Creative Commons (BY NC CA) license granted by the authors. First published on February 01, 2015

Last Modified on February  02, 2015. Please keep us updated if you adopt this model and make improvements.

Authors:  Tiberius Brastaviceanu

Problems in peer production

work in progress…

 leave a comment please, in the doc on or the blog

Based on a work within SENSORICA.


Contents

Introduction

What is peer production?

The value system

Value creation and its distribution

Rewards or benefits

Material resources

Immaterial resources

Environment of peer production

The community

Synergy of autonomous entities

The network of oasis serving a nomad society

Problems

Participation

Disorientation

Information overload



Introduction

Peer production comes in different types and the problems that may occur in process are determined by the type of the value created, by the resources used, by the value capturing mechanisms, as well as by other contextual factors. In order to understand the problems that can occur in peer production we need to build some understanding.

First, we will define peer production. Second, we will try to understand value systems, what type of value is produced, how it is distributed, how work is rewarded, and how resources are used. by following two examples, one about open source software and the other one about open source hardware. Third, we will examine different environments where peer production takes place.

What is peer production?

See Wikipedia definition of peer production and of commons-based peer production.

See also the P2P Foundation page on the same topic.

The p2pValue project has also produced a Theoretical perspective for commons-based peer production document.

In summary, CBPP is an emerging and innovative model of collaborative production frequently taking place or supported through a digital platform. It agglutinates a set of diverse areas of activities and cases that tend to be characterized by peer to peer relationships (in contrast to the traditionally hierarchical command and contractual relationships, and with limited mercantile exchange), and/or results in the (generally) open access provision of commons resources that favor access, reproducibility and derivativeness. [from p2pValue project]

NOTE: There is the subtle distinction between peer production and commons-based peer production, that we will address below.

The value system

A good resource on value and on  value systems on the OVN wiki.

Value creation and its distribution

The type of value created determines the type of processes required, the type of resources needed, the governance, etc.

For example, FLOSS (Free and Open Source Software projects) produces use value in the form of computer programs. Code is immaterial, which is very different from material products like a pumpkin or a smartphone. The means of production usually require very low cost tools, such as a personal computer. The development costs are largely driven by labor, which can be geographically dispersed. Since the spatial constraints are greatly relaxed, the need of central physical location where people work together is reduced, which would come with greater costs, would require a heavier governance and perhaps a legal wrap around the organization. An open source software community can be informal from a legal perspective. Temporal constraints are also greatly relaxed, since collaborating on code doesn’t require synchronicity, most of the time. People can work whenever they can on different parts of the code, and coordinate through various communication channels. The number of individuals who can collaborate on the code is only limited by the coordination. These characteristics make peer production of open source software a natural for a long tail process. Developers can literally swarm a project. Since the production activity can reach large scale in participation, new phenomena can emerge. For instance, the production process can become highly redundant, meaning that at any given time there can be enough individuals who can complete a particular task, to reach a probability close to 100% for the task to be completed in effective time. Hence only light governance is needed, since the need for coercive measures to force someone in particular to complete a task. Furthermore, errors or vandalism can be effortlessly reversed in the case of immaterial things. Hence there is no strong need for policing the system, and the rules of conduct are also relaxed. If there is a disagreement on a development path to take, the process can be forked, which splits the team, thus reducing the amount of resources allocated to each development branch. But this is usually not a significant loss, since developers can still alternate between branches, and because there is an abundance of developers across the planet. Therefore, decision making processes can be lighter.

Once the code is made, it doesn’t require care, maintenance or replenishment. It can be stored somewhere, and only updated from time to time, which is part of the continuous development process. Moreover, immaterial products have negligible costs of reproduction and distribution.

Open source software cannot be sold by itself, because it is available for free, for whoever wants it, whenever it is needed. Revenue can be generated by offering services around it.

Content production systems like Wikipedia follow almost the same pattern.

Things are very different with material products. Some of them impose very strong spatial constraints, like agricultural products for example. Plants grow in a specific geographical location, therefore collaboration in the production process must be localized. They also impose strong temporal constraints, which go with the lifecycle of the plants, and that requires synchronicity, or a tighter temporal coordination. Moreover, since land is used for growing plants, the production process is tied to the forms of property that we apply to land. Food production is also highly regulated. All that implies specific requirements on governance and raises some serious legal concerns.

Mobile phones can be designed and their mechanical, electronic and optical properties can even be simulated using computers. The design process has been almost entirely virtualized and can follow the same pattern as open source software development. At the design level, open source hardware development has a lot in common with open source software production. The difference arises when we go from the virtual to the material, from the design to the prototype, and to the material product. The pressure to localize the prototyping process increases with the costs of the needed physical infrastructure. Prototyping a smartphone requires a physical location (a lab) and specialized equipment.

Rewards or benefits

Not all peer production processes are self-sustainable. Some are supported by other entities, they are parasitic. In other words, the actors have a separate source of revenue, or the resources needed in the process are supplied by entities that engage in other self-sustaining economic activities. These are also called volunteer processes or gift economies. Some of these settings don’t offer direct tangible rewards and their outputs are freely distributed as creative commons or public goods. That is not to say that actors don’t get value out of their activities, which can come in the form of reputation, sense of belonging, satisfaction from participating in the creation of social value, experience, etc. But this picture is mostly valid for small operations.

The situation is much more complex in larger and more diverse ecosystems like  Linux, which is open source operating system (software) available for free. The product (the Linux OS) is the the resultant of a collusion of multiple interests (individual, commercial, academic, ...). The incentive system can only be understood at the ecosystem level, taking into consideration the reality of every actor. We are going to explore below this type of environment in more detail. For now, we need to understand that Linux cannot be understood simply as a parasitic system, since many of the actors that take part in development (value creation) seek a direct reward by incorporating the the Linux OS into their offerings, adding value to their products. The same can be said about Android, a mobile OS.

Since such networks or communities don’t need to deal with redistribution of tangible rewards their infrastructure and their governance can be simplified.

There are many models for capturing tangible value from peer production. The most common example in the realm of software is RedHat, a company that offers services around Linux. This is a hybrid model, mixing a very diverse open network that creates open source software, with classical commercial entities that captures value by offering services. In some cases the value capturing mechanism is dominated by a single large entity. In others (see Tikiwiki), it is widely distributed throughout the community, in which case actors are at the same time independent entrepreneurs, acting as consultants, and programmers (creators of value). For simplicity, we make abstraction of support roles such as community building, communication and branding, coordination and facilitation, etc.

A similar hybrid structure has been successfully implemented in the realm of hardware, as you can see with Arduino and 3D Robotics. In this case, the classical commercial entity incorporates the functions of manufacturing and distribution. The open community is the locus of innovation, incorporating functions such as customer support, early adoption and feedback, etc. Usually, capacity of production and access to market in these settings are dominated by a single commercial entity, brakes the p2p symmetry.  

Material resources

In order to understand peer production in its purest form we need to look at Bitcoin, a p2p value exchange platform that informs us about the fundamentals of pure p2p processes. Bitcoin is built on an entirely decentralized infrastructure that allows p2p exchanges of value. It is a network of computers (or mining machines), owned and operated by individuals (or organizations). The output of this network is use value, i.e. a platform to exchange value in a p2p fashion, a payment system, a currency. Bitcoin has its own incentive system for building and maintaining its infrastructure: minors are rewarded with bitcoins, which in turn are used to expand the mining capacity and to maintain the computers used in these operations. MaidSafe (the new decentralized Internet) uses the same system of incentives to deploy its network, by issuing a coin that will be used to reward nodes (computers owners) for storage and routing of data, which are essential functions of the network. Bitcoin and MaidSafe produce use value.

Now we need to see how this applies to systems that produce exchange value in the form of products (including services) that can be sold.

Entirely distributed resource systems

In peer production, individuals can be seen as nodes of a network. They become sources of value in a process of production by relying on their intellect and wisdom, by coordinating their labor, and by using resources (brought by the same individuals) like tools, equipment, physical space, and money. By analogy with the Bitcoin value exchange system, the purest form of peer production is entirely decentralized in terms of assets. There are no common pools of shared physical resources, i.e. owned by an association, or by a trust (a custodian) which would guarantee access to a set of individuals, under some rules.

Sensorica affiliates, in collaboration with other partners united under the OVN banner, are building network resource planning tools (NRP) that allow individuals to share their material assets used in a particular process as a form of investment, with specific rules of access/use, set by the owners of these resources. For example, the equipment is tagged with QR codes that point to a database containing the inventory of all the available resources. The owners of these resources set the access/use rules. Only individuals that meet the access requirements defined by the access rules can use the equipment in a lab context, in a process or a project. The picture below shows an example of a 3D printer located in the Sensorica Montreal lab.

In reality, it is very difficult to undertake complex hardware development projects in a pure peer fashion. Moreover, it is also not clear how such a pure peer production environment would evolve, since some network affiliates might decide to exclude access based on reasons that are not aligned with keeping the processes open and with sharing, which are required to increase efficiency. In other words, some personal motives might undermine collective output.

Nondominium property

On top of a distributed resource planning system, SENSORICA has also implemented the concept of pool of shareables, which contains a mix of privately owned physical assets and assets that take the nondominium form of property. Items in this second class are legally owned by a trust, a custodian that guarantees access to all the affiliates, under some rules, through a custodian agreement. Historically, these nondominium resources have been purchased by grants or have been donated.

[Within SENSORICA, the affiliates are organized as a loosely connected non-registered association, which is compatible with the requirement large scale p2p practices, having non-contractual relationships, being fluid, dynamic and globally distributed.]

The nondominium form of property is centralized under the custodian, making the peer production process less pure and more complex. Access is widely distributed under the custodian agreement. Once we put physical resources into a common pool, we need to put in place mechanisms for maintenance and replenishment, we need a consensus on access rules, etc.

Maintenance of assets under the nondominium form of property can be done voluntarily by network affiliates, but history shows that this alone cannot be a viable solution. In order to incentivize maintenance activities, a percentage is extracted from commercial use by network affiliates of these assets, and paid to those who provide the maintenance labor and the parts, as needed.

Lab space is treated as nondominium and lab expenses are paid using the same scheme. SENSORICA affiliates practice a 5% tax on commercial activity in order to pay the necessary expenses (rent, internet access, insurance, etc.) and maintenance. This tax is administered by the custodian. The consensus is that this percentage will be adjusted according to the needs, by a democratic decision.

In case the 5% tax extracted from commercial activities is not high enough to cover all the expenses and costs to maintain the infrastructure the network has two options: secure some type of funding (grants, crowdfunding, etc.), or ask network affiliates for contributions.

A hybrid model

A hybrid model is under implementation within SENSORICA, which allows private property to be transferred into nondominium. The example is the co-purchasing of a 3D printer. A group of individuals co-fund the acquisition of an expensive piece of equipment to share the load and the risk. The participants sign a limited co-ownership arrangement. They provides access to the equipment to other network affiliates, under the condition to pay a percentage of revenues generated by its commercial use to pay back the equipment, plus 20% interest. The investment made by those who co-purchase the equipment is treated almost like a loan to the network, without the obligation to pay it back entirely if not enough commercial activities are generated. Once the equipment is paid plus the 20% interest, ownership is transferred to the custodian, transferring the private property (co-ownership) into the nondominium form of property. The access rules change with this transfer, opening it more widely to network affiliates.

Contributions to infrastructure development and maintenance are treated as loans to the network, to be repaid back plus 20% interest from commercial activities. If the network never generates enough revenue from commercial activities the debt doesn’t need to be paid back. The role of the 20% interest is to incentivize contributions.

Sensorica has 4 years of recorded history of contributions. One of our biggest struggle has been to establish a culture of participation.

SENSORICA's initial development has been  fueled by grants, R&D tax credits, and personal loans. This is a centralized way of sourcing the development, in the sense that a few individuals collaborated to obtain funding, which was managed by the network.

Other models based on shared private property

We can find other arrangements in the world of peer production. Another common example is to form a co op around the operation. Physical assets are owned by the co op. Individuals are members of the co op, producers or users. This arrangement is less adapted for p2p, since membership is a contractual relation, and it makes the organization less fluid or dynamic.

[Add link to my critique of the coop]

Immaterial resources

The digital commons is the most widely used form of property for digital assets. According to their associated license, they come with the possibility to be forked and remixed. Some licenses make them viral, meaning that they propagate the “openness” to everything else that is built on top of them.

Digital assets are immaterial, therefore they require no maintenance, have very low storage and distribution costs.

The Creative Commons family of licences is very popular these days, and it is spreading in the realm of hardware too.

Some licenses impose restrictions on commercialization. Some limit the use to non-commercial, others limit the commercial rights to a community. Licenses determine the value capturing mechanisms and can be used strategically, in order to make development sustainable.

Environment of peer production

In the previous section we saw how the value system shapes the peer production system and the environment in which it takes place. People that engage in peer production are not all clear about these differences. As they build their processes and the environment, they borrow ideas from other initiatives without a profound understanding about their effects once implemented in their specific context. For example, I’ve seen groups trying to implement decision making processes that aren’t suitable for their type of activities. I also saw processes break down because they are not properly supported by appropriate tools and methodologies. Facing problems and not understanding where they come from leads the inexperienced to jump to the conclusion that peer production doesn’t work. We still don’t have all the answers, but the jury hasn’t come out yet.

Surveys show that peer production can take place in different types of environments.

The community

Most peer production systems are understood as communities, or tribes. This implies that they are built around a shared identity, culture and set of norms or governance. Individuals may have different levels of autonomy.

The p2pValue project has identified three types of communities[1]:

  1. Virtual communities that mainly have interaction through an online platform.
  2. Communities and projects that are supported by digital platforms but have a local focus and a face/to/face interaction.
  3. Localized communities whose main interaction is face-to-face.

[From the research produced by p2pValue] Still, some peculiarities can also be distinguished between the three categories such as looking at the subject addressed as their mission and target. The results of both Task 1.4 survey to participants and Task 1.1 Statistical analysis of cases, point this. In Task 1.4, regarding the outputs of the community, global online communities produce mainly commons that are shared with the whole society. Global communities with a local focus show however a major inclination to share the output among the community members (more than 50%) rather than with the whole society, even if the latter is also important (more than 30%). In the case of local communities, outputs are also shared as commons, both for the benefit of members and the rest of society, in similar terms. While according to Task 1.1 statistical analysis of cases, an analysis of the most salient words in the description of the mission of the case (as available in the case website), show the differences between digitally-based cases versus digitally-supported communities. Digitally-based cases use frequently words related to technology (the 5 most frequent terms: open, free, software, project and source). While digitally supported cases use words connected to “people” (5 more frequent terms: community, open, people, network and platform).

The community setting has an influence on how participants view themselves in relation with others, and in relation with their activities. The community paradigm implies that members strongly identify with the community’s mission, to which they are dedicated. This further implies a larger commitment or obligation.

The community also transfers value to its members: community building, objective accomplishment, and reputation, as identified by this p2pProject study.  

Some potential problems

Some of the problems encountered in peer production in community settings can be related to community issues. For example, the sense of alienation can result from bad facilitation, or lack of attention.

A reputation crisis can be triggered by a lack of mechanisms for capturing, surfacing and attributing one’s contributions. Ego can get in the way of progress. Social relations can stall projects. Individual insecurities can hinder communication and sharing. A lack of proper conflict resolution mechanisms can affect the vibe.

The feeling of non accomplishment can emerge from project failure, which can come from bad coordination, bad planning, lack of resources, bad timing, etc.

What make communities strong can also kill them. For example, open communities need to constantly broadcast their success stories, as well as their problems and their needs, in order to insure a steady inflow of resources. Open communities have higher turnovers than classical closed organizations, companies. Therefore, in order to insure the replacement of those who turn their attention elsewhere they need to be extroverted. In my experience I observed a slow, quasi imperceptible slide towards a tribe or club mentality, as social bonds strengthen over time and the common identity and culture take shape. This turns the community, which must stay open, inward, reducing its ability to regenerate, as some members slowly but inevitably leave. The newcomer starts to be viewed as a stranger and the community becomes less inclusive, less welcoming.

Another big problem is that the tribal mentality that settles in puts communities in a mode of competition. They see each others as distinct, as rivals, and members who go from one community to another are not highly regarded. Loyalty is elevated to the highest rank, and that is contrary to the open movement, which is about mixing of ideas and collaboration.

Synergy of autonomous entities

One way to create coherent outputs is to coordinate the assembly of independent modules. These modules may have a life outside of a specific peer production context. The actors behind them can have multiple interests, which can be relevant to a particular peer production context, but not specific or exclusive to it. If the same module is used in different value streams, it can lead to a multiplication of positive outcomes, and perhaps a multiplication of rewards. Actors that are attracted by a particular peer production environment seek to converge their individual projects with the requirements in this particular context in order to contribute to the common value stream, without necessarily being alienated. Guillaume Rouyer calls that a synergy of autonomies.

This synergy of autonomies implies that one cannot require participants to work one something specific - autonomy. Moreover, this also implies that everyone keeps the right to

use their contributions in other contexts - open licence. Conversely, a particular peer production process can integrate modules developed in other contexts, even without permission from their creators - remixing. We must therefore distinguish between a particular peer production context from the ecosystem which supports it, and acknowledge the fact that value creation can only be understood from the larger perspective of the ecosystem.

This is more suitable for immaterial production, software, content and others.

This setting imposes less constraints on the individual than the community setting. There is less need to strongly identify with the values or the mission of the peer production setting. The culture is less defined, and the rules are less strict.

In terms of values, the sense of accomplishment can be more important than community building. Reputation can still be built, if a sound system for capturing and evaluating contributions is implemented. This setting offers less support to the individual than a community.  

NOTE: This section was inspired by a communication of Guillaume Rouyer (Assemblee Virtuelle). It speaks to my previous distinctions between collaboration, cooperation and coordination. I defined collaboration as a joint action in which individual goals are strongly aligned (climbing a mountain together). Cooperation is a joint action to attain a common goal, but individual goals might not be well aligned (working in a corporation, the employee may want money for personal use and the employer may want to increase market share, increase profits, to gain power, etc.). Coordination is almost an accidental joint action. Individual goals are diverse and uncorrelated, but a common output can be created by assembling together interoperable individual outputs that may not even be initially intended for the common goal. Most peer production systems are a combination of the three.

Some potential problems

Integrating independent modules into larger systems requires interoperability, which in turn requires widely adopted standards. Standards suffer from coordination problems. The Arduino ecosystem is a great example of interoperability that emerged around the core microcontroller board, the heart of Arduino-based applications. Nevertheless, other competing platforms have been created in parallel. A certain level of diversity within the ecosystem is always needed, in order to allow the system to evolve. Too much diversity is detrimental.

One thousand individuals throwing one thousand bricks can make a pile of bricks, not a house. The process of integration itself needs high levels of coordination and must be supported by specific tools to insure the organization of the input into a coherent output. Quality can also be a concern. Wikipedia is a good working example.  

Since these systems don’t incorporate all the aspects of a community, permanency needs to be established by good content management systems, which misses tacit knowledge and wisdom.

These systems can be very dispersive if there is no effective mechanism to trace a vision and make it readily available to everyone. Focus can be lost.

This setting is more suitable for peer production systems that create immaterial things.

The network of oasis serving a nomad society

This vision was conceptualized by myself towards the end of 2014, in an effort to provide a solution to problems observed in the ecosystem of open initiatives in and around Montreal.

In this setting, we identify loci of peer production, visited temporarily and perhaps repetitively by individuals, actors, who are majoritarily nomads. A locus of peer production consists of a set of resources: physical space, equipment and tools, a virtual platform, a pool of competencies, a cluster of interests, an assemble of knowledge commons, a concentration of specific know how, ... These loci can also represent an opportunity for rewards or benefits. The metaphor used is the oasis. Since freelancing is on the rise, the individual developer or worker is seen as a nomad, hopping from an oasis to another. These nomads transport with them new knowledge and skills that they acquire in different places. Some of them visit a number of places regularly, depending on their own value creation processes, which also provides a sense of familiarity. Also, reputation can be built informally, if they visit the same places regularly. Standard reputation mechanisms and protocols can make reputation transportable across vast networks of places.

Some individuals are faithful to one location. They become the keepers of the oasis. Their main roles are maintenance, accommodation, guide, facilitation, and coordination. The keepers create use value. The nomads create exchange value.  

Capturing mechanisms must be put in place so that nomads can deposit value to allow the oasis to sustain itself. A universal value accounting system can be used to allocate some of the exchange value back to the loci of peer production. Peer production processes can span a network of locations.

Some potential problems

The biggest problem I foresee in this setting is to establish capturing mechanisms, to make sure that those who pass through and use the infrastructure actually contribute something to make it sustainable. This requires coordination between many spaces, based on a widely adopted value accounting system, in order to trace back use of resources from different spaces, going upstream value chains, which deploy across the network of spaces.


Problems

Participation

beyond immediate self-interest

The chart below shows the cumulated financial contributions by SENSORICA affiliates on a 4 years span. Tactus is a company created by SENSORICA affiliates to acquire funding to build the community and advance R&D projects. The other two major contributors are co-founders of the SENSORICA OVN. This picture shows that most of the financial burden to bootstrap this peer production network is carried in major part by two individuals. The SENSORICA mailing list, its principal communication channel, has 100 members in 2015.

Financial contributions within SENSORICA from February 2011 to February 2015

See more on SENSORICA’s Dashboard

Let’s now look at how SENSORICA affiliates contribute with their time for building the network. The graph below doesn’t include time spent on projects.

See more on SENSORICA’s Dashboard

We can compare this distribution with a distribution of time invested in the Mosquito project, the oldest SENSORICA project, see graph below. Time investment in a project is motivated by individual incentives. Time invested in building and maintaining the community benefits the collective, increases the potential of the entire network and, in doing so, it increases the potential for every affiliate. The data shows that SENSORICA affiliates generally don’t make the connection between sustaining the network and individual benefits.

See more on the Project’s value accounting page.

In August 2014 I initiated a transition to self-sustainability, trying to take SENSORICA off grants, to be entirely supported by its affiliates or by crowdfunding. The graph below shows how cash flows into SENSORICA OVN. loans are cash infusions from affiliates (myself only) to top off the monthly bills in order to maintain SENSORICA’s infrastructure. The presence of this form of contribution means that other affiliates don’t contribute enough cash to infrastructure maintenance to pay the bills. In SENSORICA’s case, one affiliate, the same one that has invested the most cash and time in network development, is lending money to keep the network afloat. SENSORICA will become sustainable when the loan form of contribution disappears, being replaced by Contributions to SENSORICA (a form of investment in the SENSORICA OVN, supposed to be paid back plus 20% interest from commercial activities) or by 5% fees (a tax extracted from commercial activities). This picture clearly shows that SENSORICA is still far from having established a culture of participation required for peer production, where affiliates feel “part of” and take initiative to develop the network. See more on Tibi’s monthly report.

From experience, at least in boothstrapping mode, the majority doesn’t seem to have the ability to see second or third order benefits that come as a result of giving, sharing or helping the community. The instinctive reaction is to go for the immediate reward, the first order action-reward, neglecting important activities that have the potential to amplify the ability to extract value from the system at a later time.

Different manifestations of the same problem observed in SENSORICA

All that has negative consequences on the entire system, which undermines the value of the network and its ability to provide as a community. This problem is quite complex, it has different layers from the tragedy of the commons to creating a very efficient and effective system for value creation and distribution. See the latest SENSORICA crisis called Sustaining the commons.

Causes of this problem

...

Possible solutions

The situation might change when the peer production network will generate enough commercial activity. Sensorica affiliates have implemented a tax on commercial activities to sustain the infrastructure. In 2015 this was set at 5%.

Most affiliates comply with this rule, but we have also observed signs of resistance and a tendency to find cleaver ways to esquivate it. This is no different from what we already observe in the current society at large scale, in the realm of government taxation for public goods and utilities.

Disorientation

Related to the illegibility of the dynamic network structure

Networks are dynamic and fluid. Newcomers need to understand who they can speak to, what roles they can take, where to contribute, etc. We need an adaptive signaling system, in the physical and the virtual space to guide newcomers. Other ideas?

When people join groups, everyone has its own approach.

Someone needed to provide orientations.

Mentoring.

Care relation - demotivation, losing interest when attention towards them is taken out. .

Information overload

p2p environments are suffering from information overload, because every node has access to all the information. How do we filter and reduce complexity, so that people  can make rational decisions without being overwhelmed with information.


[1] Open reference