THE NONPROFIT MARKETING PROCESS AND FUNDRAISING PERFORMANCE OF HUMANITARIAN ORGANIZATIONS: EMPIRICAL ANALYSIS */NEPROFITNI MARKETINSKI PROCES I UCINAK PRIKUPLJANJA SREDSTAVA HUMANITARNIH ORGANIZACIJA: EMPIRIJSKA ANALIZA

Abstract

This paper links the fundraising success of nonprofit organizations to their marketing process. Financial and non-financial dimensions of fundraising performance are included in the study, as to demonstrate both aspects of the intended outcomes. The empirical part of the paper is based on research focusing on Croatian humanitarian organizations and employs the Structural Equation Modeling (SEM) methodology to assess the hypotheses regarding the positive influence of the nonprofit marketing activities on two dimensions of fundraising performance. In addition, the paper discusses and empirically verifies the influence of fundraising feedback on the (re)definition of marketing activities. Implications for nonprofit marketing and management practice(s) are discussed, along with recommendations for future research./Cilj ovog rada je povezati uspjesnost fundraisinga sa marketinskim aktivnostima neprofitnih organizacija. Istrazivanje obuhvaca financijsku i nefinancijsku dimenziju performansi uspjesnosti fundraisinga kako bi se prikazala oba aspekta ishoda procesa fundraisinga. Empirijski dio rada proveden je na uzorku neprofitnih organizacija u Hrvatskoj. U svrhu procjene hipoteza o pozitivnom utjecaju marketinskih aktivnosti na uspjesnost financijskih i nefinancijskih performansi fundraisinga u istrazivanju je koristena metoda modeliranja strukturnih jednadzbi (SEM). U radu se kriticki analizira i empirijski potvrduje povratni utjecaj uspjesnosti fundraisinga na (re)definiranje marketinskih aktivnosti. Dodatno su, na temelju rezultata istrazivanja, prikazane implikacije za marketinske i menadzerske prakse neprofitnih organizacija kao i preporuke za buduca istrazivanja.

Headnote

This paper links the fundraising success of nonprofit organizations to their marketing process. Financial and non-financial dimensions of fundraising performance are included in the study, as to demonstrate both aspects of the intended outcomes. The empirical part of the paper is based on research focusing on Croatian humanitarian organizations and employs the Structural Equation Modeling (SEM) methodology to assess the hypotheses regarding the positive influence of the nonprofit marketing activities on two dimensions of fundraising performance. In addition, the paper discusses and empirically verifies the influence of fundraising feedback on the (re)definition of marketing activities. Implications for nonprofit marketing and management practice(s) are discussed, along with recommendations for future research.

Key words: Nonprofit marketing process, Management, Fundraising performance, Humanitarian organizations, Croatia

1.NONPROFIT MARKETING PROCESS AND FUNDRAISING PERFORMANCE - THEORETICAL BACKGROUND

Nonprofit organizations, particularly humanitarian organizations, demonstrate the misunderstanding of the marketing concept and mostly focus on sales and promotional activities (Dolnicar and Lazarevski, 2009). Another objection to the use of marketing tools and techniques is reflected in the inadequate 'social image' of marketing, which is perceived to be an inadequate tool for the social sector, as it is primarily driven by profit motives. The marketing orientation toward research and the satisfaction of end users' needs is, often, ' conveniently' ignored, or perceived from the viewpoint of profit sector managers (Sheth and Sisodia, 2005).

The majority of nonprofit organizations are focused on their beneficiaries (users) and the satisfaction of their needs. A problem arises, if the beneficiary focus is only declarative, which is often due to the inadequate understanding of the importance of the strategic analysis, as the first step of the strategic marketing process (Andreasen and Kotler, 2008). Nonprofit organizations which implement marketing orientation are focused on all of their key stakeholders, which consequently leads to better understanding of stakeholders needs and organizations' performances (Modi, 2012).

The marketing orientation, as well as derived marketing activities, of nonprofit humanitarian organizations requires its application both to beneficiaries, and to donors, in order to avoid wasteful fundraising activities and concentrate on those, who are willing to support an organization (Sargeant and Woodliff, 2008; Srnka, Grohs and Eckler, 2003).

However, with the sudden growth of nonprofit/social sector organizations, competing for scarce resources (financial and human), the resulting competition has re-emphasized the need to target adequate stakeholder segments and establish a positioning vis-à-vis the competitors. This means that the traditional practice of emphasizing promotion and distribution in the nonprofit marketing mix becomes a 'trap' for inflexible organizations (Novatorov, 2010; Dolnicar and Lazarevski, 2009; Stater, 2009; Pope, Isely and Asamoa Tutu, 2009; Sargeant and Wymer, 2008).

Those organizations could be further limited in their management process and strategic marketing implementation, if there is a prevailing belief that a mission change would be unacceptable as it is defined in advance and cannot be changed or adapted to market needs since that would change the core of existence of nonprofit organization (Dolnicar and Lazarevski, 2009).

The last step of the strategic marketing process requires the performance to be measured and corrected (Sawhill and Williamson, 2001 ; Herman and Renz, 2004; Poister, 2003; Keating and Frumkin, 2003). This process mostly depends on organization's marketing orientation - capability to recognize, react/adapt and use all changes in organizations environment (Abdulai Mahmoud and Yusif, 2012).

There are many difficulties in measuring the success of nonprofit organizations, including the 'non-monetary character' of their performance, difficulties in assessing the mission and objectives, the multiplicity of stakeholders, etc.

However, those can be addressed by the multiple constituencies of nonprofit performance (Herman and Renz, 2004), an endeavor, which supports the notion of using the same marketing approach to address the needs of both the beneficiaries/users and donors, along with numerous other stakeholders (beneficiaries).

With regard to donors and beneficiaries, the marketing approach and planned activities should be different, but complementary. Nevertheless, author concentrates on the donor dimension of the overall nonprofit strategy and proposes a generic fundraising model and links it to the nonprofit marketing activities, which has not been done before in an adequate manner (Knowles and Gomez, 2009; Stater, 2009; Hart, 2008; Andreoni, 2006; Heinzel, 2004; Bennett, 2003).

Specifically, there is a lack of comprehensive studies, since the majority of empirical research relates to particular aspects of fundraising, especially the behavior and motives of individual donors (Sargeant and Woodliff, 2008).

As marketing function in nonprofit organizations is often executed from managerial perspective, the point of marketing orientations or satisfaction of all stakeholders' needs is jeopardized. To avoid the trap of marketing orientation misunderstanding, "classical" Kotler's (1999) approach to marketing process is used in this paper. Accordingly, marketing activities refer to analysis, planning, application and control as key components of marketing management process.

2.NONPROFIT MARKETING PROCESS AND FUNDRAISING PERFORMANCE - TOWARD A MODEL

The lack of funds is identified as a fundamental problem in the implementation of nonprofit marketing, followed by the lack of staff and basic marketing knowledge (Pope et al., 2009), which has been confirmed in the Croatian context as well (Pavicic, Alfirevic and Ivelja, 2006). On the other hand, some organizations perceive marketing to be a 'wasteful' activity and an unwanted source of expenses, regarding it as being unnecessary for the realization of objectives (Bennett, 2007).

Along with the previously mentioned traditional emphasis on price and distribution in the nonprofit marketing mix (Dolnicar and Lazarevski, 2009), it is unfortunate that research to date has not placed the analysis of fundraising performance in the nonprofit strategy context.

A review of established practices leads to the selection of the following financial fundraising indicators (Barrett, 2005): FACE ratio (fundraising expenses, administrative expenses and overall expenses ratio) and expense per collected monetary unit (Sargeant and Shang, 2010). The desired FACE ratio is usually set at the 35% level, although this established practice is viewed critically by Sargeant and Shang (2010). The definition of nonfinancial fundraising indicators is equally important, but even harder to conceptualize, because of a multitude of values and results to be achieved by diverse nonprofit organizations.

The literature includes many of those, such as: number of employees and volunteers, trust of the wider public, satisfaction of beneficiaries, quality of service, public awareness about the problem, perceived reputation of organization, clarity and acceptance of reasons for support, dedication, satisfaction and lifetime of donors (Sargeant and Shang, 2010; Andreasen and Kotler, 2008; Bennett, 2007; Bryson, 2004; Poister, 2003; Balabanis, Stables and Philips, 1997).

Once the financial and non-financial dimensions of the fundraising performance have been established, it is easy to create a conceptual model (see Figure 1), which removes the previously mentioned limitations of the existing research. Presented model is a generic one, as it concerns marketing activities and both dimensions of fundraising performance, as well as accommodates the use of feedback by means of managerial controlling.

The following hypotheses are based on the theory review:

H 1. Marketing activities of a humanitarian nonprofit organization have a positive influence on fundraising performance.

The following sub-hypotheses are based on the two dimensions of the fundraising performance:

H 1.1. Marketing activities have a positive influence on the financial dimension of the fundraising performance.

H 1.2. Marketing activities have a positive influence on the nonfinancial dimension of the fundraising performance.

There is little research on the role of feedback and the overall influence of managerial control in achieving fundraising success (Bennet and Savani, 2011; Hsieh, 2010; Dolnicar and Lazarevski, 2009; McGee and Donoghue, 2009; Sargeant and Woodliff, 2007; Bulla and Starr-Glass, 2006; Bennett, 2003). The idea of feedback, certainly, calls for the re-definition of marketing activities, if the expected fundraising performance has not been achieved. Some partial research has been conducted so far, including a study on how the uniqueness of the fundraising environment shapes the product creation in the nonprofit sector (Stater, 2009). Once again, the author was not able to identify any studies that deal with the feedback about the generic nonprofit marketing process, based on the previous fundraising performance. This justifies the second conceptual hypothesis:

H2. Changes in fundraising performance serve as feedback for the (re)defmition of nonprofit marketing activities.

3.OPERATIONALIZATION OF RESEARCH VARIABLES

3.1.Operationalization of the nonprofit marketing variable

The operationalization of the nonprofit marketing variable by using the constructs related to its fundamental stages: analysis, planning and implementation, is based on the established theoretical foundations of the strategic marketing planning process (Sargeant and Jay, 2004; Andreasen and Kotler, 2008; Mcloughlin and Aaker, 2010; Gillian and Voss, 2013; Haig, 2005). Purpose of this research was to determine marketing impact on fundraising success in specific nonprofit organizations - humanitarian, whose practices and beliefs notably differ from other business-like organizations. Accordingly, previously mentioned components of "classical" Kotler's (1999) approach to marketing management in nonprofit organizations were used instead of a well-known and commonly used marketing orientation approach. Nevertheless, the popular market orientation scales MARKOR (Kohli, Jaworski and Kumar, 1993) and the MKTOR (Narver and Slater, 1990) were also consulted, as the stages of nonprofit marketing were operationalized.

The items introduced for the operationalization of strategic analysis included: regularity of internal and external environment analysis, taking into consideration multiple stakeholders, analysis of the existing situation and future perspectives, as well as awareness of the implementation importance. Items related to strategic marketing planning included: relationships of the analysis and planning stages, inclusion of employees and volunteers in the planning process, time dimension of planning, development of the system for measuring the implementation of the plan, taking into consideration multiple stakeholders, coordination of plans with a mission and objectives, and awareness of importance of the plan for future performance.

The operationalization of the marketing implementation stage is related to the organizational mission, objectives, segmentation, positioning and the marketing mix. Items related to the mission are modified from Bennett (2007) and the estimate of employee and management, if opposing the mission. The marketing objectives items include specificity, measurability, relevance and time determination, as proposed by Sargeant and Jay (2004). The segmentation items are somewhat modified from the measurement model provided by Srnka, Grohs and Eckler (2003) and address the existence of segmentation, as well as the segmentation criteria. Positioning is operationalized in terms of organizational perception, in terms of the specific case(s) for support for different donor segment(s) (Ibid.). They are further described by items, related to donor perception and attitudes, availability of information about the mission, objectives and case(s) for support, possible perception according to the area of nonprofit activity, beneficiaries and the perceived marketing success rate. Elements of the marketing mix are operationalized according to the conceptual foundations, as described by Andreasen and Kotler (2008) and McLeish (2011), as well as by the measurement model developed by Lai and Poon (2009).

3.2. Operationalization of the fundraising performance variable

Both financial and non-financial dimensions of fundraising were included in the operationalization. The financial dimension used a modified approach, originally introduced by Bennett (2007), including items related to: organizational income, total amount of donations, donations per donor, total expenses, administrative expenses and expense per collected monetary unit of donations. The non-financial fundraising dimension was operationalized in accordance with the theoretical foundations, proposed by Sargeant and Jay (2004) and Sargeant and Shang (2010) and consists of items related to: the number of repeated donations, the increase of donors, employees, volunteers and beneficiaries, satisfaction and dedication of donors.

3.3. Operationalization of the feedback variable

The operationalization of the feedback variable was conducted by referring to the concepts of organizational learning and control (as a part of marketing management process). The existing measurement models, related to collection, distribution and interpretation of information and organizational memory, were used (Flores et al., 2010; Dimovski et al., 2008; Lopez et al., 2005). The role of managerial control was operationalized by using the following items: regularity of control, structure of performance data collected, sharing of information, performance data availability, frequency of marketing control, undertaking corrective measures, attitudes toward corrective measures and feedback-based changes in marketing.

4.METHODS AND THE RESEARCH SAMPLE

4.1. Population and sample

The sample of Croatian humanitarian nonprofit organizations, out of the projected population of 400 organizations (as suggested by the three previously interviewed experts), was created by using the snowball sampling approach, since the official registries of nonprofit organization in Croatia1 proved to be unsuitable. Namely, they produce results for 52,659 formally registered organizations (in February 2015), although the majority of those are either inactive, or only conduct occasional activities. The 'snowball' sampling method is justified, considering that the research was carried out on a relatively small and generally unavailable population of ' high-capacity' humanitarian nonprofit organizations, realizing a significant part of their income from donations. A similar approach has been already used in regional nonprofit research (Alfirevic, Pavicic, Najev Cacija, 2014). After two calls for participation via electronic mail and additional telephone contacts, 97 completed survey questionnaires were collected, four of which were not valid. The total effective response rate in this research is 23%, which is considered an acceptable rate for research related to nonprofit organizations.

One limitation of this research lies in the relatively small number of cases (93), which is close to the lower practical limit for designing the model, consisting of five to seven manifest variables, if the 'rule of the thumb' of seven to ten cases per manifest variable in a model is used (Bentler and Chou, 1987; Macroulides and Saunders, 2006.) However, it should be added that Structural Equation Modelling (SEM) has been successfully used on even smaller samples and in specific scientific fields, such as management and marketing (Mottner and Ford, 2005; Browne et al., 2002; Gignac, 2006). The size of this sample is in accordance with Hayduk et al. (2007), who criticize the generally accepted rule on the absolute necessity for sample size to be larger than 200 cases, if SEM is to be applied.

Levene's test for homogeneity of variances and the ANOVA test for independent samples were used to test the possible existence of statistically significant differences between responses of organizations that responded immediately and those responding during later stages of the survey (for all proposed manifest variables). In the sample of 93 organizations, the first 30 were classified as the 'early' response group, and the last 30 survey questionnaires to be received were classified as the 'late' response group. Upon review of values of Levene's test for homogeneity of variances (p>0.05) and pvalues of the ANOVA test for independent samples (p>0.05), it can be concluded that there is no statistically significant difference in variance between responses of early and late interviewees and no bias, due to several 'waves' of data collection.

4.2. Measurement scales, items parceling and statistical assumptions

The values of Cronbach coefficients for manifest variables indicate a high level of internal consistency, as demonstrated by Table 1.

Considering that the reliability of measurement scales is acceptable, composite manifest variables were created by computing the mean values of items (item parceling), representing individual manifest variables. Procedure of manifest variables creation, as a mean value of belonging statement, is justified in cases when area of interest is widely defined (Hall; Snell and Foust, 1999), which is the case in this research investigating influence of marketing activities on fundraising success since influence of analysis, planning, application and control on financial and non-financial fundraising objectives is investigated without intention to identify and clarify individual components of proposed manifest variables, i.e. wide components of marketing activities or fundraising success. Item parceling (by any method) is acceptable for relatively small samples, for testing models with a higher number of parameters, whereas the method of total disaggregation (statement as indicator) would very probably lead to unacceptable fit indices (Rocha and Chelladurai, 2012). The advantage of item parcels as an indicator of latent variables is related to the reduction of the number of observed variables in the model (Coffman and MacCallum, 2005). In addition, data transformation is achieved with variables that do not follow the normal multivariate distribution (Sterba, 2011). From the theoretical/cognitive viewpoint, the research should be credible and logical, which relates both to the methods and the results. This is in line with the liberal-pragmatic standpoint, stating that the researcher should have the freedom to define indicators in models with latent variables (Rocha and Chelladurai, 2012).

Furthermore, Little et al. (2002) stress the need to consider the purpose of the research. If the objective is to understand the relations among latent variables, then statements or sets of statements are simply a tool, enabling the researcher to create the measurement model. In addition, if there is no intention to investigate the dimensionality of relations among statements within the measurement model, the item parceling is more than justified (ibid.). Considering all arguments, total aggregation as the item parceling method is used in this research. With this approach, eight manifest variables are created, calculated as the mean values of the related statements, in the range from 3 to 25 statements. Items with a low level of internal reliability, as measured by the Cronbach alpha value, were omitted from the measurement model. An overview and the description of composite manifest variables is shown in Table 2.

Considering that item parceling might improve the normality of the manifest variables' distribution, formal normality checks for eight manifest variables were conducted. Only three, out of eight, satisfy the requirements of Shapiro-Wilk's normality test. The presumption of normality is important, so that the model parameters can be estimated by using the maximum likelihood method. If the presumption of normality is not met, parametric ratings will still be unbiased and asymptotically consistent for sufficiently large samples. Taking into consideration the law of large numbers and the central limit theorem, the value of the t-test asymptotically moves closer to the value of normal distribution, even if the input variable is not distributed normally. A distribution normality check was further conducted by analysis of values of skewness and kurtosis indices for each manifest variable. It was found that the absolute values of skewness and kurtosis were within acceptable limits of -/+ 3 for skewness and -/+ 10 for kurtosis (Kline, 2011). Consequently, the use of structural equations modelling is formally acceptable.

The check for univariate outliers was conducted on the basis of z-value, i.e. the difference between the measurement results and the arithmetic mean for all measurements, expressed in standard deviation units, with the limiting value of 3.29 (Field, 2009). The presented results of analysis indicate the existence of only two separate cases of outliers in manifest variables FB1 and FB2 (individual cases 13 and 8), which were, thus, excluded from further analysis. Checks of bivariate and multivariate multi-collinearity were conducted, by analyzing the tolerance threshold and VIF (variance inflation factor) of manifest variables. The results show that there is no bivariate multi-collinearity among the variables, while the tolerance threshold and VIF indicators show there is no multivariate multi-collinearity problem (tolerance threshold is higher than 0.20, while the VIF value is lower than 10).

5.RESEARCH RESULTS

The SEM model, used to analyze the relationship between marketing activities and the fundraising performance, is illustrated by Figure 2. There are six manifest and two latent variables, with marketing activities (MA) representing the exogenous latent variable, while fundraising performance/success (FS) is the endogenous latent variable. Factor loadings are determined by using the maximum likelihood method. Errors associated with manifest variables represent measurement errors, while those associated with latent endogenous variables (residual errors) represent inaccuracies of forecasting the endogenous factors by using the exogenous factors (Byrne, 2010). Measurement errors and residual errors are labeled as e1, e2,...en.

Results show that 13 parameters were estimated in a model, while the chisquare-value is 15.338 with eight degrees of freedom. The value of the chisquare/df ratio is satisfactory (1.917) (Schermelleh-Engel et al., 2003) and the pvalue is higher than 5%, which confirms that the model is appropriately specified, i.e. statistically significant. Parameter estimates, i.e. the suitability of the model, is described by the value of goodness-of-fit indicators, as demonstrated by Table 3.

RMSEA does not fall within the satisfactory limits for acceptance of the proposed model and the recommendations for model modification were checked, as shown in Table 4.

The recommendation to create a correlation between measurement errors e3 and e4 and el and e6 could be implemented. Considering that errors show the unexplained part of variance (accidental errors and errors caused by unknown/unexplained factors), it is necessary to theoretically review the effects of acceptance for the proposed modification.

Each modification of the proposed model, i.e. changes of connections between parameters, as well as the addition of new connections, has to be justified by strong theoretical or practical reasons (Byrne, 2010). In some cases, if a correlation of measurement errors is recommended, it can be justified, since 'forcing' disconnectedness of large measurement errors is rarely appropriate for actual data (Bentler and Chou, 1987). The correlation of measurement errors can be ascribed to the overlap of items, if similar or equal ones are repeated in the questionnaire (Byrne, 2010), or if they are perceived as such by respondents. In this case, the correlation of measurement errors of ANL and PLN is theoretically acceptable, since the marketing activities are causally associated and mutually influenced. In addition, many actors in nonprofit organizations can not differentiate between strategic marketing analysis and planning as separate activities, which justifies the correlation of measurement errors e3 and e4 (for manifest variables ANL and PLN), as demonstrated by Figure 3.

A slight change of factor loadings occurs, which confirms that both chisquare test results for the structural model, as well as the goodness-of-fit parameters for the model (see Table 5) are appropriate. The chi-square test results show that 14 parameters were estimated in a model, while the chi- square-value is 7.516, with 7 degrees of freedom. The value of the chi-square/df ratio is 1.074, while the p-value is higher than 5%, which confirms that the model is appropriately specified, i.e. statistically significant.

All other goodness-of-fit indices for the model are appropriate and serve to demonstrate that the model can be accepted. RMSEA (root mean square error of approximation) is lower than 0.05, which requires the check of the p-value for RMSEA (PCLOSE). Its value of 0.525 also falls within the limits for the estimate of a good model suitability.

The indicators of multivariate kurtosis and critical ratios (CR of multivariate kurtosis) are also checked, considering that the univariate distribution normality does not necessarily imply that distribution is characterized by multivariate normality. In this case, the value of multivariate kurtosis CR is far below the lower level of acceptability of 5 (Byrne, 2010). The critical ratio of Mardia's coefficient of multivariate kurtosis is 0.887, which also satisfies the requirement of a value lower than 1.96 (Gao et al., 2008). Estimates of parameters (including non-standardized values, standard errors, critical ratios and p-values, calculated by using the maximum likelihood method), are shown in Table 6.

Within the measurement model MA, multiplicator 1 is allocated to variable IMP2, so that other parameters of the measurement model, i.e. their multiplicators, can be scaled accordingly. In the measurement model FS, multiplicator 1 is allocated to variable FS FO. Upon review of the nonstandardized values of the estimated parameters, it is obvious that they are all significant on an empirical level of significance 0.01.

Critical ratios are estimated on the level of t-, or z-test (higher than 1.96 is significant and suggests a significant contribution to the model), which is the case for this model. Table 7 provides values for total, direct and indirect standardized effects.

It is obvious that marketing activities (MA) have a positive influence on fundraising performance (FS), because the standardized value of the estimated parameter, which shows intensity and direction of variable association, equals 0.674 (if MA increases for one standard deviation, FS will increase for 0.674 standard deviations). If non-standardized indirect effects are reviewed, a positive indirect connection is established: marketing activities influence the financial (0.555), as well as the nonfinancial dimension of the fundraising performance (0.620). The established empirical relationships lead to the conclusion that both H1, as well as both sub-hypotheses (H1.1 and H1.2), should be accepted.

Another SEM model has been created, in order to analyze the relationship between the fundraising performance and the redefinition of marketing activities, involving feedback (FDB) as a mediator variable. This model, along with the factor loadings, is shown in Figure 4.

There are eight manifest variables and three latent variables in the model, with FS being the exogenous latent variable, while FDB and MA are endogenous latent variables. Factor loadings are estimated by the maximum likelihood method and their values can be described as satisfactory. As the model modification, related to inclusion of the correlation between measurement errors e3 and e4, was previously accepted, the same modification will be included in this model, as well. In this way, consistency of the analysis is achieved and both structural models are comparable. Chi-square test and the goodness-of-fit indices (see Table 8) were calculated for this model. Chi-square results show that 20 parameters were estimated in the model, while the chisquare-value is 24.762, with 16 degrees of freedom. The value of the chisquare/df ratio is 1.548, while the p-value is higher than 5%. Therefore, the model is appropriately specified, i.e. statistically significant. All other goodness-of-fit indicators are within acceptable limits and show good suitability.

The multivariate normality can be established, since the value of the critical ratio of multivariate kurtosis is far below the lower level of acceptability of 5 (Byrne, 2010). The value of Mardia's coefficient of multivariate kurtosis is 0.808, which also indicates multivariate normality. Estimates of parameters (including non-standardized values, standard errors, critical ratios and p-values, calculated by using the maximum likelihood method), are shown in Table 9.

Within the measurement model MA, multiplicator 1 is allocated to variable ANL, so that other parameters of measurement model, i.e. their multiplicators, can be scaled accordingly. In the measurement model FS, multiplicator 1 is allocated to variable FS NFO, while, in the measurement model FDB, multiplicator 1 is allocated to variable FDB 1. All estimated parameters are significant, either on the significance level of 0.01, or 0.05. The exception is the relationship of the latent variable FS (fundraising performance/success) with the latent variable MA (marketing activities). Namely, the presumed relationship is not statistically significant and the non-standardized influence of FS to MA is relatively small.

The overview of the structural model shows there is no direct influence of fundraising performance on the (re)definition of the marketing activities, while the indirect influence, via the mediator variable FDB, proves to be statistically significant. Values of total, direct and indirect standardized effects for this structural model are provided in Table 10.

The value of direct influence, i.e. the standardized value of the estimated parameter related to the direct (and statistically insignificant) relationship between fundraising performance (FS) and marketing activities (MA) is 0.240, while the standardized value of the indirect (and statistically significant) effect is 0.439. The standardized value of the total effect equals 0.678. Therefore, the model can be interpreted in terms of FS increase for one standard deviation, leading to an increase of MA for 0.439 standard deviations, via the feedback (FDB) mediator variable. The obtained empirical results show that hypothesis H2 can be accepted as well.

6.CONCLUSION, IMPLICATIONS AND GUIDELINES FOR FUTURE RESEARCH

The analysis of two structural models has demonstrated that both hypotheses are acceptable, i.e. that the entire conceptual model is relevant. This opens up new dimensions for research into nonprofit marketing and strategic management, since this is the first empirical confirmation of the notion that, if it is to reach a high level of performance, nonprofit fundraising should be implemented in the context of the comprehensive nonprofit marketing process. Limited, 'quick-fix' approaches to fundraising, usually arising from the dire need to address the financial crisis in an organization, are, thus, expected to fail, which can also be attributed to the lack of feedback-based marketing improvement. Specifically, this study has empirically demonstrated the relevance of organizational feedback, operationalized by means of the traditional controlling and organizational learning mechanisms.

The creation of new models for measuring organizational performance and new strategies for the entire nonprofit/social sector contributes to its further development, with a special emphasis on fundraising. Fundraising is not only a prerequisite for survival in the crisis-prone nonprofit environment. It has also reached the mature stage, in which it needs to be perceived as an exchange of values. Donors do not just contribute financial means, but satisfy their own needs in the fundraising process, regardless of their nature (Andreasen and Kotler, 2008). A large number of nonprofit organizations do not have a marketing-based approach to fundraising and try to motivate donors to donate in order to satisfy the needs of the organization. This empirical research provides extensive evidence that such an ad-hoc approach does not work. In fact, the exact opposite applies: fundraising specialists need to investigate the needs of target groups of potential donors and propose actions (giving) satisfying the donors' needs (ibid.).

A significant practical implication of this study is related to the questionable viability of small nonprofit organizations, without adequate expertise in nonprofit marketing and fundraising management. Those are likely to be 'pushed' to ad-hoc, unsuccessful fundraising by the lack of financial funds and the decreased giving patterns in the environment, characterized by a high level of uncertainty and a lack of economic growth. Our results imply that a 'vicious circle' might be the resulting outcome for such organizations, additionally fueled by the lack of adequate feedback mechanisms. Thus, additional investments into the human resources and the expertise in nonprofit marketing/management/fundraising skills seem to be the key to future survival, although the costs of training and marketing are often considered to be 'superficial' and easy to eliminate (without visible effects). This study shows that there might be invisible effects, leading to the creation of the 'vicious circle' and the ultimate failure of a large section of the nonprofit/social sector, which still subscribes to the notion of 'amateurism' as the prescribed development path.

In future research, the suitability of proposed structural models for nonprofit organizations from other fields of nonprofit activities (other than humanitarian), as well as from other countries, should be tested. It would also be desirable to examine the influence of individual marketing activities on both dimensions of the fundraising performance. In addition, the suitability of our model(s) should be analyzed for the case of organizations, which acquire the majority of their income through membership fees and social entrepreneurship, since they have a higher degree of resemblance to profit sector organizations. Finally, future research should also take into account the influence of beneficiaries/users - both on the formulation of marketing activities and fundraising performance, as well as on the established patterns of the marketing - fundraising variables in nonprofit marketing.

Sidebar

NEPROFITNI MARKETINSKI PROCES I UCINAK PRIKUPLJANJA SREDSTAVA HUMANITARNIH ORGANIZACIJA: EMPIRIJSKA ANALIZA

Sazetak

Cilj ovog rada je povezati uspjesnost fundraisinga sa marketinskim aktivnostima neprofitnih organizacija. Istrazivanje obuhvaca financijsku i nefinancijsku dimenziju performansi uspjesnosti fundraisinga kako bi se prikazala oba aspekta ishoda procesa fundraisinga. Empirijski dio rada proveden je na uzorku neprofitnih organizacija u Hrvatskoj. U svrhu procjene hipoteza o pozitivnom utjecaju marketinskih aktivnosti na uspjesnost financijskih i nefinancijskih performansi fundraisinga u istrazivanju je koristena metoda modeliranja strukturnih jednadzbi (SEM). U radu se kriticki analizira i empirijski potvrduje povratni utjecaj uspjesnosti fundraisinga na (re)definiranje marketinskih aktivnosti. Dodatno su, na temelju rezultata istrazivanja, prikazane implikacije za marketinske i menadzerske prakse neprofitnih organizacija kao i preporuke za buduca istrazivanja.

Footnote

* A previous version of this paper has been presented and discussed at the 11th International Conference "Challenges of Europe: Growth, Competitiveness and Inequality", organized by the Faculty of Economics Split, in May 2015. The author wishes to express the gratitude for discussions, comments and all suggestions to two anonymous reviewers and participants of the conference "Challenges of Europe: Growth, Competitiveness and Inequality", who contributed to the final version of this paper.

1 https://registri.uprava.hr/#!udruge; https://banovac.mfin.hr/rnoprt/ (Accessed on 14. February 2015).

References

REFERENCES

1. Abdulai Mahmoud, M. and Yusif, B. (2012): Market orientation, learning orientation, and the performance of nonprofit organisations (NPOs), International Journal of Productivity and Performance Management, 61 (6), 624-652.

2. Alfirevic, N., Pavicic, J., and Najev-Cacija, Lj. (2014): Performance of non-profit organizations: Empirical contrasts between privately and publicly funded Croatian humanitarian organizations. Economic annals, 59 (200), 115-129.

3. Andreasen, A. R., and Kotler, P. (2008): Strategic marketing for nonprofit organizations, Prentice Hall, Upper Saddle River.

4. Andreoni, J. (2006): Leadership giving in charitable fundraising. Journal of Public Economic Theory, 8 (1), 1-22.

5. Balabanis, G., Stables, R. E., and Phillips, H. C. (1997): Market orientation in the top 200 British charity organizations and its impact on their performance. European Journal of Marketing, 31 (8), 583-603.

6. Barrett, W.P. (2005) The 200 largest U.S: Charities, Forbes, http://www.forbes.com/2005/11/18/largest-charities-ratings_05charities_land.html [Accessed 11.4.2013]

7. Bennett, R. (2003): Factors underlying the inclination to donate to particular types of charity, International Journal of Nonprofit and Voluntary Sector Marketing. 8 (1), 12-29.

8. Bennett, R. (2007): The use of marketing metrics by British fundraising charities: a survey of current practice, Journal of Marketing Management. 23 (9-10), 959-989.

9. Bennett, R., and Savani, S. (2011): Sources of new ideas for charity fundraising: An empirical study, Creativity and Innovation Management. 20 (2), 121-138.

10. Bentler, P. M., and Chou, C. P. (1987): Practical issues in structural modeling. Sociological Methods and Research. 16 (1), 78-117.

11. Browne, M. W., MacCallum, R. C., Kim, C. T., Andersen, B. L., and Glaser, R. (2002): When fit indices and residuals are incompatible. Psychological Methods, 7 (4), 403-421.

12. Bryson, J. M. (2004): Strategic planning for public and nonprofit organizations: a guide to strengthening and sustaining organizational achievement, Jossey Bass, San Francisco.

13. Bulla, M., and Starr-Glass, D. (2006): Marketing and non-profit organizations in the Czech Republic. European Journal of Marketing, 40 (1-2), 130-144.

14. Byrne, B. M. (2010): Structural equation modeling with AMOS: basic concepts, applications, and programming, Routledge, New York.

15. Coffman, D. L., and MacCallum R. C. (2005): Using parcels to convert path analysis models into latent variable models. Multivariate behavioral research, 40 (2), 235-259.

16. Dimovski, V., Skerlavaj, M., Kimman, M., and Hernaus, T. (2008): Comparative analysis of the organisational learning process in Slovenia, Croatia and Malaysia. Experts Systems with Applications, 34, 3063-3070.

17. Dolnicar, S., and Lazarevski, K. (2009): Marketing in non-profit organizations: an international perspective. International Marketing Review, 26 (3), 275-291.

18. Field, A. (2009): Discovering statistics using SPSS (and sex and drugs and rock 'n ' roll), Sage publications, London.

19. Flores, L. G., Zheng, W., Rau, D., and Thomas, C. H. (2010): Organizational Learning: Subprocess Identification, Construct Validation, and an Empirical Test of Cultural Antecedents. Journal of Management, 38 (2), 640-667.

20. Gao, S., Mokhtarian, P. L., and Johnston, R. A. (2008): Nonnormality of data in structural equation models. Transportation Research Record: Journal of the Transportation Research Board, 2082 (1), 116-124.

21. Gignac, G. E. (2006): Self-reported emotional intelligence and life satisfaction: testing incremental predictive validity hypotheses via structural equation modeling (SEM) in a small sample. Personality and Individual Differences, 40, 1569-1577.

22. Gillian, D. A., and Voss, K. (2013): A proposed procedure for construct definition in marketing. European Journal of Marketing, 47 (1/2), 5-26.

23. Haig, B. D. (2005): An abductive theory of scientific method. Psychological Methods, 10 (4), 371-388.

24. Hall, R. J., Snell, A.F. and Foust, M. S. (1999): Item Parceling Strategies in SEM: Investigating the Subtle Effects of Unmodeled Secondary Constructs, Organizational Research Method, 2, 233-258.

25. Hart, T (2008): e-Philanthropy: Leveraging technology to benefit charities and donors. In Sargeant, A., and Wymer, W. (eds.), The Routledge Companion to Nonprofit Marketing, Routledge, New York.

26. Hayduk, L., Cummings, G., Boadu, K., Pazderka-Robinson, H., and Boulianne, S. (2007): Testing! testing! one, two, three - Testing the theory in structural equation models! Personality and Individual Differences, 42, 841-850.

27. Heinzel, H. (2004): Philanthropy and fundraising in Western Europe within a framework of change, New directions for philanthropic fundraising, 46, 101-120.

28. Herman, R. D., and Renz, D. O. (2004): More theses on nonprofit organizational effectiveness: results and conclusions from a panel study, in: Sixth international conference on' International society for third sector research', Proceedings, Toronto, July 11-14, 2004, pp. 11-14.

29. Hsieh, J. (2010): Strategic stakeholder orientation and performance consequences - a case of private nonprofit performing arts in the US. International Journal of Nonprofit and Voluntary Sector Marketing, 15, 13-27.

30. Keating, E. K., and Frumkin, P. (2003): Reengineering Nonprofit Financial Accountability: Toward a More Reliable Foundation for Regulation. Public Administration Review, 63 (1), 3-15.

31. Kline, R. B. (2011): Principles and practice of structural equation modeling, The Guilford Press, New York.

32. Knowles, P., and Gomes, R. (2009): Building Relationships with Major-Gift Donors: A Major-Gift Decision-Making, Relationship Building Model. Journal of Nonprofit and Public Sector Marketing, 21, 384-406.

33. Kohli, A. K., Jaworski, B. J., and Kumar, A. (1993): MARKOR: A measure of market orientation. Journal of Marketing Research, 30 (4), 467-477.

34. Kotier, P. (1999): Upravljanje marketingom: analiza, planiranje, primjena, kontrola, Informator, Zagreb.

35. Lai, C. A., and Poon, J. P. H. (2009): Location, Marketing, and the Financiai Fiexibiiity of Nonprofit Performing Arts Organizations in Second Tier Cities. Journal of Nonprofit and Public Sector Marketing, 21 (2), 160-183.

36. Little, T. D., Cunningham, W. A., and Shahar, G. (2002): To parcel or not to parcel: Exploring the question, weighing the merits. Structural Equation Modeling, 9 (2), 151-173.

37. Lopez, S. P., Montes Peon, J. M., and Vasquez Ordas, C. M. (2005): Organizational learning as a determining factor in business performance. The Learning Organization, 12 (3), 227-245.

38. Marcoulides, G. A., and Saunders, C. (2006): PLS: A silver bullet? Management Information Systems Quarterly, 30(2), iii-ix.

39. McGee, S., and Donoghue, F. (2009): The Conundrum of Fundraising in Nonprofit Organizations: A Story from Ireland. Journal of Nonprofit and Public Sector Marketing, 21, 367-383.

40. McLeish, B. J. (2011): Successful marketing strategies for nonprofit organizations, John Wiley and Sons, New Jersey.

41. McLoughlin, D., and Aaker, D. A. (2010): Strategic market management: global perspectives, John Wiley and Sons, New Jersey.

42. Modi, P. (2012): Market orientation in nonprofit organizations: innovativeness, resource scarcity, and performance, Journal of Strategic Marketing, 20 (1), 55-67.

43. Mottner, S., and Ford, J. B. (2005): Measuring nonprofit marketing strategy performance: the case of museum stores. Journal of Business Research, 58, 829-840.

44. Narver, J. C., and Slater, S. F. (1990): The effect of a market orientation on business profitability. Journal of Marketing, 54 (4), 20-34.

45. Novatorov, E. V. (2010): A critical appraisal of the concept of nonprofit services marketing. Service Science, 2 (3), 136-143.

46. Pavicic, J., Alfirevic, N., and Ivelja, N. (2006): Oblici i mehanizmi potpore i suradnje s neprofitnim organizacijama. Zbornik Ekonomskog fakulteta u Zagrebu, 4, 217-238.

47. Poister, T. H. (2003): Measuring performance in public and nonprofit organizations, Jossey Bass, San Francisco.

48. Pope, J. A., Isely, E. S., and Asamoa Tutu, F (2009): Developing a marketing strategy for nonprofit organizations: An exploratory study. Journal of Nonprofit and Public Sector Marketing, 21 (2), 184-201.

49. Rocha, C. M., and Chelladurai, P. (2012): Item parcels in structural equation modeling: an applied study in sport management. International Journal of Psychology and Behavioral Sciences, 2 (1), 46-53.

50. Sargeant, A., and Jay, E. (2004): Building donor loyalty: The fundraiser's guide to increasing lifetime value, Jossey Bass, San Francisco.

51. Sargeant, A., and Shang, J. (2010): Fundraising principles and practice, John Wiley and Sons, New Jersey.

52. Sargeant, A., and Woodliff, L. (2007): Building Donor Loyalty: The Antecedents and Role of Commitment in the Context of Charity Giving. Journal of Nonprofit and Public Sector Marketing, 18 (2), 47-68.

53. Sargeant, A., and Woodliff, L. (2008): Individual giving behaviour: A multidisciplinary review, in Sargeant, A., and Wymer, W. (eds.), The Routledge Companion to Nonprofit Marketing, Routledge, New York.

54. Sargeant, A., and Wymer, W. (2008): Introduction: The growth of a discipline, in Sargeant, A., and Wymer, W., (eds.), The Routledge Companion to Nonprofit Marketing, Routledge, New York.

55. Sawhill, J. C., and Williamson, D. (2001): Mission impossible? Measuring success in nonprofit organization. Nonprofit Management and Leadership, 11 (3), 371-386.

56. Schermelleh-Engel, K., Moosbrugger H., and Hans Müller, H. (2003): Evaluating the Fit of Structural Equation Models: Tests of Significance and Descriptive Goodness-of-Fit Measure. Methods of Psychological Research Online, 8 (2), 23-74.

57. Sheth, J. N., and Sisodia, R. S. (2005): A dangerous divergence: marketing and society. Journal of Public Policy and Marketing, 24 (1), 160-162.

58. Srnka, K. J., Grohs, R., and Eckler, I. (2003): Increasing fundraising efficiency by segmenting donors, Australasian Marketing Journal, 11 (1), 70-86.

59. Stater, K. J. (2009): The impact of revenue sources on marketing behavior: Examining Web-promotion and place-marketing in nonprofit organizations. Journal of Nonprofit and Public Sector Marketing, 21, 202-224.

60. Sterba, S. K. (2011): Implications of parcel-allocation variability for comparing fit of item solutions and parcel solutions. Structural Equation Modeling, 18, 554-577.