Rebalancing interests and power structures on crowdworking platforms

Disclaimer: This study was completed with the support of the German service sector union ver.di. We would like to thank the participating platforms and their communities for the opportunity to conduct a survey. We would also like to thank the jovoto platform for the implementation of an idea contest. Thanks also go to Prof. Dr. Christian Fieseler, BI Norwegian Business School and Prof. Dr. Wrona, TU Hamburg-Harburg, for their comments and reviews. Any errors remain the responsibility of the authors.

Introduction and objectives

Working on virtual platforms, whose function is to match up the skills of sellers with assignments set by buyers, is a new phenomenon. In uncertain economic times in which long-term planning is not possible, tasks, resources and skills must be marshalled, ad hoc and at short notice. The resulting “pull economy” differs from the traditional “push economy”, which allocated these resources on the basis of supposedly accurate plans and sought to develop them (Hagel et al., 2010, pp. 31 ff.; Bauwens et al., 2012, pp. 29 ff.). Now, traditional companies increasingly use platforms to quickly access needed skills from external sources, boosting their capacities (Al-Ani, 2013, pp. 89 ff.). This “cooperation” between the platform and the hierarchy is needed to solve a pressing dilemma for the firm: on the one hand, companies are subject to cost pressures due to globalisation and hyper-competition; on the other hand, there is a lack of innovative abilities in the traditional lean hierarchies: cost cutting has limited their capacity to create ideas and to conduct experiments (Hamel, 2007, pp. 45 ff.). In this situation, it is precisely the innovative, cost-conscious and almost inexhaustible power of the crowd, accessed via platforms without negatively impacting the company’s cost structure, that seems to have become a new core competency of the firm (explore & exploit). Not surprisingly, a variety of pull platforms, which integrate the crowd’s labour and innovative power into the company’s value chain, have emerged (Hagel et al., 2010, pp. 61 ff.).

While a number of reflections on the mechanics and functioning of these commercial platforms have emerged in recent times, we lack deeper insights into the new employment relations that have appeared as a result.1 The present study examines what determining factors crowdworkers consider relevant in their work and, above all, what support measures they anticipate to be useful in improving their working relationship with the platform.

Prelude: the rise of the P2P platform

In the software industry of the 1990s, salaried programmers came together in their spare time to work on interesting, unpaid projects on custom-built platforms. These employees switched roles after the working day finished and mutated into “free producers”, into peers, who, together with other free producers, could do more, not just for themselves, but above all with other people (Benkler, 2006, p. 9). New principles of cooperation emerged: the intrinsically motivated peers worked in a self-directed manner and autonomously sought out their challenges.2 On the basis of an often given “core”, thousands of programmers sought and found opportunities to participate in such open source projects all around the world. The results of these collaborations were made available free of charge as commons, and everyone was invited to participate in improving them (Bauwens, 2005). These new peer-to-peer (P2P) collaboration platforms had extremely flat organisational structures.3 This does not mean that there were no hierarchies. However, within the P2P environment, these levels are used to secure the mostly open access to the platform and to protect the contributors; they are not intended to enforce exclusivity: ‘An important clarification is that when we say that peer to peer systems have no hierarchy or are not centralized, we do not necessarily mean the complete absence of such characteristics. But in a P2P system, the use of hierarchy and centralization, serve the goal of participation and many-to-many communication, and are not used to prohibit or dominate it.’ (ibid). The working principle of these platforms can be defined as a meritocracy: peers are mainly assessed by other peers based on their contributions. This feedback alone allows them to rise to a new role with more responsibility (ibid). If an individual’s contribution diminishes, the role is relinquished. A paradoxical principle arises: a highly individualised form of organisation, in which individuality plays a major role while the individual plays a small part.

Platforms as partners and competitors of the firm

The obvious success of P2P platforms (Wikipedia, Linux, etc.) also led to a growing interest among companies in this form of organisation and especially in the capacities of peers, who usually worked free of charge.4 To create a needed interface between companies and peers, commercial platforms have emerged, which have “interpreted” or rather re-interpreted the basic principles of P2P platforms (self-identification, self-management, meritocracy) in a commercial context. Covering a wide array of needs, these platforms differ quite significantly in their functioning and thus also in the kind of crowdwork that takes place on them (Al-Ani, 2015, pp. 11 ff.; Al-Ani and Stumpp, 2015, pp. 24 ff.):

• Efficiency platforms: these platforms assign small-scale work packages from contracting companies to interested crowdworkers. The work packages range from less complex tasks, such as those offered through the well-known platform Mechanical Turk, to the evaluation of seminar papers, and even complex medical reports. 5

• Innovation platforms: on these platforms, companies advertise issues to be solved using the framework and logic of “competitions”; the “winners” are rewarded with a bonus. This category includes scientific-problem-solving platforms like InnoCentive, creative platforms like jovoto and software development platforms like TopCoder.

• Agency platforms: on these types of platforms, mediated resources – and not work outcomes – are in the foreground. Agency platforms source particular skills or rather employees for contracting companies. The employees sourced in this manner subsequently act as freelancers rather than crowdworkers within the traditional client enterprise: they are not managed via the platform’s algorithm, but directly via the firm’s hierarchy. Examples of this include the platforms Elance, Twago, oDesk or Jomondo.The path from these agency platforms to “real” crowdsourcing platforms is not a very long one. It is obvious that an extension of the business model sooner or later may involve taking on the task of managing freelancers for the client.

In contrast to the original P2P platforms, the “commercialised” platforms led to significant changes for the participating peers, who mutated into crowdworkers who:

• Are paid for their activities, at least for a bid (winning a tender) or if a predefined goal is achieved (e.g. finding a software error).

• No longer collaborate with others in a self-managed manner, but are instead individually allocated, directed and remunerated based on the procedures or algorithms of the commercial platform: self-management is repressed and is only possible within narrowly outlined steps, which leads to a power asymmetry at the crowdworkers’ expense. The algorithm determines what work will be accepted and carried out under what conditions and thus restricts contingency significantly: “Working conditions in digital workplaces appear to be characterized by a smaller tolerance for failure and mistakes. More strikingly, platforms that mediate digital work severely limit the scope and potential outcomes of negotiations. Workers can accept tasks and supply results, and requesters can define and allocate tasks and accept or reject results. These basic platform settings also limit the negotiation opportunities of requesters, but workers systematically find themselves in a weaker bargaining position because the power to sanction perceived misbehaviour rests squarely with the requesters.” (Bucher, Fieseler and Hoffmann, 2015, p. 14 f.).

• Interact with a platform company which intermediates between the client and the crowdworker.

This said, platforms do not necessarily have to confine themselves to being sourcing partners for companies. Platforms are increasingly becoming competitors to private sector companies. Here, the combination of the peers’ work capacity with particular assets owned by the independent producers (cars, rooms, etc.), results in a “new” product or service that is provided to the market through the platform (Manjoo, 2015). This can create unique and disruptive situations on the market, since from one day to the next, both the peers’ working capacities and their thus-far commercially unused assets are brought onto the market (Uber, Airbnb, etc.) via platforms. These compete with existing providers in the transport, hotel and media industries quite successfully, since these platforms are often able to draw on the high commitment of the crowdworker, who has mutated into a micro-entrepreneur, as well as the lower overhead costs, high flexibility and innovative customer interfaces (ibid.). Clearly, this “flexibility” is also due to the policy of these platforms, which often do not treat their producers as employees in the traditional sense. In extreme cases the commercial platform accepts no liability for the actual market transaction, but charges a commission of between five and 20 per cent for initiating and mediating the agreement between the producer and the client.6 It is therefore not surprising, that despite the ongoing “monetisation” of the original P2P idea, commercial platforms are often still regarded as a neutral party positioned between clients and crowdworkers.7

Lessons learned so far: history remains

Due to the widespread impact of new forms of working practice and organisation, there has been considerable interest from institutionalised research, and in particular from a vanguard of popular science and business management writers.8 On top of that, the P2P movement has institutionalised itself as a complex of interests on its own platforms, which encourage and support the development of scientific content in the spirit of P2P (P2P Foundation, n.d.a.). In recent times, the growing strength of some platforms, which have emerged as competitors to companies, has added an important topic to the research agenda.9

The findings of these studies can be summarised as follows: on P2P platforms such as Wikipedia and open software platforms such as Linux, intrinsic motivation, self-management and self-identification with tasks are dominant factors. Workers who have “escaped” from hierarchical organisations grasp the opportunity to finally do the things they want and thereby enjoy the appreciation of other peers.10 Although P2P platforms are aligned with the principles of self-management and self-identification, they are not free of hierarchies, and do indeed have procedures, algorithms and management principles. However, the function of such control loops differs from the function of similar systems in hierarchical organisations, because the mechanisms are not there to establish exclusivity, but rather intend to achieve meritocracy, openness and cooperation: “[…] P2P is a democratic process of full inclusion based on the idea of equipotency. It believes that expertise cannot be located beforehand, and thus general and open participation is the rule. But selection immediately sets in as well, since the equipotency is immediately verified by work in the project. Thus there is a selection before the project, and a hierarchy of networks is created, where everyone finds his place according to demonstrated potential. Within the project, a hierarchy is also immediately created depending on expertise, engagement and the capacity to generate trust. […] Peer to Peer is not anti-hierarchy or even anti-authority, but it is against fixed hierarchies and ‘authoritarianism’ […].” (Bauwens, 2005). Commercial platforms have adopted the structural elements of P2P systems and have begun marketing their peers-turned-producers’ services and focusing even more closely on customers’ requirements. Despite this restriction of self-determination, even on commercial platforms like Mechanical Turk and Uber, elements of the P2P paradigms still seem to be influencing the motivations of the crowdworkers: fun and learning experiences, for instance, are still of considerable importance and often outweigh commercial considerations.11 Yet, this transformation of the commons based P2P production into a commercial enterprise begins to provoke reactions of the producers. There are some initial indicators that the dissatisfaction with the existing power asymmetries on commercial platforms will increase with the length of time spent on the platform.12

Design of the study

This study is based on an explorative two-step process, which is composed of expert interviews in the form of a workshop and a quantitative online survey.

Opening the field: the expert workshop

At the beginning of the research project in June 2015, an expert workshop was conducted for platform representatives, representatives of trade unions and scholars to address the phenomenon of crowdworking. First, the typologies of individual commercial platforms were analysed, and scenarios were outlined for trade union support of crowdworkers based on international experience (Al-Ani and Stumpp, 2015). Based on this discussion, the strategy for the present study was developed. We selected two platforms – an efficiency and an innovation platform, and subsequently two target groups of crowdworkers – creatives and IT developers and testers – to cover the greatest possible range of different crowdworkers and platforms. These target groups were to be asked about their motivations and their requests for improved working conditions on the respective platforms. The intention was also to examine whether there are concrete demands for unions to optimise these conditions.

Quantitative online survey

Participants in the online survey were crowdworkers from the platform jovoto.13 and an anonymised efficiency crowdworking platform. jovoto is an innovation platform for creative people, particularly those who design objects and create communications strategies. The jovoto community has an international scope. The sample in this case is made up of a total of 20 nationalities. By contrast, crowdworkers on the efficiency crowdworking platform test software for applications, websites, desktop applications, etc.. Here, one could assume that community members have a computer science background. The community on the efficiency platform is internationally oriented, but for practical reasons only German-speaking users were surveyed. Participants were approached in the course of a notification on jovoto, which was sent to the entire group. On the IT platform, the respondents were contacted via the community managers. The survey was conducted in the period from 15 September 2015 to 30 October 2015. The sample size for the efficiency platform was n = 93 and for jovoto, it was n = 72 crowdworkers. The participants in the sample from the efficiency platform are henceforth referred to as the IT crowd, while the participants from jovoto are referred to as the creative crowd. The questionnaire used in this sample consisted of both open-ended and also Likert-type scale questions. We asked participants to fill out questions about their sociodemographic information, current working situation, their motivations and also their expectations about support mechanisms for crowdworking.

Fun and learning effects, the main drivers of crowdworking: data analysis and results

Sociodemographics and time spent crowdworking

Crowdworkers in our two samples are predominantly male, young and well-educated. The proportion of male crowdworkers in the sample is 67.7 per cent for the IT crowd and 51.4 per cent for the creative crowd. Most of the crowdworkers are up to 34 years of age. The 50+ age group, by contrast, is a minority. The differences between the two groups, the IT crowd and creative crowd, are relatively low in the age groups represented.

For the educational levels, too, the results are clear. On both platforms, the crowdworkers are well-educated and have either completed a university degree or have at least a higher secondary-level certificate that qualifies them to enter university.

Figure 1: Employment status of crowdworkers (in per cent)

What is particularly notable for the two groups is that a very high share of these individuals work on standard contracts elsewhere – 45 per cent of IT crowdworkers report this status and about 35 per cent of the creatives (see Figure 1). In addition to explaining the higher payment for IT services, this may also explain the better income situation of IT crowdworkers: while only 18 per cent of the creative crowd earn a net household income of more than 3.000 euros a month, at least 31.2 percent of the IT crowd have earnings above this level.

There are other differences that are apparent in the monthly investment of time. The creative crowd invests significantly less time in hours per month than the IT crowd (see Figure 2). Meanwhile, a large percentage of users within the IT crowd invests eleven to 40 hours per month. Again, this is probably also an indicator that platform work is becoming or has already become part of the work- and lifestyle in the IT sector.

Figure 2: Time spent crowdworking per month (in hours)

Motivations of crowdworkers

Surprisingly, however, the socioeconomic differences have no fundamental impact on the respondents’ motivations. On both platforms the motivating factor “fun” is ranked most important, followed by “additional income” (see Figure 3). On the one hand, this result may be an echo of the P2P traditions, but on the other hand, it may be explained by the feedback that only a minority of interviewed persons can earn a living by working on these platforms (see Figure 3).

Figure 3: Motivations of the crowdworkers (in percent): the scaling of the response options ranges from “I agree completely” to “I strongly disagree” (1–4). The chart shows the cumulative responses “I agree completely” and “I somewhat agree”.

In addition to the almost identical motivation present in both groups – “to gain appreciation” – in both platforms, learning new skills was also a very prominent factor. Creative crowdworkers seem to use these platforms to an even greater degree to build up their reputation and to instrumentalise or practice entrepreneurial knowledge. Obviously, the opportunity to build up a market reputation on the platform is more important for this group than for the IT crowd.14 Here, a number of participants mentioned access to interesting customers, brands and projects. Clearly, the advantage of innovation platforms is that crowdworkers have the opportunity to obtain access to projects or clients and brands that they would normally not encounter in their regular employment relationship. The relevance of platform work for hedonistic motivations and also for learning new skills enabled by meritocracy is also reflected here in the comments of the respondents (see Table 1). In addition, the higher temporal and geographical flexibility of this work is evidently a further advantage.15

How can crowdworking conditions be improved?

When asked how other actors can support the work done by crowdworkers, the respondents, not surprisingly, largely assessed all the listed possibilities as reasonable – albeit to a greater or lesser degree. This generally positive assessment may have something to do with the participants’ relatively brief experience, which limits their ability to assess the costs and benefits of these supporting strategies.16 In addition to providing ways of evaluating platforms, this especially involved developing their own community, which, as a kind of shadow organisation, would reintroduce the connections between crowdworkers that were suppressed by platforms.17 Just establishing an independent rating system to track the reputation of crowdworkers, found slightly less support.

Figure 4: Strategies for supporting crowdwork (in per cent)

What is initially surprising in this analysis is the finding that the respondents – and here the IT crowd is a bit more supportive – strongly favoured more transparency and a certification of algorithms of platforms. However, when one considers the central role of algorithms in the control of crowdworkers, this requirement is more than understandable. And the fact that the IT crowd viewed this issue as the most urgent comes as no surprise because of the familiarity of this group with IT functionality.

An open response field gave the respondents an opportunity to express their expectations towards supporting mechanisms for crowdworkers. It is striking that the respondents have a critical attitude towards rating mechanisms due to the risk of fake reviews. It is also obvious that there is a demand for a system to determine a fair wage for the crowdwork done. Besides considering the technical requirements, there is also a demand for work appreciation and a kind of lobbying work for crowdworkers (see Table 2).

Is there a potential role for traditional trade unions?

Trade unions could act as partners of crowdworkers and help them fulfil some of the mentioned requirements. When crowdworkers think about trade union support, they often think of tasks like providing advice, which is already offered today.18 The fact that there is a demand for a neutral or conflict-mediating body indicates that the above-described dilemmas of corporate platforms and their power asymmetries have become more apparent. This situation may act as an incentive for trade unions to extend their traditional range of services and skills; in order to take on these roles, unions have to understand and deconstruct the mechanics and algorithms of the platform in the same manner as they evaluate firms’ business processes and IT solutions today. These new skills, in turn, could then be used to provide certification for platforms by the union, an idea which is supported by nearly one-third of respondents (see Figure 5).

Figure 5: Expectations of a union (in reply frequencies)

However, there is another traditional role of trade unions that is apparently not demanded by crowdworkers in our sample: that of advocacy. The lowest expectations related to support for worker self-organisation and the union organisation of groups of crowdworkers. This may have something to do with the “de-institutionalised” self-organisational abilities of crowdworkers, who can act via shadow organisations or rather, shadow platforms that are very similar to the original P2P (self-)organisations (Bauwens, 2005).

At this point, it seems justified to speculate about the abilities of crowdworkers to build bargaining power effectively. Some initial experiences, for instance with the Amazon Mechanical Turk platform, seem to show that self-organisation is a rather complex undertaking: self-organisation manoeuvres between the risk of being too passive and not offering enough stimulation to members to join in (stall) or having to deal with too many different interests and collapsing under the weight of contradictions (friction).19 Still, as crowdworkers clearly acknowledge their position on the platform as not being equal to the platform owners and customers, we can expect self-organisation experiences and bargaining processes to increase over time.20 If currently specialised organisations like unions are finally able to connect with these self-organised bodies, as happened in the Uber California cas.21, this will, we suspect, only be possible after intensive rearrangements and change processes in the political sphere that surrounds the digital economy, of which unions are a part of.22

Conclusion and discussion

From a crowdworker’s perspective, platforms today operate as a kind of learning laboratory: people upgrade their reputations and knowledge, which they can then use and monetise in self-employment, but also in standard employment. Work on platforms thus becomes part of an individualised learning path for people who are continuously learning “on the job”.23 Of course, pay plays an important role, but not in the sense of earning enough to live but rather as additional income to secure and support the crowdworkers’ existing living situations. To that extent, a discussion that overemphasised the precariat would narrow down the issue too much, because the motivations for and the effects of crowdworking today are in very different areas, and crowdworking in fact compensates for certain deficits in traditional working life (rigid, inflexible learning opportunities, economic inequality, hierarchical organisations, division of labour, restrictive development opportunities, seniority and elitism, etc.). However, on this point, it is interesting to note that the criticism of crowdworking has not led to a related discussion of the shortcomings of traditional companies and educational paths. It is these very deficits that make crowdworking attractive, despite its limitations and asymmetries.

The novelty of crowdworking suggests that the crowdworkers’ demands concerning the fairness of the platforms, as well as the search for and the development of alternatives will increase in the next few years. Over time, the ability of the crowd to analyse asymmetries and to develop counter-strategies will also expand. It should also not be forgotten that, given its propensity for self-organisation, the crowd always has the inherent potential to develop its own business models, if the dissatisfaction with what its members encounter is too great. For those working in the virtual world, the option to exit always means building something of one’s own, perhaps even with the help of trade unions.24 Alternatives arise when one decides to flee: “To resist is to create!.25 Viewed in this manner, we suspect that commercial platforms do not necessarily have to participate in a rat race, but instead may attempt to increase their attractiveness by offering crowdworkers additional functions (social benefits, training, etc.).

The transparency of platforms requires the development of analytical and evaluation skills as well as new rules for the participants of the digital economy. The fairness of the distribution and of the procedures requires the ability to decode and evaluate algorithms. As recent examples from the automotive industry show, however, opening the relevant programming is a delicate and difficult matter, because abuse is possible and competitive advantage may be lost.26 In this regard, trade unions could be active as neutral and independent bodies: they need to develop skills in decrypting algorithms so that they can provide competent advice and mediation and thus act as certifiers of platform procedures. This would counter the danger of competitive disadvantages for the opening of platforms and would signal to crowdworkers that the algorithm is based on fair criteria.27

Knowledge about the effects of algorithms, however, is of little use, if crowdworkers cannot increase their bargaining power to assert their demands to the platform owners. For crowdworkers, the community is obviously an important element. It promises protection, or rather help and assistance, or at least the feeling of being part of a greater whole, which is composed of different individuals who share similar life situations (Bollier, 2014). Furthermore, considering the difficulties of imposing regulations on entities such as platforms – which use a workforce from various regions of the world and whose “location” can easily move from one place to the other – the importance of self-organisation in the virtual space and at physical locations becomes a key factor.28 At the same time, crowdworkers do not seem to consider the organisational capacity of unions, which may have something to do with the novelty of the relationship between crowdworkers and the unions, but also with this group’s inherent capacity for self-organisation. Initial experiences with platforms like Amazon Mechanical Turk and Uber, however, show that self-organising crowdworkers for political conflict is a rather complex undertaking. From this perspective, trade unions have the opportunity or challenge to support these self-organisational capacities with appropriate services. For the moment, it seems unlikely that crowdworkers will use unions as the dominant site of self-organisation. What is more realistic is a scenario in which trade unions are a part of an alliance of groups that represent worker interests and enter into cooperations with specific crowdworker organisations and ad hoc movements that will, in all likelihood, develop once this form of employment will have expanded.29 Understanding the dynamics of such a scenario would require further research.