Photo by Thought Catalog on Unsplash

Five Ways in Which Platform Business Models Influence Workers’ Well-Being

A summary of the ILO’s comprehensive report on the future of work

The International Labour Office (ILO) published a comprehensive Future of Work Research Paper offering a valuable overview of the platforms’ business model design choices and their influence on workers’ well-being. I summarized some of the main points of the paper for Exponential View, a weekly newsletter you should subscribe to if you’re interested in the impact of exponential technologies on the near future of society.

Two management techniques are part of the framework that determines whether a worker is exploited or empowered: lean startup methodology and management by metrics:

• Lean startup methodology: frequent changes to the platform’s design choices and policies may adversely impact consumers or producers who spend considerable amounts of time or money to participate on the platform only to realize that the workings of the platform have changed.
• Management by metrics: in addition to managing internal product development with metrics, platforms can also manage their ecosystem through metrics. Ride-hailing platforms, for example, use metrics to manage their drivers who are obliged to achieve or avoid a given threshold as a pre-condition of continued participation on the platform.

Framework for exploitation

The presence of these five characteristics indicates platforms propensity to contribute to worker exploitation:

1. Removal of free agency

• Information asymmetry between platforms and workers limits free agency for workers by preventing them from accessing information that would help them choose profitable interactions on the platform.
• If drivers were operating as entrepreneurs, they would tend to choose the rides which delivered maximum financial outcomes at minimum cost. To empower drivers as entrepreneurs, the platform would have to provide them with all the information needed to make a decision about accepting a ride. However, by forcing the driver to accept a journey with scant information, the platform reduces free agency for the worker
• Platforms may change the presentation of information after having established trust with platform participants.
• Power imbalances on platforms also occur when platforms rely on user input to develop machine learning models. These models are used to minimize market failure in future interactions by influencing how tasks are completed. If the driver training the algorithm is given a sub-optimal route without being informed about it, the platform has exploited information asymmetries to its own benefit.
• Workers who are managed by these algorithms often have a limited understanding of how they function.

2. Reduced bargaining power and rights

• Workers are likely to have less bargaining power when the potential worker base is large and when workers are more easily substituted. Hence, in the case of standardized or low-skilled work, the power balance shifts significantly away from workers. Meanwhile, platforms that mediate high-skilled work cannot afford high turnover among their workers and are therefore likely to create policies that are more worker-friendly.
• Some platform policies endeavour to create a superior experience for consumers, passing on the costs to workers. For example, if a passenger forgets something during a ride, Uber does not pay drivers for the time and effort required to return a passenger’s lost property.
• Some platforms encourage workers to compete among themselves such that only the winner is paid (99designs).
• Platforms that exert greater control over the terms of exchange determine how much workers get paid, and require confirmation of adherence to certain service standards, as well as surveil the tasks performed. Such controls deprive the worker of freedom of choice to participate and erode free agency, opening the door to exploitation.
• Lyft even tells drivers what to say: it instructs them on how to greet passengers in line with the platform’s brand.

3. Making workers subservient to the platform

• An Uber driver offered payment guarantees based on high acceptance rates is required to maintain those high acceptance rates for UberPool as well.
• Deliveroo platform requires workers to respond to new orders within 30 seconds, the only electronic option available being: “Accept delivery”. The delivery address is not revealed until the food is collected from the restaurant, at which point the only way to cancel the order is by contacting the driver support line directly. At that point, any refusal to deliver is recorded, to the detriment of the worker’s reputation.
• As the workers lack easy access to consolidated action through unions or legal representation, the balance of power favours the platform, which can introduce policy changes with scant risk of a legal challenge
• Several labour platforms, especially those tracking service delivery, gather large quantities of data about individual worker’s behaviours, both during the active working time and during more passive periods of participation
• Limiting collective action.

4. Making workers dependent on the platform

• In order to increase multihoming costs for workers, platforms prevent workers from transferring their reputation to other platforms. If the worker were to move to a new platform, they would have to invest time, effort, and money in building their reputation from scratch. In this manner, platforms effectively control a worker’s career, not just the allocation of their next job. Lack of reputation portability may also reduce a worker’s ability to find non-platform work.

5. A platform doesn’t allocate risks and rewards fairly

• Ride-sharing platforms encourage workers to be available on the platform, to ensure ride availability for consumers, without any reciprocal guarantee of work.
• Platforms tend to pass on to workers the cost of any liability for unforeseen circumstances.
• With its wealth of data, the platform often intervenes as an arbiter in case of conflict between the customer and the worker. These arbitrations may conveniently minimize the platform’s liability while passing on the cost of resolving a conflict to the worker.
• Worker ratings are one factor that can determine workers’ search result placement, and hence their visibility to consumers. Creates a feedback loop.

The report proposes a set of solutions

• Several jurisdictions have taken an extreme approach of entirely banning platforms which do not comply with existing regulations. This is unlikely to prove to be a nuanced or sustainable solution;
• Some scholars argue that traditional regulation, when applied to platforms, will lead to over-regulation, thereby curtailing all benefits that labour platforms create. Some proponents of eliminating regulation go so far as to suggest that, because the interests of the platform are intrinsically aligned with those of the workers, platforms will naturally be motivated to invest in worker protection;
• The central issue for regulators seeking to empower workers is platform control, not worker status;
• Changing worker status from independent contractor to employee could well improve the social benefits and insurance coverage;
• To combat worker exploitation by platforms, the goal of regulation should be the enablement of worker agency and a reduction of platform control;
• Regulation of labour platforms must seek to restore worker power that has been eroded on labour platforms, while employing new models for regulation that apply to the platform economy;
• An expandable and effective regulatory framework for platforms must be centred around the regulation of data;
• A regulatory framework for platforms would best be structured as multi-sided coordination between three stakeholders: the platform, the regulator and the workers (and their representatives). As a result, should enable worker empowerment through data access and usage rights and enable:
  → Unionization
  → Worker-centered information across different platforms
  → Metrics-based bargaining
  → Reputation portability and dual reputation.

Full paper available here.

If you liked this, you’ll enjoy the weekly newsletter Exponential View. Get on the list here → http://www.exvw.co