Toward Overcoming Suboptimal Provision of Free and Open-Source Software: The Case of CiviCRM

[Abstract] This text examines the suboptimal provision of free and open-source software (FOSS) using the example of CiviCRM, a software for constituent relationship management that has been developed worldwide for and by non-governmental organizations (NGOs) for 20 years, that holds significant market potential within the EU, and that is already used by thousands of NGOs across the EU. The suboptimal provision of CiviCRM is understood in this text neither as a temporary deficit nor as the result of individual failure, but rather as a stable development path emerging from rational individual decisions made by the actors involved. Analytically, a distinction is made between the provision of software functionality and the capacity to appropriate the software. The text is structured in four steps: it first describes an observable suboptimality in usage and diffusion, then explains this as an emergent result of rational decisions in the sense of the collective action problem as described by Mancur Olson, and subsequently discusses institutional governance approaches, drawing on Elinor Ostrom, for overcoming this suboptimality. The aim of this analysis is to make visible the welfare potential of the software that has not yet been realized and to identify points of leverage through which this potential can be unlocked under realistic institutional conditions. In an additional note, the existing governance structure of CiviCRM is reconstructed and assessed in terms of its scope, representativeness, and capacity. In addition, it includes a risk analysis with regard to the geopolitical and legal embedding of central elements of CiviCRM’s governance in the United States, as well as an exploratory attempt to approximate the scale of suboptimally bound resources and welfare potentials. Finally, the existing situation is interpreted not only as a structural problem but also as a strategic opportunity: particularly for implementation agencies, which can contribute to the transformation of the existing development path through standardized services, scaling, and constructive integration into governance structures.

The comparatively extensive length of this text is deliberate. It serves the reconstruction of the argument and the systematic linkage of empirical observations with the analysis of institutional arrangements, as well as their respective theoretical classification. The text is conceived as an analytical reference framework; individual sections can be read independently of one another.

Table of Contents

1. Levels of suboptimal provision of free and open-source software

In an increasing number of societal domains, digital infrastructure is built on free and open-source software (FOSS). It is developed collectively as a collective good and promises its users long-term sovereignty—unlike software systems that are protected as exclusive goods through property and licensing regimes and that invariably create economic dependencies. Particularly within civil society organizations, FOSS often enables digital participation and collaboration in the first place, especially because the comparatively higher costs and poorer conditions of use of proprietary systems would not allow for this.

However, the everyday use of FOSS is accompanied by tensions, which manifest themselves partly in the performance of the software, but even more so in its concrete capacity for appropriation, maintainability, and further development. These tensions are examined in the following using the example of CiviCRM.

CiviCRM is an open-source software for managing contacts and activities, developed for and by non-profit organizations worldwide for 20 years. As of early 2026, CiviCRM has over 13,000 active installations (see https://mastodon.social/@civicrm/115961405957089637). The software is typically used in non-profit organizations as a tool for operational tasks such as handling bulk email distribution, generating serial letters, planning appointments and activities, or managing processes and cases. The ability to manage contacts and the relationships between them often serves users as a powerful address book and for grouping target audiences. In addition, there are functions for secondary processes within an organization, such as fundraising, issuing and managing donation receipts, membership management, and the organization of events—from invitation and registration processes to participant management, fee billing, and the issuance of participation confirmations.

Those who work with CiviCRM—whether as users or as external facilitators of its implementation—often have an ambivalent experience. On the one hand, there is a feature-rich and highly capable system that makes the complex work of a non-profit organization manageable in the first place. In this sense, “the system does what it is supposed to do.”

Nevertheless, a sense of friction often arises. Installing and maintaining the software is demanding. The quality of features and usability varies considerably across the extensive system. New users require a long time to become productive. Even experienced users regularly rely on external support. Users report that they are theoretically aware of certain functions but avoid them in practice—not because the functions are unsuitable, but because their application is demanding. A low frequency of certain usage scenarios leads to insufficient learning of correct usage. A lack of intuitive usability and equally insufficient documentation repeatedly necessitate costly re-appropriation. In all these cases, recurring project-based workarounds, repeated training needs, or even the substitution of functions that are in principle available by other additional software systems become the norm rather than the exception.

This experience is not a marginal phenomenon resulting from isolated misconfigurations or unfortunate project trajectories. It recurs across types of organizations and contexts of use. It occurs both in small associations and in large, resource-rich organizations. And it persists even when the system has been operated stably for years.

Notably, this friction is rarely described as a “malfunction.” Hardly anyone would claim that CiviCRM is fundamentally defective. Rather, the impression arises of a system that is convincing in its domain logic, but whose use must be continuously mediated and supported. Working with the system is often inseparably linked to working on the system.

In addition to the aspect of user experience, a second, no less striking observation emerges. Measured against its capabilities and the breadth of possible use cases, CiviCRM is still used by surprisingly few civil society actors.

The market of potential users—journalists and media professionals, social entrepreneurs and public figures, politicians at all levels, initiatives, citizen participation projects, associations, parties, trade unions and union-related organizations, foundations, charitable institutions, and many more—is significantly larger than the actual user base would suggest. This is even more so if one considers that the software can be used not only at the central level of organizations but also independently within their subsidiary units. For the German-speaking region, it can be roughly estimated in early 2026 that there are probably no more than around 1,000 active installations in use. At the same time, the German-speaking region—alongside the United Kingdom—is one of the most developed markets in Europe to date. Market penetration therefore clearly falls short of its potential.

This limited diffusion calls for explanation. It can hardly be explained solely by a lack of demand or by superior competing offerings. The requirements addressed by CiviCRM are widespread. Adequate alternatives hardly exist, and where they do, they are mostly proprietary systems whose pricing targets upper market segments with substantial resources. Beyond CiviCRM and its proprietary challengers, what remains is fragmentation: patchworks of different software systems and numerous media discontinuities—if software is used at all.

Thus, the impression emerges that a fundamentally available opportunity to strengthen organizational capacity, increase the efficiency of civil society action, and ultimately support democratic participation is only realized to a limited extent. The potential exists—but it is not being fully exploited. The question is: why?

In the following analysis, I consider CiviCRM as a good in the sense of the theory of goods in political economy. I address the question of the optimality of its provision. With regard to FOSS, using CiviCRM as an example, I distinguish between two levels of provision:

At a first level, the provision of software functionality can be examined. This concerns whether a system is capable of fulfilling its intended tasks: whether it provides the necessary data structures, maps core business processes, operates in a technically stable manner, and can be maintained with reasonable effort. Even at this level, a structural suboptimality of provision becomes apparent. The provided functionalities are primarily aligned with the requirements of those actors who bear the costs of procuring and further developing the software. For these actors, the functions they procure are sufficiently good: they enable the fulfillment of their own tasks and thereby justify their investment. Functions and qualities of functions that are of lesser importance to the procuring actors, however, are only partially considered or remain underdeveloped. This is particularly the case where procuring actors are able to compensate for the deficits of the procured good through their own professional or technical competence, such that the additional effort required in usage appears lower than the effort of procuring a more optimal good whose usage would require less effort.

What is “good enough” for the procuring actors in this way need not be equally sufficient for other user groups—especially when these others are unable to compensate for the deficits of the procured good due to lower levels of competence.

From this perspective, CiviCRM is sufficient. The system is powerful, technically stable, future-proof, and covers a wide range of core use cases. Existing deficiencies are specific enough that they do not undermine the primary purpose of use. Nevertheless, the provision of software functionality cannot be considered optimal, as it is shaped by selective utility orientations of the procuring actors, which systematically underprovide requirements of other user groups beyond this circle.

The second level of provision concerns the capacity for appropriation of the software—more precisely, the ability of users to appropriate both the usage and the possibilities of the software themselves, to use it independently on a long-term basis, and to expand their use autonomously within the possibilities provided by the system.

At this level, the question is not whether functions exist, but how accessible they are to users: how easily it can be recognized which functions exist, how understandable the underlying concepts and terminology are, and to what extent the system supports users in developing an independent mode of use.

At this second level, provision is far less optimal than at the first level. Available documentation often refers to the version in which it was originally created. It is also highly fragmented, as different authors, with different motivations, have produced texts or videos on using the software at different times, in different places, and for different versions—if at all. As a result, search engines and cloud-based AI services often provide references to outdated problem-solution pairs or speculate about possible modes of use that have since become obsolete due to changes in the software. A technically secondary but practically relevant aspect is the ongoing change in translations of terms and phrases used in the user interface.

Not only the use of the software, but also the available documentation often presupposes prior knowledge. User interfaces and functional logics are based on conceptual assumptions that cannot easily be derived from use alone. The appropriation of the software is generally not a linear, step-by-step learning process. Once-recognized logics cannot be reliably extrapolated, as the logic implemented in one area does not necessarily apply to another area within the same software. As a result, usage repeatedly depends on external knowledge. Autodidactic appropriation is in principle possible, but it requires considerable additional effort. The knowledge gained requires maintenance and often remains tied to individuals and their specific competencies.

In the case of CiviCRM, it becomes apparent that suboptimality at the first level has consequences for the second level, while underprovision at the second level also exists independently. The functional orientation toward the requirements of those who bear the costs of development enables solutions with many configuration options that are manageable for these actors due to their competencies. Other user groups, however, face both the underprovision at the first level and at the second level, as they experience the solution space itself as a problem and are unable to operate it effectively.

The fact that the software is licensed as free and open-source software does not in itself mean that it can easily be procured and appropriated. First, it requires the competence to install and continuously maintain it. Second, the available functions remain difficult to access, may not be fully usable, and may remain dependent on external support in the long term.

The limited capacity for appropriation is not merely a problem of usage but a structural barrier to access. It affects not only existing users, but even more so potential new users, for whom the effort required for implementation and initial exploration acts as a deterrent. As a result, the diffusion of the otherwise highly capable and socially relevant CiviCRM system remains below what would be socially desirable and technically and economically possible.

This diagnosis implies neither individual failure of specific actors nor a fundamental critique of the software or its community. Rather, it points to a structural tension between selectively available software functionalities, the difficulty of their appropriation, and the diffusion of the software, which remains below its potential in terms of its societal impact. Why this tension is not accidental and why it stabilizes despite good intentions is the subject of the following sections.

2. Suboptimal provision as a result of rational decisions

It is tempting to read suboptimality as an accusation: as an attribution of blame toward those who carry the development of the software. However, this would be inappropriate and would do justice to no one.

From personal observation, I can confirm that the credo of CiviCamp 2025, held in the Netherlands, holds true: “Came for the software, stayed for the community.” Without its community, the software would likely not have reached its twentieth anniversary in 2025, nor would it be in use worldwide. Without this shared sense of responsibility, the software would possess neither its current technical maturity nor its demonstrable future viability.

Nevertheless, the described underprovision exists. The observed effects can occur in political-economic terms even when all participants act rationally to the best of their knowledge. Suboptimal provision does not result from intent, but is an emergent effect of rational decisions.

The development of software such as CiviCRM continuously requires prioritization decisions. Time and human resources are limited. The demands placed on the software vary in urgency. It is therefore natural to invest first where— from the perspective of those bearing the costs—the greatest immediate benefit is expected: in fixing errors or adapting to the specific requirements of real-world projects. The revision or expansion of existing functionalities, as well as the addition of entirely new features and measures to improve the capacity for appropriation, are typically addressed last.

Improvements have low urgency as long as the software does what it is supposed to do for the cost-bearing users. Expanding the functional scope generally requires at least one cost-bearing user or organization with a sufficiently strong incentive to finance the development, publish the extension under the GNU Affero General Public License Version 3 (AGPLv3), and commit to long-term maintenance responsibility for that part of the software. As soon as the expected costs of achieving improved or extended functionality exceed the budget and anticipated benefits for those willing to bear the costs, such functionality is not procured. This can be observed concretely in forums and Git repositories wherever feature requests labeled “need funding” remain inactive.

While investments at the functional level are relatively attributable and have short-term effects, this does not apply to investments in the system’s capacity for appropriation. User-oriented documentation, the establishment of consistent terminology, the standardization of intuitive user interfaces, or the creation of systematic learning paths unfold their effects with delay. Although they would benefit a broad, partly future user base and are therefore strategically attractive for the community, already competent users have little incentive—given their limited resources—to invest in guiding usage through user-oriented documentation, maintaining such documentation continuously, or identifying their own blind spots resulting from habitual use. Under these conditions, rational behavior leads to measures that improve the capacity for appropriation being acknowledged as desirable, yet regularly treated as secondary in concrete decision-making situations.

All participants are affected by these phenomena. Software developers focus on solving specific technical problems. Actors who support the implementation of the software through consulting, configuration, and user training align their work with the requirements of paying clients. User organizations focus on maintaining their own operational capability. In none of these cases is there a motive that would suggest a deliberate neglect of collective interests. Rather, the rational decisions of all participants aggregate into a systemic outcome that results in suboptimal provision from the perspective of the overall system—a pattern characteristic of the provision of collective goods.

It is noteworthy that this form of suboptimality is stable. It does not produce immediate crises or system failures. It forms a robust, albeit inefficient equilibrium. This very stability makes it difficult to recognize suboptimality as such. It does not appear as a problem, but as the normal state: the system remains usable, projects are implemented, organizations continue to function—even if the quality of provided functionalities varies or remains insufficient, the capacity for appropriation remains low, and the diffusion of the application remains far below its market potential. This is a case of rational suboptimality.

3. Specialized solution for a few – or infrastructure for all?

3.1 Suboptimal, but stable

Suboptimal provision is not a temporary condition, but a stable development path. Limited quality and performance at the level of software functionality, limited capacity for appropriation, and consequently comparatively low diffusion of the software reinforce each other.

Within this path, the implementation and use of the software remain permanently dependent on intermediary actors. New users join only hesitantly. Parts of already existing functionalities remain unused—not because there is no use for them within the organizations using the software, but because the costs of appropriation are too high and they can usually only be accessed with external guidance.

From a transaction cost perspective, the central scarcity lies not in the software itself, but in making it usable. For new users, significant costs arise in understanding, adapting, and learning the system—costs that are not offset by the absence of licensing fees or by the formal openness of the software. These costs are real and for many organizations not manageable. This is why, in some cases, the costs of appropriation themselves become the subject of dedicated funding programs, for example by foundations such as the German Foundation for Civic Engagement and Volunteering.

The transaction costs associated with the quality of functionalities and the costs of appropriation are accepted as an inherent part of using the software. In some cases, staff turnover within a user organization leads to the loss of established practices, as user knowledge is strongly tied to individuals. In handovers to successors, sometimes only the operation of core functions essential for the organization’s survival is transferred—or the handover fails entirely, making renewed external consulting and training necessary. As a result, the software remains a promise—a kind of magical black box—whose potential is often not fully realized.

On the supply side, the agencies offering implementation services further reinforce this development path. The welfare potential of the good CiviCRM is not realized also because—and as long as—the capacities of the already limited number of implementation agencies remain fully utilized over months and years. This capacity constraint coincides with relatively high market entry barriers, so that no significant number of new implementation agencies emerges, while existing agencies themselves grow only minimally. Although the number of employees increases steadily, the growth over multiple years remains—when aggregated—in the low double-digit range in absolute numbers, even if this may represent a doubling for individual agencies. Recruitment of new personnel often occurs endogenously: new implementation consultants frequently emerge from the pool of users who have previously become familiar with CiviCRM through practical use and gradually grow into advisory roles based on this experiential knowledge.

Heuristically, it can be concluded that the currently small number of competent implementation providers in the German-speaking region is not an expression of limited demand, but the result of high entry and scaling barriers. The impression of a narrowly bounded circle of permanently fully utilized actors therefore reflects less the actual size of the market than a structural narrowing of market access.

It is noteworthy that the market thus remains underprovided not due to a lack of demand, but due to the lack of scaling on the supply side. The sustained capacity utilization of implementation agencies does not act as a growth impulse, but as a structural brake on growth. The existing scarcity thus reproduces itself.

However, this development path is by no means dysfunctional. The provision is sufficient to enable what is essential: the continuous operation of core processes within the organizations that are served. This is why it is stable. At the same time, this path limits the reach of the system: usage remains concentrated among those actors who possess sufficient resources, competencies, or access to external support. The possibilities that the system fundamentally enables are thus realized only in part of the potential field of application.

3.2 Unleashing civil society

The development path of stable suboptimality is functional because it makes the procurement of CiviCRM possible in the first place and thereby ensures the existence of the software. In this sense, it fulfills a necessary condition for its emergence and continued existence. However, this path does not promise further development, but merely stabilization at a persistently limited level.

Within the CiviCRM community, however, developments are emerging that point beyond this state. Approaches such as containerization through Docker to enable one-click installations, the redesign of user interfaces through new themes such as the “Riverlea” template framework, or the simplification of interfaces through extensions like “Ducktape” indicate a development path that is no longer solely aimed at preserving the status quo. Rather, it is directed toward increasing the depth of use among existing users while simultaneously expanding the overall diffusion of the software.

The effects of such a shift extend beyond CiviCRM itself and have a societal dimension. From a political-economic perspective, the current development path can be described as an under-realization of a collective good. The welfare potential arising from the use of CiviCRM as an infrastructure for civil society self-organization is, under the conditions of the current path, largely unrealized. The decisive factors are high transaction costs and costs of appropriation, which limit access to the software and increase the cost of its use.

The welfare gain would consist in activating previously latent potentials and thereby overcoming the path-dependent loss of efficiency.

The costs that currently prevent this welfare gain are not an immutable feature of the software. They are the result of a path-dependent development logic that functions sufficiently well for the actors who sustain it and stabilizes precisely for that reason. The associated loss of efficiency is systematic: it does not result from individual misjudgments, but from a structure in which existing arrangements are reproduced as long as they fulfill their minimum functions.

Leaving this path does not imply a radical break, but an increase in efficiency. Declining entry barriers and a consolidated provision of functionalities would make it possible to realize a larger share of the currently unused capacities for action, both within individual organizations and at the societal level. Resources that are currently tied up in work on the system as a result of suboptimal provision could increasingly be directed toward substantive work in pursuit of organizations’ core missions—and this across a significantly larger number of organizations and with a greater number of users. The societal output of civil society activity would increase.

That such a development appears difficult to imagine today is itself an expression of the described path dependency. It is based on the widespread assumption that the high initial costs of adoption and use of CiviCRM are structurally unavoidable and would necessarily scale proportionally with broader usage. This assumption overlooks the fact that efficiency gains consist precisely in decoupling growing usage from a proportional increase in mediation, implementation, and support costs. A broader use of the software therefore does not necessarily imply linearly increasing revenues for implementation agencies.

The difference between the two described development paths concerns not only technical or organizational questions, but the horizon within which suboptimality is either accepted as an unavoidable side effect or understood as a condition that can be shaped. Since the CiviCRM community is driven by a shared commitment to work for public-benefit purposes, an orientation toward broader and more effective use of the software is already inherent.

At the same time, even members of the CiviCRM community are often unclear as to why, despite shared objectives, high levels of expertise, and continuous engagement, systematic underprovision persists at the level of software functionality and the capacity for appropriation. The explanation lies not in the intentions of the participants, but in the structural conditions of collective provision. These conditions are the subject of the following theoretical discussion.

4. Collective goods and rational underprovision: the core problem

Suboptimal provision requires explanation. How is it that it persists, even though the software has been developed for 20 years, even though the needs are clear, even though expertise is available, and even though continuous work is being done on the further development of CiviCRM? It seems implausible that the cause of underprovision can be attributed to persistent individual misjudgments, lack of effort, or lack of insight. It is far more plausible to assume that this is a structural problem that can be explained using the theory of collective goods, as developed in particular by Mancur Olson.

Olson addresses the question of why certain goods, whose benefits accrue to many, are provided at levels below what would be socially desirable in both qualitative and quantitative terms. His central insight is that underprovision, under conditions of collective good provision, is the unintended result of the interplay of rational decisions by individual actors.

The core problem can be formulated simply: if the costs of providing a good must be borne by individuals, while the benefits accrue to many or even all—because no one can be excluded from the use of the good once it has been provided—a high collective return is generated, but also a strong incentive to use the good while, if possible, avoiding participation in the effort required to provide it. Even actors who explicitly recognize the collective benefit of a good are thus placed in a situation in which it appears rational to avoid or postpone their own investments as long as others might bear the costs.

According to Olson, a collective good is provided at all only if the actor bearing the costs achieves a sufficiently large share of the benefit despite the inevitable use by others—at least enough to justify the costs incurred.

Suboptimal provision arises precisely because the procuring actor provides the good only to the extent that their own demand for benefit is satisfied. The cost-bearing procuring actor will cease provision as soon as the marginal benefit of additional features of the good no longer justifies the marginal costs incurred.

Crucially, this stopping point is not random, but structural. The collective good is not provided up to the point at which collective benefit would be maximized, but only up to the point at which the private marginal benefit of the cost-bearing actors equals their individual marginal costs. The difference between these two points constitutes the systematic core of underprovision.

In the case of software, this stopping criterion is particularly pronounced at the level of comprehensibility and usability. While software functionalities generate immediate benefits for the procuring actors themselves, investments in documentation, consistent interaction logics, or improved user interfaces primarily generate benefits for others. The private marginal benefit of such measures is low for those bearing the costs, even though the collective benefit would be high. From Olson’s perspective, it is therefore rational to stop provision at the point where one’s own use is secured—even if the software remains difficult to access for other user groups.

At this point, the needs of those who use the collective good without contributing to its costs are not yet satisfied, but those who are not involved in procurement are also unable to bear the marginal costs required to acquire additional beneficial features. Moreover, their incentive to contribute to these marginal costs is weakened by the expectation that another actor—whose marginal benefit would be higher—will bear these costs. In this way, a stable suboptimal provision of collective goods emerges.

Olson points out that this problem intensifies with the size of the collective, since the individual contribution of a single actor becomes statistically less significant—an effect that is particularly pronounced in successful open-source projects.

Applying this reasoning to CiviCRM makes it clear that central aspects of the software have the character of collective goods. This applies to the level of software functionality and especially to the level of the system’s capacity for appropriation. Improvements in documentation, consistent terminology, coherent business and interaction logics, intuitive user interfaces, and optimal training materials would benefit a large number of users. However, with regard to the marginal costs of their provision, actors tend to rely on others, while their own marginal benefit remains sufficiently high to instead make do with existing solutions and benefit from the available good without further investment. Moreover, the costs of such collectively beneficial improvements are real and compete directly with other, more urgent expenditures, so that goodwill alone is usually insufficient to finance their provision.

Under these conditions, it is rational for users of the software to prioritize investments in project-specific adaptations, the resolution of acute functional problems, or extensions that directly benefit a clearly defined group of users. Such measures are economically sound. By contrast, investments in collectively effective qualities—particularly in consolidating already procured functionalities and in improving the capacity for appropriation—are structurally disadvantaged. The resulting underprovision does not arise despite rational action, but precisely because of its consistent application.

It is noteworthy that this underprovision does not lead to instability or system failure. Olson shows that collective goods can be provided on a long-term basis even when this occurs at a limited level. As long as a minimum level of functionality is ensured, no acute pressure arises to fundamentally change the structural conditions of provision. In this way, inefficient equilibria stabilize: the system remains usable, fulfills its core functions, and thereby continuously justifies the individual decisions made—even though its potential remains far from fully realized.

Olson’s theory is therefore well suited to explaining the suboptimal provision of CiviCRM. Underprovision does not resolve itself. The issue is how to deal with it, given that it arises from rational action. As a collective good, FOSS is systematically provided only up to the point of private utility maximization of the contributing actors and necessarily remains below the socially optimal level of provision.

5. Governance as institutional handling – and the risk of its failure

5.1 Governance according to Ostrom: institutional handling instead of solution

How can this problem of underprovision be addressed institutionally? How can suboptimality at least be mitigated?

The governance approach developed by Elinor Ostrom can be applied to this question. Ostrom’s work is not a counterpoint to the analysis of Mancur Olson, but rather a complementary perspective. While Olson primarily focuses on non-rival collective goods, where one person’s use does not diminish the benefit available to others, Ostrom examines so-called commons, where use is rival, such as drinking water reservoirs or fisheries. What is decisive in Ostrom’s contribution is that observed underprovision does not arise solely from the properties of the good itself, but depends substantially on the design of governance.

Ostrom does not begin with the causes of collective underprovision, but with the question of the conditions under which collective action remains possible despite persistent incentive problems. Governance—understood as the totality of formal and informal rules, procedures, and practices—does not, in this sense, represent a solution to the collective action problem described by Olson, but rather a mode of its institutional handling. Through governance, contributions are coordinated, expectations stabilized, and conflicts rendered manageable, without eliminating the underlying incentive problems.

Governance therefore does not necessarily lead to an optimal provision of collective or commonly shared goods. It can, however, improve the conditions. This distinction is central. Governance operates within the limits described by Olson; it can structure decisions, but cannot guarantee that collectively desirable outcomes will be realized to a sufficient extent. The benchmark of governance is therefore not its mere existence, but its effects.

For CiviCRM, a governance structure exists. The development of the software is embedded in organizational responsibilities, decision-making processes, and publicly accessible discourses. The community has institutional arrangements that enable coordination and ensure continuity. However, whether and to what extent these arrangements are capable of addressing precisely those aspects of the collective good that are structurally underprovided remains uncertain. Ostrom’s perspective suggests that CiviCRM should not be treated as a monolithic common good. Governance must instead explicitly distinguish between sub-commons whose usage and contribution logics differ: (1) core code and architecture, (2) modules and extensions, (3) documentation, (4) UX/UI patterns, and (5) knowledge resources such as best practices and use-case references. These sub-goods have different user groups, cost structures, and externalities. A further development in the spirit of Ostrom would therefore initially consist in making these boundaries visible and establishing distinct governance spaces for these sub-goods, rather than implicitly subsuming them under the core process or market logics alone.

Ostrom should not be read normatively in this context. Governance is not a promise to overcome collective underprovision, but a mode of dealing with it. If this distinction is not observed, governance risks becoming a symbolic response to a persistent provisioning problem.

A note on Governance structures of CiviCRM

Global level

The formal sponsoring organization of CiviCRM is CiviCRM LLC, based in the United States, specifically in San Francisco. It provides the legal and organizational framework of the project, including trademarks, domains, and central infrastructure.

The official project website is https://civicrm.org. It bundles information on releases, community structures, events, and governance-relevant processes.

The technical development of CiviCRM is publicly accessible and fundamentally organized in a decentralized manner, but it does not take place uniformly on a single platform. While central repositories—including the publicly visible repository at https://github.com/civicrm—primarily serve distribution, documentation, and low-threshold onboarding, the authoritative development work, including merge decisions, reviews, and release preparation, takes place primarily in the self-hosted development environment at https://lab.civicrm.org.

Feature requests, bug reports, and technical discussions are distributed across both platforms and are sometimes mirrored, sometimes deliberately kept separate. Decisions emerge less through formalized voting procedures than through procedural consolidation: through ongoing discussion, technical review practices, and ultimately through the actual integration of code into the core of the software.

Influence on further development is primarily exercised by those actors who continuously contribute in the form of code, maintenance, review, infrastructure work, or organizational coordination. The governance of the software is thus organized in a contribution-based and performance-oriented manner; it is not designed as a representative system of the entire user base, but implicitly follows the structures and requirements of the development process itself.

Forums and communication platforms

For exchange and coordination, the community uses Mattermost as its central discussion forum, accessible at https://chat.civicrm.org. In addition, the actively used Q&A platform “CiviCRM Stack Exchange” is available at https://civicrm.stackexchange.com/. Previously, the forum at https://forum.civicrm.org, which is still searchable as an archive, was used.

These formats serve to make topics visible and to enable informal coordination.

European level

At the European level, there is no independent legal or formal governance body with binding decision-making authority. Coordination takes place primarily through networks and recurring community events. Of central importance are the CiviCamps. CiviCamps are open, self-organized community conferences that facilitate exchange between developers, implementation actors, and users. Information on CiviCamps can be found at https://civicamp.eu/.

CiviCamps function as forums for articulating problems, discussing priorities, and initiating joint initiatives, but they have no formal decision-making authority. Participation requires physical or organizational presence.

German-speaking region

In the German-speaking region, the association Software für Engagierte e.V. (https://www.software-fuer-engagierte.de) assumes an intermediary role. The association acts as an organizational sponsor for community activities, supports exchange and qualification, and is involved in organizing CiviCamps. It serves as a point of contact for civil society organizations and implementation actors in the German-speaking region. Membership and active participation are voluntary and dependent on resources; a formal representation of all users is therefore not established.

Ostrom’s principle of polycentric, nested governance (“nested enterprises”) would suggest not leaving these levels as loose networks, but developing them into thematic or sectoral units with clearly defined responsibilities—for example, “CiviCRM in the political sphere,” “membership-centered use cases,” or “CiviCRM for foundations.” Such units could develop context-specific standards, reference setups, and learning paths in a binding manner and feed them into overall governance. In this way, maturation within the core could be achieved without fragmentation through isolated special solutions.

Risk assessment in light of political developments in the United States

In light of current developments in the United States, it is reasonable to ask whether CiviCRM could be endangered by the location of its sponsoring organization in the US. After all, the operational processes of numerous civil society organizations, both within and outside the United States, rely on the use of CiviCRM. Putting CiviCRM under pressure could therefore represent a lever of interest for political adversaries seeking to exert pressure on user organizations as well. Due to the copyleft nature of the GNU Affero General Public License, however, the risk of political blockage is legally negligible. A fork of the software could be created at any time, regardless of whether CiviCRM LLC is able to cooperate, US servers are blocked, or GitHub access is restricted. Regardless of the operational capacity of the LLC or the blocking of servers hosting code repositories, the already licensed and distributed code remains forkable.

The risk of a blockage scenario by the United States arises only if the governance structure is too weak to handle the coordination costs of a necessary fork. The potential vulnerability of the international community in light of developments in the United States does not lie in legal constraints, but in a possible lack of institutional redundancy.

The existence of a certain level of problem awareness is reflected in the introduction of a self-hosted GitLab instance by CiviCRM LLC and the discussions surrounding the physical location of this instance.

5.2 The economic weight of usage: an approximate quantification

The question of the adequacy of governance structures gains analytical sharpness when one considers the economic weight of existing use. No systematic statistics are available for CiviCRM in this regard. Nevertheless, the scale of committed resources can be roughly estimated on the basis of plausible assumptions.

If one assumes—deliberately conservatively—1,000 instances, each used by an average of five full-time staff members, this yields an order of magnitude of around 5,000 professional users. If one further assumes—again rather conservatively—average implementation costs of €7,500 per instance, this corresponds to a one-time committed initial capital of approximately €7.5 million. If one also calculates with monthly maintenance costs of around €150 per instance (hosting, maintenance, updates, recurring consulting), this results in an ongoing commitment of around €1.8 million per year. These amounts are not recoverable, but are amortized exclusively through continued use.

This rough quantification marks only the lower bound. Depending on the actual number of productive instances, the depth of implementation, the amount of internally committed working time, and the extent of external support, the real capital commitment and cost commitment are plausibly higher. At the same time, the non-monetary investments—capacity-building, process adaptation, data maintenance, and integration—are often even more decisive for the stability of use than the pure monetary amounts. It is precisely these non-monetary investments that generate substantial switching costs and bind organizations to the system over the long term.

 

Even under cautious assumptions, this yields committed organizational capital amounting to millions—and under more realistic assumptions (more instances, higher implementation budgets, more internal working time, and multi-year use), plausibly in the tens of millions.

Governance-relevant deficits—especially in documentation, terminology logic, and UX consistency—translate directly into corresponding productivity losses. Even small frictions scale: if 5,000 professional users lose on average only 30 minutes per week due to deficits, particularly at the level of the software’s capacity for appropriation, this corresponds to over 100,000 working hours per year—in this model calculation, 130,000 hours. Even with a conservatively estimated hourly cost of €40, the resulting welfare loss amounts to more than €4 million annually.

For many user organizations, CiviCRM is not simply interchangeable software, but part of their institutional infrastructure. Use is stable because it builds on investments already made and because switching would involve risks and costs. These switching costs are not merely abstract. Data migration (including histories and relationships), process remodelling, training, parallel operation, and risk buffers frequently amount—roughly and conservatively estimated—to between €20,000 and €50,000 per organization. This makes exit rationally unattractive for individual users, even when governance is experienced as inadequate.

There are two decisive differences between CiviCRM as FOSS and proprietary competing products. First, the use of CiviCRM does not entail lock-in effects intentionally engineered by the software vendor. In this respect, operating the software does not carry the same cost risk, because vendors cannot exploit switching costs through pricing policy, and an increase in the number of users or an expansion of the functional areas in use does not lead to price increases driven by licensing rules. Quite incidentally, this also has advantages in terms of data security, because it avoids incentives toward account sharing driven by cost avoidance.

Second, as FOSS, CiviCRM has a participatory governance structure in which user organizations can in principle participate—and, from the perspective of the following analysis, would also have to participate. User organizations are less mere “users” of a free good than de facto co-investors in an infrastructure whose value depends on long-term development and governance decisions.

Against this background, however, a mismatch becomes visible. While the economic and institutional weight of use is considerable, the participation of users in the governance of the project remains surprisingly limited. The existing governance structure therefore represents only to a limited extent those organizations that in fact bear a large share of the costs of use and whose operational capacity depends on the quality of the system.

This discrepancy makes clear that the governance of CiviCRM is not about symbolic participation or community identity, but about the institutional representation of considerable committed resources. It is plausible to assume that a large share of users, despite their economic weight, are not integrated into governance processes. The structural causes underlying this presumed non-participation therefore require clarification.

5.3 De facto non-participation as a consequence of the lack of imaginability of governance

The limited participation of users in the governance of CiviCRM cannot be adequately explained merely as the result of individual cost-benefit calculations. Even those who have committed substantial resources to the use of the software frequently do not participate in its governance. While the theory of collective goods explains why underprovision remains stable, it does not explain why users themselves stay away even where they have invested considerable resources in use. Non-participation therefore points to an institutional problem.

This problem lies less in a lack of affectedness than in a prior barrier: participation presupposes some conception of where decisions emerge, how they can be influenced, and what role users can take within them at all. Where such orientation is absent, non-participation becomes customary—regardless of how great the committed interest may be.

For many user organizations, it remains unclear how the governance of CiviCRM works and in what way decisions come about. It also remains unclear what alternative governance arrangements might look like and what differences they would make. Above all, there is no comprehensible idea of what role users themselves could assume within existing or possible structures.

As soon as governance becomes imaginable to users as relevant to action and accessible, the problems described by Mancur Olson come into play: the question then arises of who bears the costs of articulation and participation while the benefits of improved governance are diffusely distributed. But if imaginability itself is lacking, this second order of the problem is never even reached—participation then fails not because of deliberation, but because of a lack of concreteness.

Under these conditions, participation is not an obvious option. Governance does not appear as a space that can be shaped, but as a background process without clear access points. Non-participation is therefore less the result of a conscious decision than an expression of a lack of orientation. Where institutional roles, procedures, and effects are not made explicit, participation remains unlikely.

This finding shifts the analytical focus. It is not a lack of public spirit or rational restraint that explains low participation, but a governance structure that, for a large share of users, is not experienced as relevant to action. This prepares the transition to the question of the conditions under which governance may be present, yet remain limited in its effects.

5.4 Representation deficit and the risk of governance simulation

Observations suggest the assumption that, despite existing governance structures, a substantial share of users remains institutionally underrepresented—even though these very users have committed considerable financial, organizational, and personnel resources to the use of the system, while not themselves having invested in the development of the software so far. This mismatch is the starting point for the risk of a mere simulation of governance.

By governance simulation, I do not mean deliberate deception, but rather a constellation in which formal structures and procedures create the impression of effective steering without actually changing the provisioning conditions of the collective good in any significant way. This constellation is not accidental, but rationally explicable. Governance simulation arises where governance is sufficient to stabilize expectations and enable coordination without triggering the costs of structural change. It constitutes an equilibrium in which formal participation exists while those actors whose everyday work and operational capacity are most affected remain only weakly integrated institutionally. Governance is present, but its effect remains limited to the organization of existing processes, which at present are above all processes of software development. Structurally, all those qualities of the good whose relevant stakeholders are absent from the governance structure are scarcely improved in any substantial way.

The criteria for assessing this failure can be derived from Elinor Ostrom’s own work. Ostrom does not measure governance by its formal existence, but by its connection to real practices of use. Decisive here are the intelligibility of rules, the assignment of roles, the actual participation of those affected, and the possibility of adapting institutional arrangements in light of experiences of use.

By these standards, governance fails in its function when decision-making processes are not comprehensible for a large share of users and do not provide clear roles for their participation. Governance then remains formal without enabling collective action in Ostrom’s sense. It organizes procedures without actually distributing responsibility. The danger of mere governance simulation exists wherever no binding decision arenas exist for specific sub-goods. In the case of CiviCRM, this would concern in particular the quality of software functionalities, continuous improvement processes advised by users, UX standards, minimum documentation requirements, and deprecation rules. Ostrom’s principle of “collective-choice arrangements” would require here that those affected by rule changes be institutionally able to participate in determining them—not in the form of non-binding consultation, but in arenas with clear competence and tangible effects. This is particularly problematic wherever governance is not tied back to the actual weight of use.

The open question arises whether the considerable capital commitment of users finds an institutional counterpart in the existing governance structures and whether roles, procedures, and metrics exist that systematically represent or protect this committed dimension of value.

This decoupling is further stabilized by the incentive structure on the side of intermediary actors. Implementation agencies that remain continuously fully utilized face little economic pressure to work toward a fundamental change in the provisioning conditions of the good. Existing demand instead confirms the functionality of the current arrangement. In such an arrangement, underprovision does not become visible as a risk, but is normalized as a manageable condition. As long as intermediary services remain scarce and in demand, there is no immediate incentive to address effectively those structural deficits whose removal would, in the longer term, devalue one’s own intermediary role.

This constellation is not an expression of particular interests or deliberate obstruction, but the result of rational behavior within given framework conditions. The overload of implementation agencies has a doubly stabilizing effect: it not only limits the expansion of the use of the good, but also reduces the likelihood that impulses to overcome structural underprovision will emerge from this group of central actors.

A potential incentive for change arises, if at all, only indirectly from the burdens of this arrangement itself. High and durably committed utilization is not infrequently accompanied by considerable consulting and coordination pressure. Project work is characterized by shifting requirements, unclear boundaries between technical support and organizational consulting, and a high volatility of demand. Under these conditions, a desire may arise to further develop one’s own business model toward more stable, more predictable, and more standardized work processes.

Such a structural change, however, would not be an automatic result of the existing market situation, but would require that the long-term advantages of lower volatility and more predictable work organization be weighted more highly than the short-term returns from persistently high demand. The resulting incentive therefore remains contingent and individual, not collectively effective. On its own, it is insufficient to systematically overcome the structural underprovision of the good.

This decoupling becomes particularly problematic wherever governance is not tied back to the actual weight of use. If no institutional connection arises between use, costs, and influence, a representation deficit emerges. In this case, governance can stabilize existing decision asymmetries and open up points of entry for particular interests without this being contrary to the rules. Governance structures then run the risk of being criticized not as a lever for addressing structural underprovision, but as a frame of reference that deflects criticism without changing the underlying provisioning conditions. More than that: those who engage within this governance structure run the risk of having their engagement accused of being a strategy for pursuing particular interests.

The danger of governance simulation does not lie in the absence of institutions, but in their superficiality. Ostrom’s approach makes this boundary visible: governance loses its function when it no longer contributes to enabling collective action, but takes over its representation. The analysis thus shifts to the question of where the existing governance structures systematically run empty: which objects of decision, forms of affectedness, and costs remain invisible—and which institutional preconditions are missing in order to be able to articulate these blind spots at all?

For precisely this reason, further development in the spirit of Ostrom is to be understood less as a shift in power than as an explicitation of rules where implicit logics of markets, reputation, and contingency currently prevail. Differentiated sub-commons, congruent roles, effective decision arenas, monitoring, graduated consequences, and low-threshold conflict resolution would not make governance “larger,” but more functional—and would thereby reduce the danger of governance simulation: not through more rhetoric of participation, but through capable institutions known to all users.

5.5 Institutional blind spots: lack of visibility of decision objects, affectedness, and costs

The preceding sections have shown that the suboptimal provision of CiviCRM can be explained in terms of the theory of collective goods and that existing governance structures institutionally handle this underprovision. From the perspective of Mancur Olson, it appears as a rationally stabilized result of individual decisions under conditions of diffusely distributed benefits. At the same time, with Elinor Ostrom, governance was introduced as a way of dealing institutionally with this structural underprovision without ever being able to dissolve its causes completely. Against this background, the question of governance is not to be understood normatively, but functionally: as a question of the institutional preconditions under which the existing welfare potential of the collective good, so far only partially realized, can be better unlocked. What is decisive is not that governance exists, but what it makes visible and available.

Three dimensions are central here: first, the visibility of decision objects; second, the visibility of affectedness; and third, the visibility of the costs associated with decisions. Accessibility presupposes roles. As long as contributions and forms of affectedness are represented only informally and meritocratically, it remains unclear who is responsible for what and who institutionally brings in whose perspective. In conformity with Ostrom, an explicit differentiation of roles would therefore be appropriate, for example in the form of functionally delimited steward roles (core, modules, documentation, UX) as well as delegate roles (implementer delegates, user delegates). What would be decisive would not be formal equality of rank, but congruence: whoever bears the consequences of documentation or UX deficits must receive rights and obligations in those areas—without thereby necessarily co-deciding on minor architectural questions.

The central deficit of existing governance structures lies in the lack of visibility of the objects of decision. For many users, it remains unclear not only who decides, but already what is being decided and which alternatives are even under consideration. Governance thus does not appear as the site of relevant decisions, but as an abstract background process.

This is also evident in practice. Users are usually neither part of association structures nor of working groups or information distribution channels that specifically concern themselves with the CiviCRM functionalities they use or with the appropriation problems of their everyday work. Accordingly, they do not learn when exactly those questions that shape their use are being discussed, and can hardly locate their own affectedness institutionally. Affectedness thus remains an individual experience, but is not translated into a category institutionally capable of connection. Neither intensity of use nor dependence on specific functionalities or appropriation services is systematically recorded or represented. Governance therefore operates without reliable feedback as to whom decisions affect and to what extent.

This opacity has a further-reaching consequence. Where decision objects and affectedness are not visible, users also cannot understand themselves as potential bearers of costs. They recognize neither where participation in costs would make sense for them nor what collective benefit such participation would have. Nor is there any visible and efficient structure through which cost-sharing could be pooled and organized effectively. “Make it happen” and “needs funding” impulses lack the supporting breadth of the possible base of a community that could be organized by the governance structure.

In addition, costs arise not only financially, but also temporally and organizationally. Efforts for re-appropriation, training, workarounds, or external consulting are indeed experienced in everyday life, but are not made visible as collective costs. Without such visibility, a basis for decisions is missing.

Thus, not only does participation in decisions fail to materialize, but so too does the possibility of financing and prioritizing collectively effective qualities in a targeted way. Investments in capacity for appropriation, documentation, or conceptual consistency do not appear as a shared project of users, but as the concern of other actors. The structural underprovision of these areas is therefore not merely accepted, but actively stabilized by the lack of visibility of decision objects, affectedness, and costs.

These blind spots concern precisely those aspects of the provision of goods that are central to increasing welfare potential. As long as governance does not make decision questions, affectedness, and possible cost-sharing explicit, it remains limited to the administration of a suboptimal equilibrium. The following reflections therefore begin with the question of which development paths appear realistic under these conditions in order to shift this equilibrium toward improved provision—whether through collective mechanisms, through marketable bypass solutions, or through levers that systematically increase visibility and comparability.

6. Two paths beyond the illusion of a complete solution

6.1 Collective procurement: an obvious but fragile solution

Collective procurement is not an abstract ideal of shared responsibility, but a concrete mechanism: several user organizations agree to jointly finance certain measures whose benefits are primarily collective in nature.

Within the CiviCRM community, tested formats for this already exist. So-called “Make-it-happen” initiatives are used to collect funding commitments for clearly defined development or consolidation projects in order to finance them jointly. Such initiatives make collective procurement practically manageable: they pool willingness to pay, formulate concrete goals, and attempt to distribute the costs of collectively effective improvements across multiple shoulders. At the same time, however, they also reveal the limits of this approach wherever the governance structure is unable to mobilize the necessary potential, since not every initiative reaches the funding required.

From a normative perspective, this path is plausible. It addresses precisely those areas of goods provision that are structurally underfinanced even though their benefits are widely recognized.

From a political-economic point of view, however, this approach remains fragile. In Mancur Olson’s perspective, the financing of collectively effective measures is inherently unstable as long as individual contributions are not linked to clearly attributable individual returns. Even well-justified projects are, under these conditions, vulnerable to a lack of participation or to participation concentrating on a small number of particularly committed actors.

The conditions of suboptimal governance described in Section 5 intensify this problem. Where decision objects, affectedness, and possible cost-sharing are not transparent, users can only with difficulty understand themselves as part of a shared financing project. Even an existing willingness to pay then finds no efficiently functioning channel through which it could be pooled and stabilized. And even if transparency were established, collective financing would, according to Olson, remain precarious.

In addition, collective procurement generally remains episodic. “Make-it-happen” initiatives may succeed in individual cases, but they depend on continuing mobilization, attention, and consent. As a durable mechanism for securing collectively effective qualities, they therefore prove unstable—above all when the governance structure itself is also superficial.

Against this background, collective procurement appears less as a viable development path than as an analytical point of reference. It makes visible which form of coordination would in theory be capable of unlocking the welfare potential of the collective good, while at the same time showing why such coordination comes about only with limited reliability under real conditions. It is precisely this limitation that forms the contrast against which a second path gains contour, one that dispenses with collective acts of decision-making and financing and nevertheless aims at improving the provision of the good.

 

6.2 Strategic consolidation through implementation agencies: a marketable bypass path

The second path does not rely on collective procurement decisions, but on individual decisions made by market actors.

Implementation agencies are intermediary actors that reduce the transaction costs of appropriation. Their business model is based not primarily on providing the software, but on translating a suboptimally procured collective good into an individually usable application. As long as these individual costs of appropriation remain high, there is, from an economic point of view, no structural incentive to eliminate the causes of this scarcity completely. Underprovision is therefore not merely tolerated, but becomes a functional component of a stable market arrangement. While governance structures can create the expectation that underprovision is being institutionally addressed, implementation agencies enable its practical compensation in exchange for payment. This, moreover, creates a very real danger of a mere simulation of governance insofar as governance processes are shaped primarily by implementation agencies. From the users’ perspective, this then constitutes a principal-agent problem: the users as principals, the implementation agencies as their agents who are difficult to control.

Implementation agencies operate in direct contractual relationships with individual user organizations. Their investment decisions therefore follow a logic that is compatible with Mancur Olson’s assumptions: they pay off where benefits are attributable and can be realized across multiple projects.

Against this background, the targeted improvement of the provision of core functionalities and of the capacity for appropriation can represent a strategic self-interest of a market participant offering CiviCRM. Whoever standardizes usage patterns—in the sense of stabilized application logics—provides comprehensible interfaces, and develops reproducible implementation logics reduces project-related adaptation costs and at the same time opens up market segments that differ from the usual business model of requirements-specification-driven customization and implementation. Capacity for appropriation then becomes a productive feature of an offering.

This approach works without collective decision-making. Improvements emerge through repeated individual decisions under similar conditions and solidify wherever they prove themselves in practice. In this way, collectively effective qualities can also arise without having first been collectively financed or decided upon. Governance structures are not replaced by this, but relieved, since decision objects become clearer and usage becomes more stable.

The effects of this path extend beyond individual projects. As the provision of core functionalities and capacity for appropriation matures, the costs of adoption and use decline. This makes CiviCRM attractive to organizations that have so far been deterred from using it by high costs of appropriation. Under these conditions, diffusion can increase abruptly, thereby raising the overall welfare potential of the collective good.

The strategic consolidation path thus describes not an ideal-typical escape from the collective action problem, but a realistic mechanism for shifting the existing equilibrium: away from project-specific special solutions and toward a more broadly usable, more easily appropriable, and thus socially more effective infrastructure.

6.3 From stable equilibrium to optimal provision: a constructive perspective

What changes when CiviCRM is procured up to the collectively optimal level of provision? This question is directed not at individual improvements, but at the systemic effects of a different logic of provision.

Optimal provision would concern not primarily the scope of functionalities, but the capacity for appropriation of the system. Documentation, terminology logic, user guidance, and learning paths would be designed in such a way that use would no longer depend on expert knowledge. The good would be provided in a form that is easy to use and generally accessible. In this way, that second level of provision—the provision of capacity for appropriation—would be fully procured, whereas under the current equilibrium it remains structurally underfinanced.

The central economic effect of such procurement would lie in the reduction of transaction costs. For organizations, the relevant costs would then no longer lie in understanding, translating, and securing the use of the system, but in transforming and expanding their own organizational processes. The effort of implementation would lose its deterrent character. Actors previously excluded would thereby become capable of action without intermediary services being strictly necessary. This effect is to be understood as a welfare gain because it frees up resources that are currently tied up in work on the system.

This would also change the market structure. Under current conditions, a market niche for implementation agencies emerges because there is a gap between the suboptimal provision of the collective good—which is free of charge for them—and the willingness to pay of potential users. This gap consists of the transaction costs of appropriation. In economics, arbitrage refers to the use of such cost or price differences by intermediary actors. In this sense, implementation agencies act as arbitrageurs by bridging, for payment, the difference between the low provisioning costs of the software and the high costs of appropriation borne by users. They do not sell the good itself, but the reduction of barriers to access. As long as this difference exists, this business model is rational and legitimate. As the capacity for appropriation grows, however, the room for arbitrage shrinks. The market does not disappear as a result, but changes its character. Value creation then shifts from necessary translation toward optional specialization, substantive consulting, and organizational development.

At the same time, the market expands. The currently small number of competent implementation agencies worldwide is not a measure of demand, but an effect of high entry barriers. As long as appropriation is possible only through implicit expert knowledge, fragmented documentation, and informal networks, the circle of providers remains small. As the capacity for appropriation increases, these entry costs decline. Additional service providers can then develop CiviCRM as an economically viable field without first having to accumulate years of project-specific knowledge. Out of a limited, knowledge-based oligopoly emerges a differentiated service ecosystem with more providers, greater specialization, and real competition. Optimal provision of the good does not narrow the market; it expands it.

From such a perspective, it is also possible to make a rough estimate of the potential market structure. For the German-speaking region, it appears plausible that even under conservative assumptions a substantial expansion and differentiation of the provider landscape would be possible. A narrow core of highly specialized implementation agencies for complex migrations and highly individualized integrations would remain and, at most, grow moderately. Alongside this, however, a much broader layer of general implementation service providers could emerge, capable of serving standardized implementations and typical use cases on an economically viable basis. This field would be complemented by a growing circle of individual consultants and trainers, as well as by complementary services such as hosting, maintenance, data protection consulting, or process consulting. The software would become adaptable for already established service providers, enabling them to offer the potential of their consulting services in connection with the software.

Heuristically speaking, under conditions of optimal provision of the good, one could expect not merely a marginal increase, but a multiple of the current number of providers: out of a small, knowledge-based oligopoly would emerge a regionally and functionally differentiated service ecosystem capable of addressing broader demand while also generating new demand in the first place.

Finally, optimal provision also feeds back into governance. Where capacity for appropriation is present, compensatory arrangements lose importance. Governance has less to conceal and can coordinate more effectively. Participation becomes more comprehensible because decision objects and affectedness become visible. The previously described danger of governance simulation loses weight, since structural deficits are no longer merely asserted or relativized, but become empirically verifiable for many through low-threshold accessibility.

6.4 Incentive-compatible transformation strategies

The following transformation strategies do not presuppose that collective procurement or governance will succeed reliably. Rather, they respond to the institutional blind spots described in Section 5.5: the lack of visibility of decision objects, affectedness, and costs under conditions of stable suboptimality. A distinction is made according to the conditions under which they can become effective: as a genuinely collective solution (in the sense of Section 6.1), as a marketable bypass mechanism (in the sense of Section 6.2), or as a supporting reinforcement of both paths. What is decisive is that none of these strategies depends solely on a willingness to cooperate.

A first lever is the collective procurement of user-oriented documentation, user-friendly interface design, and user-friendly onboarding paths. What is decisive here is to make the costs of appropriation visibly lower and thereby to realize an efficiency gain with collective effects. These qualities can be understood as infrastructure whose benefits unfold across many projects, but whose financing is difficult to sustain individually. Their realization appears more likely if collective procurement mechanisms actually take hold (in the sense of Section 6.1). For implementation agencies, there is a rational self-interest in this lever because of the prospect of reduced consulting effort on basic issues, so that resources on the customer side can increasingly be directed toward more advanced questions. For user organizations, dependencies and long-term usage costs are reduced. If collective financing fails to materialize, however, the question arises of what can then be achieved through entrepreneurial action in the sense of Section 6.2.

A second lever lies in making quality visible through comparable criteria, benchmarks, or certifications. Visibility here does not mean symbolic certification, but the creation of decision-relevant comparability: qualities must be presented in such a way that they can in fact be taken into account in selection, procurement, and prioritization decisions. Monitoring could be organized as a commissioned community function that makes underprovision institutionally visible. This lever can be realized both collectively (in the sense of Section 6.1), for example through jointly defined quality criteria and benchmarks, and as a market mechanism (in the sense of Section 6.2), if individual providers strategically make certain qualities visible and comparable in order to differentiate themselves from alternative offerings. The effectiveness of this lever depends not primarily on formal governance, but on whether quality information is institutionally actionable—that is, whether it systematically enters into decision-making, procurement, and investment processes. It remains ineffective if opacity persists on the demand side and qualitative differences cannot be perceived or operationalized as a relevant decision dimension.

A third lever concerns the design of governance itself, in particular the institutional strengthening of, and equal standing for, contributions to the software’s capacity for appropriation alongside contributions to software functionality or technical work. This lever is strictly bound to functioning governance structures (in the sense of Section 6.1). Without binding rules for prioritization, collectively binding quality criteria for the appropriability of interfaces and usage logics to which software releases are subjected (“release gates”), and documentation requirements, it cannot be ensured that corresponding contributions will be delivered reliably. Ostrom’s principle of graduated sanctions can here be translated into rule-bound quality discipline without moral attribution: for example, “no release without minimum documentation,” an explicit “experimental” status for insufficiently documented functionalities, or reduced prominence in the UI for unmaintained modules. The point is not punishment, but the institutional handling of rule deviations through clear, graduated consequences.

A market-based substitute is not possible here, because strategic design decisions as well as decisions about feature or merge requests, about interface design, or about translations define the free and open-source software as a set of conditions that are then given for downstream actors in the value chain, such as implementation agencies, consultants, or users. Governance can at most be relieved by market developments here, but not replaced by them.

A complementary institutional lever lies in low-threshold conflict-resolution mechanisms. Where conflicts are currently “resolved” in practice through standstill, silent withdrawal, or informal power games, a mediation instance would help—for example in conflicts between architecture and UX, core interests and implementer interests, or short-term customer logic and long-term maintenance of the common good. What is decisive is a low transaction burden: conflict resolution must be easier than withdrawal.

Public funding and procurement policy act as external reinforcers. They are neither purely collective nor purely market-based, but intervene from outside the system. Their effectiveness depends on whether funding and procurement criteria actually oblige the actors raising funds to produce open, reusable, and documented results. Such allocations of funds can both stabilize collective procurement (in the sense of Section 6.1) and favor marketable consolidation (in the sense of Section 6.2), but they do not develop an autonomous steering effect without institutional compatibility with the decision-making, coordination, and usage structures of CiviCRM governance.

A fifth lever is the standardization of recurring use cases. Reference setups, a clear separation between standardization and specialization, and reproducible implementation paths can emerge independently of collective decisions. This lever is to be assigned primarily to the marketable bypass path (in the sense of Section 6.2). Individual implementation agencies can pursue it out of self-interest in order to shorten project durations, lower entry barriers, and open up new market segments. Collective coordination is helpful for this, but not necessary.

In summary, it becomes clear that only some of the transformation strategies depend on functioning collective procurement and governance. Only certain levers remain effective even when collective solutions remain precarious or fail to materialize. But this asymmetry is also good news: after all, it opens the possibility of gradually shifting the existing equilibrium even if an institutional breakthrough fails to occur. There is justified hope that a more optimal provision of the good will arise not through a single mechanism, but through the interplay of collective and market-based dynamics under realistic conditions. Visibility thus proves to be an institutional lever. It does not replace governance, but it can relieve and effectively complement it by bringing decision objects, affectedness, and costs out of invisibility. Only where quality becomes comparable can collective and market-based dynamics begin to take effect at all.

7. Systemic risks of stable suboptimality

The following reflections summarize the problems of sunk costs and path dependencies, the underrepresentation of existing and potential users, the possibility of mere governance simulation as an emergent risk, and the scale of unrealized welfare potential once again as systemic risks of stable suboptimality.

The stability of existing conditions rests to a significant extent on sunk costs. The adoption and use of CiviCRM involve financial expenditures, organizational adaptations, and the accumulation of person-bound knowledge that can be reversed only to a limited degree after the fact. Under these conditions, remaining within the system is often more rational than switching, even if the capacity for appropriation is experienced as inadequate. Stability thus arises not from satisfaction, but from cost avoidance. Suboptimal structures gain de facto legitimacy without being based on active consent.

Governance structures primarily reflect those actors who are able to contribute on a sustained basis. User organizations with limited resources, by contrast, remain weakly represented. Perspectives concerning capacity for appropriation, learnability, and everyday usability are institutionally harder to articulate than functional or technical requirements. In addition, there is a systematic blind spot: the interests of those actors who do not use CiviCRM because they fail at the barriers of appropriation or abandon implementation altogether. These potential users remain invisible precisely because their absence can be misunderstood as a lack of demand. Governance thus orients itself primarily toward existing use and narrows its field of vision to present interests.

From the interplay of sunk costs, stable use, and asymmetric participation emerges the risk of governance simulation. What is meant is a constellation in which formal structures and procedures suggest effective steering without addressing the structurally underprovided qualities of the collective good. Governance inspired by Ostrom can in this case conceal the underprovision explainable by Olson rather than address it. Functioning systems are particularly susceptible to this: because they generate no acute pressure to act, a status quo consolidates that appears legitimate from the perspective of those involved, while at the same time limiting its own reach.

The scope of these risks becomes visible when viewed in light of unrealized potential. A limited number of productive installations with high committed investments stands against a much larger circle of potential users whose use never materializes. The forgone efficiency gains, coordination possibilities, and effects remain invisible because they were never realized. Suboptimality thus appears not as loss, but as a possibility that has so far failed to materialize. For those, however, who benefit from the suboptimal arrangement, this not-yet-realized possibility appears like a promise of inexhaustible market opportunities.

These risks cannot be resolved through individual reforms. They are the expression of structural tensions within stable systems. The task of governance therefore consists less in producing optimal conditions than in their continuous institutional handling. Governance proves itself not through harmony, but through its capacity for institutional self-observation. Where this vigilance is lacking, suboptimality threatens to become an invisible normality.

8. Proximity to everyday use as a strategically advantageous position

The preceding sections have shown that the suboptimal provision of CiviCRM is neither an anomaly nor a transitional state. It is the result of stable structures characteristic of collective goods. However, this does not imply that change is only possible through large-scale reforms, new governance bodies, or collective efforts of great magnitude. Rather, room for action arises where decisions are already being made on a daily basis: in the practical use of the system.

Such change can originate in particular from implementation agencies. Those who regularly implement, adapt, and operate CiviCRM have a precise understanding of where users encounter difficulties, which questions recur, and which usage patterns prove effective across different organizations. This proximity to everyday use constitutes a practical observational position from which strategic decisions can be derived.

Instead of treating each implementation as a unique case, it can be effective to consciously focus on a stable core and on well-reproducible use cases. When frequently used scenarios are clearly delineated, reliably implemented, and comprehensibly documented, something gradually emerges that benefits many: a shared frame of reference. The capacity for appropriation then improves not in the abstract, but in concrete terms along real usage situations that others can recognize and adopt.

Such an approach does not stand in opposition to collective governance as discussed in the preceding chapters. On the contrary, it can even relieve governance. Where usage patterns become more stable and more comprehensible, questions, interests, and conflicts can be articulated more clearly. Governance then has to engage less in fundamental debates about diffuse problems and can instead focus on manageable, tangible issues.

It is important, however, not to misunderstand this path as a substitute for governance. Strategic consolidation does not eliminate the structural underprovision of collective goods. It can, however, mitigate its consequences and make them more visible. It expands real possibilities of use while at the same time creating the conditions under which collective questions can be meaningfully posed in the first place.

From this perspective, change appears neither as a heroic act nor as a task that can simply be delegated. It emerges from the interplay of practice and institution, of individual decisions and collective attention. A broader and more effective use of CiviCRM is then not a distant ideal, but the result of many limited yet directed steps.

The decisive point is therefore not whether structural tensions can ever be fully resolved. What matters is whether they are perceived. Vigilance here means not accepting suboptimality as given, but repeatedly making it the starting point for learning—both in the further development of the software and in the further development of its governance.

This text was submitted by me in an initial version to the European Commission as a response to the Call for Evidence within the framework of the European Open Digital Ecosystem Strategy Initiative, see https://ec.europa.eu/info/law/better-regulation/have-your-say/initiatives/16213-European-Open-Digital-Ecosystems/F33370460_en

The text was originally drafted in German. This translation was produced using ChatGPT.
I reserve the right to make further revisions and additions.