For platform operators and clients

Are you a platform operator or client who wants to make sure your platform, or the tasks you post to someone else’s platform(s), meet high ethical standards? This page is for you.


The information is divided into five sections:


Advice for successful platform work

Calculating fair pay

Dispute resolution

Compliance with data protection regulation

Compliance with regulation for online intermediation services (EU) or online marketplaces (California)


If you are a platform operator or client (or potential client) and you would like more detailed guidance on ensuring that your processes are fair, please feel free to contact us at


Do you already offer the best working conditions? Then a membership in the Crowdsourcing Code of Conduct might also be an option for you!

Advice for successful platform work
Employment classification


The question of employment classification is one of the most controversial in platform work. Many platform workers are required to agree that they are self-employed, not employees of the platform or client. However, if clients or platform operators control when and how workers do work — even indirectly — this may be inappropriate. For example, if workers are given special incentives to work at certain times or do specific tasks, this could indicate an employer-employee relationship. There have been quite a few lawsuits about this. There is no clear answer, and the legal situation is changing in some jurisdictions as a result of court decisions or new legislation. Our general advice to platform operators and clients is that if workers are classified as self-employed, they should have the following freedoms to minimize legal risk to the platform and/or client:

  • Right and practical possibility to set own prices and negotiate these with clients
  • Right and practical possibility to negotiate contract terms with clients and/or platform (especially for example regarding usage of intellectual property)
  • Freedom from material and psychological incentives or disincentives to perform work (e.g., declining to accept offered work should not affect a worker’s reputation, access to future work, or payment for future work)
  • Freedom from non-competition or non-circumvention clauses in contracts
  • Freedom from non-competition or non-circumvention clauses in contracts
  • Freedom to choose own tools and working methods
  • Freedom to delegate work to others

A work arrangement in which the platform proposes that a worker is self employed but does not provide these freedoms normally accorded to self employed persons is legally risky for the platform and/or client.


Fair payment


Even when platform workers are properly classified as self-employed, platform work is real work and should therefore be paid fairly — at the very least, with minimum wage in the worker’s location.


If there is no minimum wage in the worker’s location, clients and/or platform operators should estimate a minimum “living wage” in the worker’s location and aim to pay at least this rate.


Even when work is paid per task rather than per hour, it is generally possible to estimate how long a task will take by piloting it with a small group of workers. For advice on setting fair pay, see “Calculating fair pay” below.


Clear and reasonable rules


The rules and timeline by which work will be evaluated and paid for should be made clear to workers in advance.


If the worker may be refused payment for work, the conditions under which this can occur — and the process by which the decision will be made — should be made clear in advance. Generally, workers should have the right to contest such decisions and to have their complaint reviewed by a qualified human employee of the platform who is empowered to revoke the decision if it is found to be mistaken.


Automated nonpayment decisions, including “voting” processes where several equally qualified workers do one task and “dissenters” are refused payment, should be avoided.


Nonpayment should always be explained, and ideally should not occur for any reason not listed in advance. Task instructions are an implicit contract between client and worker: if the worker follows instructions, they should be paid. Clear instructions — that address possible complications — are the client’s responsibility!


Test tasks and workflows


Tasks should be “piloted” with a small group of workers to ensure task instructions are clear and there are no technical problems that could lead to wasted time or unusable work. Workers should never pay the price of unclear instructions or technical problems not in their control.


Prompt, relevant, respectful communication


Customers and platform operators should respond to worker communications promptly, politely, and substantively. Clients should plan time to answer worker questions and make changes to instructions, technology, and evaluation processes based on worker feedback while their tasks are being worked on.


Form messages can be used, but overreliance on automated communication can lead to irrelevant answers to important questions — and, as a result, unusable work.

Transparent billing


Platform operators should provide detailed billing information to clients, including how much of the client’s payment is paid to workers and how much is kept by the platform; each worker’s location, the payments they received, and the time they spent on each task; and information about tasks for which workers were refused payment, and why.

Calculating fair pay — an example

An annotation task is to be completed by German platform workers, and should therefore pay at least German minimum wage, 9.19 € per hour. First, the client or platform generates a reasonable time estimate for how long the task takes an average qualified worker. To do this, they may “pilot” the task in-house to get a rough estimate — but in-house workers may be more qualified than platform workers, and therefore may do the task faster. So a second “pilot” with platform workers can be helpful. Depending on the capabilities of the crowdsourcing platform, the qualifications of these workers can be specified and they can be asked to complete the task without interruption.

Suppose the “pilot” is run with 10 in-house employees who do the task without interruption, and they take the following times (in minutes):


8 9 10.5 12 15 19 20 20 20 21


Suppose the online pilot with platform workers is done with 25 workers with the following requirements: (a) fluent speakers of the task language, (b) experienced with this kind of task, (c) they are told they should complete the task without interruption, and (d) the task monitors if they appear to have stopped working. Of the 25 workers, 4 are interrupted and their times are removed from the final data. The remaining 21 workers do the task with the following times:


11 12 18 19 20 21 22 23 24 24 25 25 25 25 26 26 26 27 29 33 34


We can set the initial hourly rate by ensuring that the fastest 75% of workers (in this case the fastest 16 workers) earn at least German minimum wage. In this case, the 16th fastest worker took 26 minutes to do the task, so the pay should be at least 26 / 60 * 9.19 € = 3.98 €.


Finally, completion times should be monitored as the task is run to ensure that the actual workers completing the tasks are not taking significantly longer for some unexpected reason. The person monitoring the tasks should be qualified to assess the situation and authorized to pay workers more or make changes to the task design or description “on the fly” as needed.


Especially the first few times a client runs a task, platform workers, like any other workers, must generally be managed. It may eventually be possible to achieve “frictionless” or “fire and forget” crowdsourcing — but this is not “the norm” the first time a task is run. Task descriptions, technology, and evaluation processes must first be tested and improved!

Dispute resolution and the Crowdsourcing Code of Conduct

Disputes arise in platform work. This is a given. Especially in situations where a client can decline payment for completed work or evaluate work performed, or where a platform operator reserves the right to suspend or close a worker’s account for a variety of reasons, it should be expected that disputes can and will arise.


The first and best step in dispute resolution is to expect that disputes will arise and to design legal terms, platform processes, and task instructions in such a way as to (a) minimize the likelihood of them arising and (b) set out clear rules about how they will be handled if they nonetheless arise.

From a regulatory point of view, the “gold standard” in dispute resolution in platform work may be set out by the EU Regulation 2019/1150, the so-called “Platform-to-Business” Regulation. (Text of the regulation is available here). This regulation sets out two levels of dispute resolution that can be invoked by a worker on a platform. First, the platform should operate an “internal complaint-handling system” to receive and handle complaints. Second, the platform should name external mediators with which they are willing to mediate disputes that cannot be resolved via the internal complaint-handling system. Even for platforms to which this regulation may not apply, this “two level” system is a useful model that can significantly reduce the risk of disputes escalating into costly and protracted legal battles. The “Ombuds Office,” described below, follows this model.

Dispute resolution lessons from Germany: the Crowdsourcing Code of Conduct and its Ombuds Office


The Crowdsourcing Code of Conduct was created in 2015 by the German software testing platform Testbirds. In 2016 it was improved with input from IG Metall, the trade union in the German manufacturing sectory. It has been signed by nine platforms and the German Crowdsourcing Association (Deutscher Crowdsourcing Verband).


The Code of Conduct commits crowdsourcing platforms to principles such as fair payment, respectful interaction, clear task descriptions, and a regulated work approval process.


Platforms that have signed the Code of Conduct participate in a dispute resolution system, called the Ombuds Office (“Ombudsstelle”).


If a worker has a dispute with a platform that has signed the Code of Conduct, they can file a complaint with the Code of Conduct Ombuds Office, managed by the trade union.


When a complaint is filed, the Secretariat contacts the responsible platform. If the dispute cannot be resolved bilaterally with the support of the Ombuds Office Secretariat, the case is reviewed by a 5-person panel composed of one crowd worker, one trade unionist, one platform employee, one Crowdsourcing Association representative, and a neutral chair.


The goals of this process are to clarify all the facts of the situation and to arrive at a solution that all involved parties can agree to. This is possible in many cases because many decisions disputed by platform workers arise either from technical errors on the part of the client or platform, misunderstandings on the part of the platform, or miscommunication.


The Ombuds Office process, in clarifying the facts, often leads to the platform operator voluntarily addressing the problem to the satisfaction of the worker.

Compliance with data protection regulation

Compliance with applicable data protection regulation is an essential part of successful and responsible platform work and should be considered by both platform operators and platform clients. Many people are intimidated by the apparent complexity of data protection regulation, but compliance does not need to be complex or burdensome.


While data protection legislation is evolving in many jurisdictions and its application to platform work is still being clarified, FairTube is in dialogue both with legal experts and with the platforms who have signed the Crowdsourcing Code of Conduct on this topic. Here we outline briefly our current understanding of some commonly arising data protection questions specific to platform work, especially those relating to compliance with GDPR (the General Data Protection Regulation), the EU’s main data protection legislation.

  • Platform workers and clients are “data subjects” under GDPR, and platform operators are “controllers.” Clients may also be “controllers” if they make decisions that affect workers, such as payment decisions and work evaluations.
  • If a worker is located in the EU, they have the rights provided by GDPR to data subjects (right to information, right of access, right to rectification, rights regarding automated decision making, etc.), and the platform and/or client has the corresponding responsibilities. This is the case even if the platform or client is located physically or legally outside the EU.
  • All data and decisions stored or taken within a labor platform with respect to a worker or submitted work are “personal data” as long as the worker they are associated with can be identified. This means that payment decisions, work evaluations, qualifications, and access decisions such as account suspension or closure are the worker’s “personal data.”
  • GDPR specifies that personal data must be processed “fairly and in a transparent manner” with respect to the data subject and must be accurate (Art. 5), and that data subjects have a right to receive a copy of data about them (Art. 15). In the context of labor platforms, this means that workers have a right to a copy of all data about them, including evaluations of their work and of them as a worker, qualifications, and payment decisions. Additionally, the processing that produces this data must be fair and transparent with respect to the worker.
  • GDPR provides a right to rectify inaccurate data (Art. 16). In the context of labor platforms, this means that workers likely have a right to contest work evaluations, payment decisions, qualifications, and other decisions such as account closure.


Platform clients concerned with platform workers’ rights should not only ask a platform, “Do you comply with GDPR?” but what the platform believes their responsibilities under GDPR to be — and then what steps the platform takes to meet them.


A few other questions from platform operators


How can I reduce the administrative burden of compliance with EU data protection law?


The short answer to this question is: You can automate, to the extent possible, the processes involved in the exercise of the rights provided by GDPR to data subjects.


To oversimplify a bit, there are two approaches to data protection compliance. The first approach is to put a boilerplate data protection statement on the website with an email address that workers and others can write to (or, at best, a web form that they can fill out) if they have inquiries related to data protection. Then a human being has to read and answer the emails individually. This is costly and time consuming.


The second approach is to systematically consider which data being created and processed on the platform are likely to be considered “personal data” under data protection law. Data such as evaluations of work, workers, and clients (i.e., reviews and ratings), qualifications, and classifications of a worker’s skills or of content are all very likely to be considered personal data. Once it is established that a certain category of data is personal data, the workflows that should be created in order to allow workers and clients and clients to exercise their rights with respect to this data can be inferred clearly from the GDPR. For example, a data subject has a right of access to all personal data about them. Therefore, logged in users should be able to see all of the information stored about them by the platform. Enabling this via a page within the platform (for example, in the user’s account settings) allows the platform operator to reduce the number of “free form” GDPR-related inquiries that must be answered “by hand.”


As a further example, GDPR establishes a right to have inaccurate or incomplete data corrected. Therefore if many inquiries involve correction of inaccurate data, the platform can reduce the number of these inquiries by adding a link “request correction/review” at the appropriate place in the platform interface. Manual review is still necessary, but the process can be streamlined by eliminating the handling of the unstructured data protection inquiry.


What about the California Consumer Privacy Act? How does it differ from EU data protection law?


While the interpretation of this relatively new California legislation is still being clarified, our understanding is that the scope of what is considered “personal information” under California privacy law is similar to what is considered “personal data” under EU data protection law, at least within labor platforms. Therefore data relating to, for example, worker or customer performance or activities on a labor platform, such as reviews of workers or customers, are probably “personal information” under California law. Our current understanding is that under California law, users have a right of access to personal information being processed about them, but not necessarily a right to have inaccurate personal information corrected.


Data protection compliance is too expensive and annoying. Can I just ban workers from the EU and California from working on my platform and call the problem solved?


You could do that. However, our experience is that data protection compliance does not have to be as expensive and annoying as many people fear. As noted above, the expense and effort can be reduced significantly through good design and automation. Additionally, consider that many other jurisdictions, including US states other than California and countries all over the world, are in the progress of expanding or updating their data protection legislation, and the European and Californian examples have “set the bar” for these efforts. Therefore it’s likely that if you adopt a strategy of avoiding compliance and limiting your worker pool to jurisdictions with more “relaxed” data protection standards, your worker pool may shrink over time.

Compliance for online intermediary services (EU) or online marketplaces (California).

 Online services that function as marketplaces or other types of "intermediaries" are gradually being regulated. In the European Union, for example, the contents of the Digital Services Act are being negotiated. In California, AB 1790, which regulates the relationship between "online marketplaces" and "marketplace sellers," went into effect in January 2020. Broadly, these laws require platforms to set clear legal terms explaining when the platform may withhold funds from sellers (or "business users"). They must also specify the conditions under which the platform may suspend or close a seller's account, whether it must issue a written statement to that effect, and specify procedures for appealing such decisions.



Compliance for online intermediary services (EU) or online marketplaces (California).


Online services that function as marketplaces or other types of "intermediaries" are gradually being regulated. In the European Union, for example, the contents of the Digital Services Act are being negotiated. In California, AB 1790, which regulates the relationship between "online marketplaces" and "marketplace sellers," went into effect in January 2020. Broadly, these laws require platforms to set clear legal terms explaining when the platform may withhold funds from sellers (or "business users"). They must also specify the conditions under which the platform may suspend or close a seller's account, whether it must issue a written statement to that effect, and specify procedures for appealing such decisions.