---

This paper contributes to the literature on Open Source (OS) software by providing empirical evidence on the incentives of firms that engage in OS activities. Data collected by a survey conducted on 146 Italian companies supplying OS solutions (Open Source firms) show that (surprisingly) intrinsic, community–based incentives do play a role but are not, in general, put into practise. We investigate this discrepancy between attitudes and behaviours and single out groups of firms adopting more consistent behaviours. Our results are in line with the literature on business models of the firms that enter the Open Source field.

 

Contents

Introduction
Intrinsic and extrinsic motivations in the Open Source movement
Discrepancy between attitudes and behaviours in Open Source firms
Conclusions

 

---

 

Introduction

Since the rise of interest of scholars in the Open Source phenomenon, one of the most intriguing questions has dealt with developers’ incentives. A growing body of literature has addressed the issue and many studies have collected empirical data on the motivations of individuals that actively participate in OS projects (Table 1). Psychological theory distinguishes between intrinsic and extrinsic incentives. A motivation is extrinsic if needs are satisfied indirectly, especially through monetary compensation while intrinsic incentives steam from the very pleasure of carrying on an activity. Empirical analyses have highlighted that OS developers show both intrinsic and extrinsic motivations.

 

Table 1: Motivations of Open Source developers.

	Motivations	Sources
Extrinsic	Low opportunity costs	Bonaccorsi and Rossi (2003)
	Monetary rewards	Feller and Fitzgerald (2002); Hertel, et al. (2003); Lerner and Tirole (2002a)
	Reputation among peers	Dalle and David (2004); Lerner and Tirole (2001)
	Future career benefits	Lerner and Tirole (2004)
	Learning	von Krogh, et al. (2003)
	Contributions from the community	Raymond (2001)
	Technological concerns	Weber (2004)
	Filling an unfilled market	Feller and Fitzgerald (2002)
Intrinsic	Creative pleasure (Fun to program)	Torvalds and Diamond (2002); Lakhani, et al. (2004)
	Altruism	Bitzer, et al., (2004); Zeitlyn (2003)
	Sense of belonging to the community	Crowston and Howison (2004)
	Fight against proprietary software	Stallman (1984)

 

Extrinsic motivations deal in particular with being paid for Open Source development [ 1]; gaining reputation as talented programmers (thus obtaining future career benefits); learning by studying the code written by others; and, benefiting from contributions from the community. However, many developers number fun to program, altruism, sense of belonging to the Open Source community and willingness to take part in the fight for software freedom among the most important reasons to carry on Open Source activities [2].

Feller and Fitzgerald (2002) have added to this literature, analysing the incentives of software companies that engage in OS activities. The aim at profiting shapes the decision to adopt Open Source–based business models (Table 2) while it has been claimed that contingent agreement with the non–written norms of the OS community simply serves the purpose to keep active the cooperative link with developers (Osterloh, et al., 2001). Behaviours violating community norms [ 3] may bring down cooperation and reduce the flow of contributions. This paper contributes to the literature by providing empirical evidence on the firms that supply OS–based products and services (Open Source firms). To the best of our knowledge, we are not aware of works that gather data on these firms.

Three main research questions are addressed:

(i) Which motivations lay at the basis of the entrepreneurial decision to set up an Open Source–based business model?

 

Table 2: Motivations of Open Source firms.

	Motivations	Sources
Extrinsic	Independence from price and licence policies of large software companies	Lerner and Tirole (2002b)
	Supplying software–related services	Feller and Fitzgerald (2002); Wichmann (2002a)
	Selling related products	Lerner and Tirole (2002a); Wichmann (2002a)
	Exploitating R&D activity of other developers and OS firms	Hawkins (2004); Lakhani, et al. (2003); Dahlander (2004)
	Software testing by users’ community	von Hippel (2002); Fink (2003)
	Hiring good Open Source technicians	Fink (2003); Wichmann (2002b)
	Lowering hardware costs	Feller and Fitzgerald (2002)
	Security concerns	Fink (2003)
Intrinsic	Conforming to the values of the OS community (not betraying the trust of other developers)	Kuster, et al. (2003); Lerner and Tirole (2002b)
	Code sharing with the community (reciprocating to sustain cooperation)	Kuster, et al., (2003)
	Fight for software freedom (reducing the market shares of large software companies)	Feller and Fitzgerald (2002)

 

Even if individuals employed by OS firms may be intrinsic motivated in their jobs, their behaviour ultimately depends on the profit orientation of the organisation they are part of. Hence, firms’ incentives cannot been regarded simply as the sum of individuals’ incentives. Why do firms allow their employees to allocate part of their working hours to the production of a collective good?

(ii) In the event that profit–oriented firms declare to attach importance to intrinsic, community–based motivations, do these attitudes generate appropriate behaviours? Do firms that declare to agree with the values of the OS community contribute to projects, maintain social links with developers and devote resources to promote the diffusion of the open standards? Or rather is there a discrepancy between attitudes and behaviours?

(iii) If attitudes do not generate consistent behaviours, is there a recognizable pattern in discrepancy? Are there respondents that adopt consistent behaviours? And if yes, what distinguishes these latter from the others? Discrepancy may impact on the long term sustainability of the Open Source as an original industrial model. Indeed, the larger the discrepancy, the higher the risk that the companies are estranged from the OS community. The ensuing reduction of the contributions from developers might threaten their survival.

 

Intrinsic and extrinsic motivations in the Open Source movement

Data have been collected by submitting a structured questionnaire to the partners or the system administrators of more than 250 Italian Open Source companies [ 4] obtaining 146 valid answers (see Bonaccorsi, et al., 2004 for the sample selection details). Data show that the offering of OS solutions in Italy is managed mainly by small companies that are born after 1998 (51 percent of the sample) and have adopted the new paradigm only recently (Table 3).

Table 3: Descriptive statistics of firms’ characteristics. Note: firms born after 2001 were asked for the change in OS turnover in the last year.

Incentives are measured through a 5–point Likert scale. Firms had to give a mark from 1 (I totally disagree) to 5 (I totally agree) to 11 items selected according to the literature. Basing on the work of Lakhani and Wolf (2003) on individual developers’ incentives, we distinguish between intrinsic (IM) and extrinsic (EM) motivations.

Our findings corroborate theoretical hypotheses. As expected, extrinsic motivations do play a leading role. When items are ranked by the mean of the scores incentives that fit well the decision processes of profit–oriented firms rank first (Table 4).

Findings on intrinsic motivations are twofold. Firms declare to agree with the values of the community (IM1) but the item dealing with the fight for software freedom (IM3) ranks at the bottom of the list while the one concerning code gifting is below the average (IM2). This gets into line with the literature regarding firms’ social motivations as simply serving the purpose of sustaining cooperation with developers. To provide further evidence, we compare data on incentives with data on firms’ involvement in community activities and check for discrepancies between attitudes (a high level of accordance with community–oriented motivations) and behaviours (the actual participation in Open Source).

Table 4: Firms’ motivations: Descriptive statistics and score distributions.

 

Discrepancy between attitudes and behaviours in Open Source firms

Discrepancy between firms’ attitudes and behaviours is tested referring to metrics of involvement in Open Source activities dealing with social links with the community, involvement in OS advertising activities, participation in OS projects (Table 5). Data show that firms in the sample carry on community–oriented activities only to a limited extent, especially as far as projects’ participation is concerned. It seem that companies basically adapt OS programs to meet customers’ requirements (firms as code takers) while little importance is attached to circulate these solutions back to the community (firms as code givers).

These findings are at odds with the general agreement with the values of OS community declared in the question on incentives: discrepancy between attitudes and behaviours is clear–cut. Firms that have chosen high (4 or 5), medium (3) and low (1 or 2) scores for proposition IM1 are compared. All the behavioural variables but the time devoted to Open Source advertising activities show no significant differences in the mean values in the three groups (Table 6).

Table 5: Firms’ involvement in Open Source activities.

Are these discrepancies a generalized pattern? Are there respondents acting consistently with their attitudes? If yes, do these firms share peculiar characteristics?

Table 6: Discrepancy between attitudes and behaviours. Note.*: p value < 0.10; ***: p value < 0.01.

Synthetic measures for attitudes and behaviours have been calculated by running principal component analyses (PCA). Three components are extracted for the behavioural dimensions but only the first one shows positive correlations with all the behaviour variables. Hence this component (B) turns out to be a valid metric for firms’ community oriented behaviours. A companion metric for firms’ attitudes is obtained by running PCA on the variables dealing with intrinsic motivations (IM1, IM2, and IM3). Only one component is extracted from the data (IM) which is high correlated with all the intrinsic incentives. The Pearson Correlation Index between IM and B is not significantly different from zero (0.096, p value = 0.249) corroborating that, in general, the agreement with the OS values does not give rise to consistent behaviours

Four groups of firms come to evidence:

1. Non Community Oriented Firms (50, 34.2 percent): negative values for both the dimensions. They have low intrinsic motivations and act in a consistent manner
2. Incognito Community Oriented Firms (13, 8.9 percent): positive values for B but negative values for IM. They behave inconsistently with their attitudes but in an unexpected way
3. Community Oriented Firms (27, 18.5 percent): positive values for both the dimensions. They declare strong intrinsic motivations and act in a consistent manner
4. Opportunistic Firms (45, 30.8 percent) positive values for IM and negative values for B. Theses firms do not practise what they preach.

The first group poses no problem. As profit–oriented organisations, these firms entered the OS field for gaining competitive advantages. They exploit collective developed software but do not take part actively in its production (firms as code takers). The percentage of the non–users of the GLP license is significantly higher in this group than in the rest of the sample (27.1 percent vs. 8.8 percent, Chi Square Test, p value=0.006). Notwithstanding that most respondents (66.0 percent) declare to attach high strategic importance to Open Source, 57.1 percent offer indifferently open and proprietary solutions [ 5] (vs. 33.3 percent of the other firms, Chi Square Test, p value = 0.010). Few firms are in the second group, so the empirical evidence is poorly informative. It is worth noticing that 10 out 13 (76.9 percent) assign a high score to EM2 proposition (contributions and feedback from the Open Source community are very useful to fix bugs and improve the software). This may indicate that, as in Osterloh’s hypothesis, their community–oriented behaviours are aimed at keeping active the links with individual programmers. The group is consistent with its profit–oriented nature and assigns low scores to intrinsic incentives.

The most intriguing groups are the third and the fourth. The former shows strong community–oriented attitudes [ 6] and behave consistently with them. In general, these firms have adopted the new paradigm from the very start (early adopters: 88.9 percent vs. 59.3 percent, Chi Square Test, p value=0.004). They all have joined at least one OS project and have social contacts with individual developers. The large majority have carried on coordination tasks and perform Open Source advertising in activities for which there is an entry budget in almost half on the cases. Surprising no significant difference emerges as far as the use of the GPL is concerned but all the respondents state to attach high strategic importance to the Open Source which, in 92.6 percent of the cases, is the most important offering.

Findings on Opportunistic Firms corroborate the hypothesis that firms’ community–based motivations are extrinsic in their nature. These firms declare to agree with the values of the Open Source community (average score of IM1: 4.5 vs. 3.4 of the other firms, p value=0.000) but their contributions in projects are scanty (Table 7). No difference emerges as far as the exclusive use of the GPL is concerned (but only only of the respondents in this group do not use this license) and in half of the cases proprietary and OS solutions are indifferently provided to the customers. The agreement with the OS values seems to be only nominal without being put into practise.

Table 7: Opportunistic firms behaviours, Chi Square Tests. Note. * p value < 0.1; **: p value < 0.5; ***: p value < 0.01.

About the characteristics of the firms in the four groups, we discuss the following hypothesis.

Hypothesis I: Firms’ attitudes and behaviours depend on the strength of the commitment in the new paradigm. Companies with a strong commitment are likely to show positive attitudes towards the community and to behave consistently (Community Oriented Firms). On the contrary firms with a weak commitment may show both negative attitudes/consistent behaviour (Non–Community Oriented Firms) or positive attitudes/discrepant behaviours (Opportunistic Firms)

Firms with a strong commitment are likely to couple positive attitudes towards the community with consisting behaviours. These companies base their being on the market on the exploitation of the code produced by the Open Source community. As a consequence the link with individual developers is so important that they really need to take part actively in community activities. Moreover, these firms are likely to be the results of the decision to enter the market by a group of individual developers who chose to exploit commercially the Open Source skills they have gained by programming within projects. So it is possible they bequeath the values and the behavioural schemes of the hacker culture to their companies.

In turn, firms with a weak commitment are likely not adopt community–oriented behaviours as their involvement in the field is scanty and based mainly on the adaptation of OS programs to meet customers’ requirements. Even these firms may want to keep active the link with the community as it might turn out to be useful independently on the focus on OS of the business model. First, independently on their size, firms’ Open Source activities might depend strongly on contributions and feedback from the community. Companies mainly focused on proprietary solutions are likely to suffer from shortage of Open Source skills and need the help of individual developers. Second, it worth noticing that customers that choose Open Source software have often strong intrinsic motivations thinking of open standards as a way for contrasting the monopoly power of large software companies. As a consequence, proprietary software is condemned not only by the OS community (West and O’Mahony, 2005) but also by individual users (Rogers, 2000).

In a companion paper (Bonaccorsi, et al., 2004) we have proposed a classification of the respondents in the sample on the basis of their commitment in the Open Source paradigm (More Open Source Oriented, MOSS, vs. Less Open Source Oriented, LOSS). Seventy–four firms have been labelled as MOSS, 64 as LOSS while eight firms, having a business model based on the provision of open solutions, have been considered Pure Open Source Firms (POSF).

Findings in Table 8 corroborates hypothesis II. The contingency table shows that in the Community Oriented group only one firm out of 27 is LOSS, 20 are MOSS while six are Pure Open Source firms. A similar distribution hold for Incognito Community Oriented Firms while more than half of the Non Community Oriented and of the Opportunistic Firms are LOSS. These latter account for the 92.5 percent of the Less Open Source Oriented firms.

Table 8: MOSS and LOSS firms distribution.

It is now of interest to study whether the 23 LOSS firms in the Opportunistic group suffer from a shortage in Open Source skills or serve customers with strong preferences for Open Source software. With respect to the other LOSS firms, the 23 Opportunistic respondents assign higher scores to proposition EM5 dealing with the availability of good IT specialist in the Open Source field (average score: 3.7 vs. 3.0, Mann–Withney Test, p value=0.035) and proposition EM6 dealing with the learning opportunities given by the new paradigm (average score: 3.6 vs. 2.8, Mann–Withney Test, p value=0.044). No significant difference emerge as far as consumers’ attitudes are concerned. No comprehensive conclusion can be made. Setting aside the small size of the considered sub–samples, further investigations are needed also to find better proxies for firms’ shortage in Open Source skills and customers’ attitudes towards Open Source software.

 

Conclusions

This paper adds to the literature by providing empirical evidence on the motivations of the software companies that enter the Open Source field aiming at profiting from the new paradigm. Using data collected by a large–scale survey on 146 Italian Open Source firms, we find out that, surprisingly, extrinsic, profit–oriented incentives couple with intrinsic, community–based motivations. In most cases, these positive attitudes are not put into practise and the very participation to the Open Source community is scanty. This discrepancy between attitudes and behaviours corroborates the hypothesis of Osterloh, et al. (2001) on the extrinsic nature of firms’ intrinsic motivations. The authors claim that, in case of firms, declaring community–based incentives simply serves the purpose of winning the trust of individual developers for receiving contributions and support from them and gaining competitive and cost advantages.

After grouping the respondents of the basis of the dimension of the discrepancy, we show that the relation between attitudes and behaviour depends on the strength of firms’ commitment on Open Source software. Firms with a strong commitment show also a strong link with the community that is expressed not only be the agreement with its values but also by the participation to community–oriented activities. Nevertheless, even companies with a business model mainly based on proprietary solutions may find it convenient to express community–oriented attitudes in order to exploit the feedback and contributions from individual developers and give a positive corporate image to customers preferring Open Source solutions.

 

About the authors

Cristina Rossi is a post–doc researcher at the Sant’Anna School of Advanced Studies, Pisa, Italy and at the Department of Electrical Systems and Automation, University of Pisa. Her research interests examine the economics of Open Source software, intellectual property rights and the diffusion of innovations under network externality effects.
E–mail: cristina [dot] rossi [at] sssup [dot] it

Andrea Bonaccorsi is Full Professor of Economics and Management at the School of Engineering, University of Pisa. His research interests treat the economics of science and technical changes and the theory of diffusion of new technologies.

 

Acknowledgements

The competent assistance of Alessandro Scateni in database construction is acknowledged.

 

Notes

1. Contrary to what most people think, it is not unusual that programmers receive monetary compensation for their OS activities (Hertel, et al., 2003).

2. Lakhani and Wolf (2003) distinguish between enjoyment–based intrinsic motivations namely how creative a person feels when working on OS projects and community–based intrinsic motivations dealing with acting consistently with the norms of a group.

3. As including pieces of open code into proprietary programs or keeping closed parts of software released to the community.

4. We asked for the filling by one opf the partners. When this was not possible system administrators were targeted.

5. Closed response question, options: exclusively OS solution; mainly OS solutions; indifferently Proprietary and OS solutions.

6. No firm assigns low scores to IM1 and IM2 while than 20 percent assign low scores IM3. High scores: 96.3 percent, 88.9 percent and 55.5 percent, respectively.

 

References

J. Bitzer, W. Schrettl, and P. Schröder, 2004. "Intrinsic motivation in open source software development," Diskussionsbeiträge des Fachbereichs Wirtschaftswissenschaft der Freien Universität Berlin, number 2004/19, at http://opensource.mit.edu/papers/bitzerschrettlschroder.pdf, accessed 1 May 2005.

A. Bonaccorsi, S. Giannangeli, and C. Rossi, 2004. "Entry strategies under dominant standards — Hybrid business models in the Open Source software industry," Social Science Research Network (SSRN) Working Paper, at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=519842, accessed 1 May 2005.

A. Bonaccorsi and C. Rossi, 2003. "Why open source can succeed," Research Policy, volume 32, number 7, pp. 1243–1258.

L. Dahlander, 2004. "Appropriating returns from open innovation processes: A multiple case study of small firms in Open Source software," at http://opensource.mit.edu/papers/dahlander.pdf, accessed 1 May 2005.

J.–M. Dalle and P. David, 2003. "The allocation of software development resources in ‘open source’ production mode," at http://ideas.repec.org/p/wpa/wuwpio/0502011.html, accessed 1 May 2005.

J. Feller and B. Fitzgerald, 2002. Understanding open source software development. Boston: Addison–Wesley.

M. Fink, 2003. The business and economics of Linux and open source. Upper Saddle River, N.J.: Prentice Hall PTR.

R. Hawkins 2004. "The economics of the open source software for a competitive firm," Netnomics, volume 6, number 2, pp. 103–117.

G. Hertel, S. Niedner, and S. Hermann, 2003. "Motivation of software developers in the Open Source projects: An Internet–based survey of contributors to the Linux kernel," Research Policy, volume 32, number 7, pp. 1159–1177.

B. Kuster, M. Osterloh, and S. Rota, 2003. "Trust and commerce in open source — a contradiction? " In: O. Petrovic, M. Ksela, M. Fallenböck, and C. Kittl (editors). Trust in the network economy. Berlin; Springer.

K. Lakhani and B. Wolf, 2003. "Why hackers do what they do: Understanding motivation and effort in free/open source software projects," MIT Working Paper, at http://freesoftware.mit.edu/papers/lakhaniwolf.pdf, accessed 1 May 2005.

K. Lakhani, B. Wolf, J. Bates, and C. DiBona, 2002. "The Boston Consulting Group hacker survey," http://downloads.planetmirror.com/pub/lca/2003/proceedings/papers/Hemos/Hemos.pdf, accessed 1 May 2005.

J. Lerner and J. Tirole, 2004. "The economics of technology sharing: Open source and Beyond," Harvard Business School, Negotiation, Organizations and Markets Unit, Research Paper Series, number 04–35, at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=620904, accessed 1 May 2005.

J. Lerner and J. Tirole, 2002a. "Some simple economics of open source," Journal of Industrial Economics, volume 50, number 2, pp. 197–234.

J. Lerner and J. Tirole, 2002b. "The scope of open source licensing," at http://opensource.mit.edu/papers/lernertirole2.pdf, accessed 1 May 2005.

J. Lerner and J. Tirole, 2001. "The open source movement: Key research questions," European Economic Review, volume 45, issues 4–6, pp. 819–826.

M. Osterloh, S. Rota, and M. von Wartburg, 2001. "Open source — new rules in software development," Working Paper, Institute for Research in Business Administration, University of Zurich.

E. Raymond, 2001. The cathedral & the bazaar: Musings on Linux and open source by an accidental revolutionary. Sebastopol, Calif.: O’Reilly.

P.L. Rogers, 2000. "Linux advocacy mini–HOWTO," http://www.tldp.org/HOWTO/Advocacy.html, accessed 1 May 2005.

R. Stallman, 1984. "The GNU manifesto," http://www.gnu.org/gnu/manifesto.html, accessed 1 May 2005.

L. Torvalds and D. Diamond, 2001. Just for fun: The story of an accidental revolutionary. New York: HarperBusiness.

E. von Hippel, 2002. "Open source projects as horizontal innovation networks — by and for users," Social Science Research Network (SSRN) Working Paper, at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=328900, accessed 1 May 2005.

G. von Krogh, K. Lakhani, and S. Spaeth, 2003. "Community, joining, and specialization in open source software innovation: A case study," Research Policy, volume 32, number 7, pp. 1217–1241.

S. Weber, 2004. The success of open source. Cambridge, Mass.: Harvard University Press.

J. West and S. O’Mahony, 2005. "Contrasting community building in sponsored and community founded Open Source projects," Proceedings of the 38th Annual Hawaii International Conference on System Sciences, Waikoloa, Hawaii (3–6 January), at http://csdl.computer.org/comp/proceedings/ hicss/2005/2268/07/22680195.pdf, accessed 1 May 2005.

T. Wichmann, 2002a. Use of open source software in firms and public institutions. Evidence from Germany, Sweden and UK. Free/libre and open source software: Survey and study, FLOSS final report — Part 1, at http://www.berlecon.de/studien/downloads/200207FLOSS_Use.pdf, accessed 1 May 2005.

T. Wichmann, 2002b. Firms’ open source activities: Motivations and policy implications. Free/libre and open source software: Survey and study, FLOSS final report — Part 2, at http://www.open-source.gr/content/modules/ downloads/Motivations_Policy_Implications_Part_II.pdf, accessed 1 May 2005.

D. Zeitlyn, 2003. "Gift economies in the development of open source software: Anthropological reflections," Research Policy, volume 32, number 7, pp. 1287–1291.

---

Editorial history

Paper received 22 February 2005; accepted 1 April 2005.
HTML markup: Susan Bochenski and Edward J. Valauskas; Editor: Edward J. Valauskas.

---

Copyright ©2005, Cristina Rossi and Andrea Bonaccorsi.

Intrinsic vs. extrinsic incentives in profit–oriented firms supplying Open Source products and services by Cristina Rossi and Andrea Bonaccorsi
First Monday, volume 10, number 5 (May 2005),
URL: http://firstmonday.org/issues/issue10_5/rossi/index.html