Positive Platforms Design

The Eight Principles

Marina Gorbis and Devin Fidler:

"Based on this research, we’ve begun to identify some principles that could guide Positive Platforms Design:

1. Earnings maximization.

We should design to optimize opportunities for those working on the platform to increase their income streams. Connections between design choices and earnings are not yet fully understood. Research has suggested, for instance, that for some types of work people do not do as well financially when the platforms set minimum wages as compared to when workers can set their own wages. Research by Arun Sundararajan and others, as well as our own observations, suggest that platforms that allow workers to organize their own small enterprises, rather than those in which workers merely serve the needs of the platform, tend to generate higher levels of incomes for platform workers. AirBnB’s design and photography services that help property owners make their listings more attractive is one example of a platform helping people maximize earnings, resulting in higher profits for AirBnB as well as for individual property owners.

2. Stability and predictability.

We are in a phase of prototyping and experimentation in platform design, a practice that is key to Silicon Valley’s style of innovation. But in the case of platforms this innovation has a direct impact on people’s livelihoods. Imagine if every month you came to work and your salary were different; this is exactly what many on-demand workers experience today. Participants in our study, for instance, described shifting pay structures with only a few days’ or no prior notice. While experimentation may be the necessary phase in platforms evolution, it is important to think about the human costs of such experimentation and build mechanisms for minimizing or compensating workers for ensuing volatility.

3. Transparency.

We need transparency at two levels: at the level of the platform algorithm itself (so that workers understand how to increase their earnings) and at the level of archived data (so that those working on platforms understand how their personal data is being used). Many people we interviewed reported how difficult it was to figure out how to maximize their earnings on platforms due to the general opaqueness of the algorithms powering them. Workers may consequently have trouble calculating out their actual hourly wages or whether it is worthwhile for them to take on certain tasks.

4. Portability of products and reputations.

People working on platforms should be able to own the products of their work and their reputation histories, and carry them from platform to platform. Platform reputations are often directly tied to earnings as well as opportunities for various types of work. This is how one research participant describes the experience of “losing” a reputation as well as the accompanying confusion when a platform was acquired by another company: “All of my portfolio links are broken now, and I don’t think people can find me anymore.”

5. Upskilling.

While traditional career ladders may not be relevant in the world of on-demand work, people still look for opportunities to increase their levels of skill and expertise. The best platforms already show those who work on them pathways for learning a particular skill and connect people to resources for advancement. Upwork, for example, not only provides forums for people to mentor and provide support for each other but also links them to free and paid courses where they can acquire desired skills.

6. Social Connectedness.

Many of today’s workers are creating communities outside of the platforms where they work to exchange tips and connect with each other. Reddit, Facebook, Google Groups, and other social media sites are becoming de facto places for this. As one person we interviewed said, “I think it’s important for me to build a relationship with the people that I work with.” Mechanical Turk workers have come together on a series of forums to not only create a sense of cohesion but also to advocate for their rights. Platform designers can make this easier by enabling and fostering such communities.

7. Bias Elimination.

Networks are at the core of what makes platforms work. Unfortunately, networks can be exclusionary (due to clustering of people with like minds and backgrounds) and polarizing (because more connected nodes tend to draw even more connections). In formal organizations, decades of labor struggles and court rulings established some basic rules and principles for non-discriminatory hiring, promotion, and so forth. We need to evolve such rules and principles in platform environments. In fact, because of vast amounts of data platforms accumulate, they may be in a good position to integrate mechanisms for surfacing bias as well as eliminating it. Models for this come from some recent startups such as Degreed which can match people to job opportunities independent of their degrees or demographic characteristics, or Unitive, which has developed software that helps spot unconscious bias in job descriptions language.

8. Feedback mechanisms.

It is hard to negotiate with algorithms, and most platforms do not have HR departments for handling everyday issues workers encounter, from late payment to unfair reviews. Platforms need to establish feedback mechanisms and equivalents of customer support services for those working on them. “If I were starting an Internet company or designing an app for something,” one of our respondents commented, “I would say that we must have phone customer service 24/7.” As platforms come to dominate more sectors of the economy, customers and workers alike will come to expect effective means for providing feedback." (https://medium.com/the-wtf-economy/design-it-like-our-livelihoods-depend-on-it-e1b6388eb752#.s40t0ip2q)