of levels of attributes as a relative part-worth value compared to the attributes' reference levels.The multinomial logistic regression creates dummy variables and uses the last level of each attribute as a reference level for each attribute.A utility of zero is assigned to the reference level.Consequently, all other part-worth utility values are analyzed relative to zero.
Figure 1(a) shows the relative individual-level and aggregatelevel part-worth utility for the attribute of brand of HO, with the IFRC serving as a reference category.From Figure 1(a) and Table 3, we notice that the average part-worth value of IFRC is higher than those of WFP and Oxfam.This result is in line with the awareness and reputation results presented in Table 2 and shows that potential donors are more likely to donate to a humanitarian operation operated by IFRC than to operations carried out by WFP or Oxfam.The part-worth value explains that when donors find that the upcoming humanitarian operation will be operated by WFP instead of IFRC, it will decrease the logodds of donating by 0.86 points while holding all other variables in the model constant.If donors find that the upcoming humanitarian operation will be operated by Oxfam instead of IFRC, it will decrease the log-odds of donating by 1.43 points while holding all other variables in the model constant.Combined with the awareness and reputation results presented in Table 2, these results suggest that donors are more likely to donate to an HO that they are more aware of and have a better reputation.
A similar situation is seen in Figure 1(b).Respondents tend to donate to humanitarian operations that involve the participation of a BC that they are more aware of and have a higher reputation.Donors are more likely to donate to humanitarian operations that are supported by Unilever and DHL compared to those supported by American Airlines.If donors find that the upcoming humanitarian operation will be supported by Unilever instead of American Airlines, it will increase the log-odds of donating by 0.56 points.If donors find that the upcoming humanitarian operation will be supported by DHL instead of American Airlines, it will increase the log-odds of donating by 0.51 points, holding other variables in the model constant.However, in contrast to Figure 1(a), the results in Figure 1(b) show a more even distribution because there are fewer differences between the part-worth values of the BCs, indicating that respondents care less about which BC will support the upcoming humanitarian operation compared to which HO initiates the humanitarian operation.
Table 3 and Figure 1(c) show that individual donors prefer to donate to humanitarian operations that are supported by services (staff and facilities) and/or products from BCs, compared to those receiving cash from BCs.If donors find that the upcoming humanitarian operation will be supported by cash donations, the log-odds of donating decrease by 2.23 points, compared to those supported by product donations.If the upcoming humanitarian operation will be supported by services from BC instead of by donation of products, the logodds of donating will increase by 0.21 points.
Table 3 and Figure 1 (d) show that donors are more likely to donate to humanitarian operations when the HO and BC have experience and history in managing the partnership.If donors find that the HO and BC have a one-year partnership history instead of having no history at all, the log-odds of donating will increase by 2.27 units.If donors find that the HO and BC have more than five years of partnership history instead of having no history at all, the log-odds of donating will increase by 2.43 units.A longer partnership history is thus preferable for donors.

Figure 1 Individual-level and aggregate-level part-worth utility profiles
The results also show that donors value information disclosure and report.If the HO provides disclosure and report to their donors, it will increase the log-odds of donation by 2.42 points in comparison with that in the absence of disclosure and report.

Choice-based conjoint relative attribute importance
Based on the individual-level part-worth values, we calculated the relative importance of attributes for each individual.This is computed by taking the part-worth utility range for each factor and dividing it by the sum of the utility range for all factors [see Equation ( 2)].When summarizing the relative importance of a group of respondents, it is best to do the calculation by first computing the relative importance for each individual, and then the overall average (Orme and Howell, 2009).
Table 4 shows the relative importance of each attribute.The result shows that the attribute "partnership strategy" exhibits the highest relative importance of 27.24%, followed by "partnership history" (23.94%) and "report and disclosure" (20.72%).Surprisingly, the brand of HO is only positioned as fourth in importance (17.51%).Respondents thus perceive which brand of BCs that support the humanitarian operation as the least important attribute when deciding whether they want to donate or not.

Choice-based conjoint attribute switching and share of preference
HOs and BCs could consider different partnership strategies, invest in longer horizons of partnerships or be more transparent to attract donors.In this section, we provide three simulations to show the impact of switching between various options.The first simulation is to see the value of when the HO and BC change their partnership strategy from the BC delivering products to delivering cash or services.The second simulation is to see the value of having long-term partnerships.The third simulation is to identify the added value of information disclosure and report.
Table 5(a) shows the base for the three simulations.For HO, we use "IFRC", for BC, we use "American Airlines", for reference partnership strategy, we use "donating products", for partnership history, we use "no partnership history" and for the disclosure and report attribute, we use "with disclosure and report".
Table 5(b) shows the value of changing the partnership strategy from BCs donating products to donating cash and delivering services (staff and facilities).Donors prefer the donations of products and services from BCs to HOs more than cash.In our simulation, if the contribution of American Airlines to IFRC switches from donating products to donating cash, the share of relative preference of this partnership in the eyes of donors will decrease by À22.47%; contrastingly, if the contribution of American Airlines to IFRC switches from donating products to donating services (staff and facilities), then the share of relative preference will increase by 4.89%.Table 5(c) shows that donors value longterm partnerships more than ad hoc partnerships.In our example, the share of relative preference will increase by 26.94% if donors find that a humanitarian operation will be managed by HOs and BCs that have at least one year of partnership history, rather than having no history at all.The result also shows that donors perceive reports and disclosure as valuable [Table 5(d)].In our simulation, it is shown that if the upcoming humanitarian operation provides no report and disclosure, its share of relative preference will decrease by 29.99% compared to that obtained when report and disclosure are provided.

Donors' preferences over partnership designs vs Actual donations
The result of the second experiment, the real-money experiment, shows that 82.5% of respondents decided to donate the endowment of e5 fully or partly, rather than keep the amount completely for themselves.Among those who donated, the majority of them donated their full endowment of e5 (88.75%), while others decided to keep some part of the money.We applied statistical tests at a significance level of p < 0.05 to test the impact of the treatment.First, we applied logistic regression to test whether the partnership strategy affects donors' decision to donate.The result of logistic regression shows that donors' decisions to donate are not affected by the partnership strategy [p ¼ 0.312].Next, we applied one-way between-subjects ANOVA to compare the effect of the three partnership strategies on the amount donated by the participants in the experiment.The result shows that there is no significant effect of the partnerships strategy on the amount donated [F(2.94)¼ 1.472, p ¼ 0.235].We also applied Tobit regression to estimate if there is any linear relationship between the part-worth value of a partnership strategy identified in stage one of the experiment and the amount donated to that strategy in stage two of the experiment.We used Tobit regression because our dependent variable has right censoring value of e5.The result from the Tobit regression shows that the part-worth value does not significantly affect the amount donated [p ¼ 0.230].This means that we do not find evidence that the preference of donors for a specific partnership strategy option is reflected in the actual amount of money donated to that strategy.

Discussion
By using CBC analysis, the results of the study show how donors perceive and value partnership design between HOs and BCs.Given the hesitancy of some HOs to initiate partnerships with BCs (Van Wassenhove, 2006;Oloruntoba and Gray, 2006;Thomas and Fritz, 2006), our research findings suggest that the positioning of BCs' contributions in humanitarian-business partnerships can serve as a significant asset for HOs in their efforts to attract individual donors.The results of the study show how donors value certain options over others when evaluating the attributes of humanitarian-business partnerships.
The findings indicate that donors value the contribution of BCs to humanitarian operations more when BCs donate staff time, facilities and products instead of only cash.So far, cash donations have been the dominant type of engagement between HOs and the business sector (Nurmala et al., 2018).The results also reveal that donors value partnerships with a long history, as well as the existence of disclosure and reporting on past performance.In fact, our study demonstrates that describing the partnership strategy, the partnership's history and the presence of disclosure and reporting are more critical to attract potential donors than which organizations (HOs or BCs) are involved in the partnership.This may indicate that donors prefer to base their donations on actual facts rather than brand-or image-related impressions.Regarding HO brands and BC brands, although they are not the main factors that affect the intention to donate, donors tend to choose humanitarian operations that involve HOs and BCs of which they have more awareness.The study also indicates that donors' pre-existing valuing of different partnership characteristics in their intentions to donate is not always aligned with how donors decide to actually donate.Our study shows that the resulting preferences based on intentions do not align well with the actual donations.This could point to an intention-behavior gap, a topic that has been widely investigated in the business sector (Padel and Foster, 2005;Carrington et al., 2010) but remains poorly investigated in the humanitarian sector.

Theoretical implications
Previous studies suggested HOs to initiate collaborations with various stakeholders, including business entities (Thomas and Fritz, 2006;Kov acs and Spens, 2009;Kaneberg, 2018;Müller-Stewens, 2019).The need for these partnerships arises from the growing complexity of humanitarian operations, making it challenging for these organizations to manage independently, along with needs to gain insights from BCs (Van Wassenhove, 2006;Thomas and Fritz, 2006;Nurmala et al., 2017).Despite the widely promoted benefits of such partnerships, research suggests that these initiatives have not been fully optimized (Nurmala et al., 2018).A significant obstacle hindering HOs from engaging with the business sector is the fear of negative media coverage that could tarnish both the reputation of BCs and the perception of their independence, potentially impacting support from individual donors (Thomas and Fritz, 2006;Kov acs and Spens, 2009;Nurmala et al., 2018).The current literature lacks in understanding how HOs can effectively navigate this situation to maximize their potential for initiating partnerships with BCs while still maintaining support from individual donors.
The results show that a strategically positioned cross-sector partnership can be a valuable asset for HOs in attracting donors and can profoundly influence individual donor contributions.
The key factors influencing donors include partnership strategy, historical performance and transparency and disclosure.Regarding partnership strategy, our study reveals that donations of services and products from BCs to HOs are more appealing to individual donors than cash contributions.This finding aligns with the concept of altruism in donation behavior (Ribar and Wilhelm, 2002;Bekkers and Wiepking, 2011), suggesting that when donors witness BCs contributing cash to HOs, they may be less inclined to donate money directly.Individual donors tend to support humanitarian operations which they believe that they will have a significant impact.Interestingly, most published partnerships between HOs and BCs in humanitarian logistics primarily involve cash contributions rather than services or products (Nurmala et al., 2018).
The study reaffirms previous research findings ( Van Wassenhove, 2006;Beamon and Balcik, 2008;Michel and Rieunier, 2012) indicating that donors prefer to support future humanitarian operations with a high likelihood of success.Our results demonstrate that donors are more likely to donate to humanitarian operations that have bigger opportunity to succeed.When BCs contribute services to a partnership, they implicitly contribute to the transfer of skills and knowledge (Oloruntoba and Gray, 2006;Beamon and Balcik, 2008;Nurmala et al., 2017).Partnerships with a longer partnership history could be perceived to have a bigger opportunity to succeed as they have better management, swift trust and coordination (Rueede and Kreutzer, 2015;Bealt et al., 2016;Dubey et al., 2019).The study shows that individual donors prefer to contribute to humanitarian operations that offer transparency through disclosure and reporting, aligning with the findings of Zhuang et al. (2014) and Haski-Leventhal and Foot (2016), who link disclosure and reporting to a higher intention to donate.Given the importance of retaining existing donors and attracting new ones (Aldashev and Verdier, 2010;Caviola et al., 2014), understanding the factors influencing donors' decisions to contribute and their preferences among donation options becomes increasingly essential.
This study also shows that actual behavior of donors may not necessarily align with their preferences.Although the study shows that there is an effect of the partnership designs on the odds to donate, this is not reflected in donors' real behavior.Most of our respondents decided to donate their full endowment, regardless of the different types of partnership design.Our study furthermore shows that there is no significant effect of different partnership designs on donors' decision behavior and on the amount donated.Similar to research in consumption behavior (Chandon et al., 2005;Padel and Foster, 2005;Carrington et al., 2010) and in physical activity (Rhodes and Dickau, 2012) and daily activity (Sainsbury et al., 2013), our study provides at minimum a first indication that the relation between stated and actual donation behavior appears to be rather weak.

Practical implications
Understanding how donors value the partnerships between HO and BC is relevant for those who devise marketing strategies by which HOs attract donations.The results of our study can help HOs to navigate the complex landscape of engaging with BCs.Our study has a number of ramifications.First, the findings show that donors prefer objective data on projects or plans over more subjective impressions, such as the general image of an organization.Showing plans and providing after-the-fact reports that present actual findings of what has been done with the money donated will build trust among donors and therefore will increase the intention to donate.Second, if an HO aims to work with the business sector, it is recommendable to invest in longstanding relationships.In relation to this, the actual content of the partnership becomes significant.This implies that if HOs intend to retain the support of individual donors, they may better agree with BCs to provide products and services to the partnerships rather than just cash (which is often opposite to current practices).Understanding these situations will help HOs to design partnerships that better support attracting donations.

Conclusion
In this study, our aim is to gain a comprehensive understanding of donor preferences regarding different designs of humanitarianbusiness partnerships in managing humanitarian logistics.Additionally, we aim to examine whether these stated preferences align with the donation behavior exhibited by these donors.Our research outcomes reveal that the design of partnerships between HOs and BCs in managing humanitarian logistics provides considerable strategic value for HOs.The strategic approach to partnerships, historical performance in partnerships and transparency and disclosure are the most prominent factors.It is essential to acknowledge that our research also highlights a notable disparity between donors' expressed preferences and their actual donation behaviors.

Strengths, limitations and avenues for future research
The combination of the use of empirical analysis on the preference of donors and simulation to evaluate the impacts of different partnership architectures is a vital strength of this study.We acknowledge that our research also has limitations, which may be addressed in future research.First, the majority of the participants in the survey were students living in The Netherlands.Culture plays a significant role in determining the behavior in making a decision (Carter et al., 2010;Yates and de Oliveira, 2016).It will be interesting to replicate the study in a different cultural context and with a broader population.Second, university students are homogenous in regard with range of age and occupation.Furthermore, we studied the alignment between donors' preference and their actual donation behavior with a relatively small amount of money.Future studies involving higher amounts of money (possibly using random lottery incentive systems) could shed more light on the robustness of our findings.

Appendix. Experiment scenario Experiment 1
Aim: The aim of the first experiment is to examine donors' preferences towards different designs of partnerships between an HO and a BC in managing the humanitarian operations.
This experiment was divided into three parts: 1 Demographic Questions: Respondents are asked to provide information about their age and education.2 Awareness and Perception Rating: Respondents rate their awareness and perception of selected humanitarian organizations (HOs) and business corporations (BCs).
Likert scale questions are used to assess respondents' familiarity with the organizations and respondents' perception on their reputation.3 Choice Tasks: Respondents are presented with a set of choice tasks using the Choice-Based Conjoint (CBC) method.They are asked to choose the best alternatives that reflect their preferences toward the presented options.
Scenario and questions:

C2 CBCscenario
We would like you to consider the following situation: Last night, a strong earthquake measuring 8.5 on the Richter scale hit Nepal.
Nepal is located in South-Central Asia.In April 2015, a strong earthquake measuring 7.8 on the Richter scale also hit Nepal, leading to more than 5,000 fatalities.
The exact number of fatalities from the latest earthquake is not known yet.However, UNOCHA (United Nations Office for the Coordination of Humanitarian Affairs) predicts that more than 500,000 people are affected by the disaster.
Assuming you've decided to donate 50 euros out of your pocket to aid relief operations in Nepal, your goal is to make a significant contribution to the affected people.You searched the internet and found the following options for upcoming relief operations to which you can donate.
If these are the only options available, which one would you choose?CBC CBC -Choice Task (Continued)

Table 1
Attributes and levels for the CBC analysis

Table 2
Familiarity and reputation of humanitarian organizations and business corporations

Table 3
Aggregate-level part-worth utility

Table 5
Changes in share of preference