Designing 21st century platform unions—part 1

I’ve been thinking a lot lately about what parallels—if any—exist between platforms and unions, and whether there are lessons that can be learned from platform design, that could benefit worker organizers who are looking for new models. This is the first in a series of three planned posts on the topic. In referring to companies as platforms, I am specifically talking about internet-based businesses that provide some kind of a marketplace for different kinds of users (called here in shorthand “buyers” and “sellers”—though of course, on some platforms a user can be both a buyer and a seller) to meet and exchange something of value (money for goods or services, ideas, advice, etc.). In this series, company platforms can refer to sites like eBay or Craigslist, as well as Mechanical Turk, Uber or Etsy. This series should not be read as an endorsement of any of the company platforms or unions mentioned. For the analysis of platforms, I am indebted to the authors of _Platform Revolutions_, and to MIT’s Sloan School for their online course of the same name, as well as Simone Cicero’s work, including this recent article.

Unions and platforms—what overlaps?

One thing to keep in mind is that, at a pretty basic level, unions and platforms are both about disrupting power in an entrenched system. The point of a union is to decentralize power in a workplace, and to put more power in the hands of union members. The point of a platform is to decentralize power in an industry, and to put more power in the hands of users. Of course, both unions and platforms can be guilty of redistributing power in ways that are non-democratic, and sometimes power just gets moved to people who already, as a group, have status or influence in our society.

Let’s look at some of the qualities of a successful platform:

Discovers unused resources & puts them to work

Platforms that have achieved success have largely done so by tapping into markets for unused resources. Whether it’s Airbnb convincing people with an extra bedroom to rent it out occasionally, Etsy tapping into the desire for artists to sell their work to a national or global audience, or Josephine finding people who love to cook for others—successful platforms have figured out ways to sell things that largely weren’t sold **at scale** before.

Combines low cost of entry with a governance model that protects users.

The beauty of most platform businesses is that anyone with a device that connects to the internet can sign up for them, and there is not generally a high cost to staying on the platform. I might sell something on eBay only once a year—but my account stays active, without costing me anything. Because eBay employs a robust rating system, where both sellers and users rate each other, buyers can see that I’m an “honest” seller, even if I don’t sell frequently.

Uses network effects to grow (but needs rapid growth on both sides, early)

Platforms that grow best grow quickly. Uber isn’t worth much to me as a rider, if when I sign up there are too few drivers for me to ever get a ride. And it’s certainly not worth my time to drive for Uber if I have to wait for hours between ride requests, because not enough riders have joined. Uber and other successful ride-sharing companies have figured out that they need to concentrate on expansion of both sides of their user experience, and cannot merely open a new market with only riders signed up.

Immerses staff in data about users

Because platform business is by its nature conducted online, platform designers need to (and are able to) constantly immerse themselves in user data. Anecdotal information doesn’t cut it, when thousands of transactions are happening in an hour. And because it’s possible to “serve” different versions of a website to a segmented audience, it’s possible to run experiments on user engagement that track whether a new design will change the way users interact with the site.

Mono-homing versus multi-homing

The concept of mono-homing is one way that platform businesses seek to keep users on their platforms. Basically, mono-homing means “I have one home for this particular interaction or practice.’ It’s pretty easy to switch from one airline to another for different trips (ie–multi-homing)—so airlines have developed reward miles, to keep you ‘loyal’ to their particular airline (hence, mono-homing). Smart phone platforms have generally adopted this tactic—it’s hard to switch from an iOS device to an Android one, because you have to replace all your apps, and learn new methods of interacting with your phone. Other platform businesses have grown through multi-homing—there is no barrier to being a driver for both Instacart and Postmates, for example, or for customers to order food delivery through both GrubHub and Yelp Eats.

And some of the parallels, of a successful union:

Reduces information asymmetry between workers and bosses

An effective collective bargaining process has the great advantage of giving workers the same information that bosses have, about both the financial health of a company, and to some degree, the strategy the employer plans to employ to grow the business. Workers with more information have the advantage of being able to make better decisions—which might involve deciding to put more of their total compensation package into retirement savings, or to request skills training, to learn to use a new technology the company is implementing.

Has a high cost of entry, combined with a democratic governance model

For most US workers, the choice to join a union comes with peril. It can involve the risk of getting fired, or having to risk going without pay for some period of time to win a strike. In some parts of the country, bosses have successfully framed unions as a polarizing force, and potential members may face social pressure from friends or family if they join a union. And finally, the actual cost of dues (and sometimes, initiation fees) are much higher than in most other social movements. (The financial cost can, of course, be offset by gains made at the bargaining table.) Unions, at least in theory and often in practice, have a democratic governance model that allows for all dues-paying members to have a vote—usually outlined in bylaws that are ratified by members.

Uses network effects to grow value

There is a fairly clear relationship between the rate of union density in the US, and the level of income inequality: as union density has declined, inequality has increased. More union members in a particular industry in a particular market (say—hotel workers in Los Angeles) will inevitably push up wages in that industry, in that region, as they bargain better contracts. However, more union members will also cause anti-union employers to raise wages, as they seek to avoid unionization. In the best times in our history, unions have also leveraged their political power and social pressure to raise wages for millions of non-union workers through minimum wage legislative or ballot initiative campaigns.

Holds data as proprietary (sometimes even within organization)

Unions have not, historically, been great users of data either for member outreach or for making a public case for why they are necessary & socially useful organizations. While most unions are committed to meticulous upkeep of their seniority list, they do not tend to track as assiduously things like average wage rates, family size, number of member children who are able to successfully attend college, or other pieces of information that might be a boon to labor economists, academic researchers, or public policy makers. In addition, many unions hold tight their member contact lists (sometimes even restricting their own staff from being able to see them). The high cost of member acquisition forces unions into this defensive crouch—but the crouch itself can provide an additional barrier to fully serving members’ needs.

Mono-homing is the default

It’s difficult to belong to more than one union. Not impossible, of course—especially in the performing arts where recent years have seen the merger of SAG and AFTRA, acknowledging that many of their members were doing work in both film and TV. There is no institutional barrier in becoming a member of UNITE HERE as, say, a school cafeteria worker and a member of SEIU as a home health aide—but the high cost of dues may prohibit part-time low-paid workers from joining two unions, if the benefits largely accrue to full-time work in both cases.

What’s the market? (where the analogy fails, but we still have something to learn)

Unlike platforms, which mostly deal in single seller/single buyer transactions, unions are generally aggregating the selling power of a large group–in this case workers in one particular worksite or industry—to one particular buyer—the company for whom they work. That transaction is conducted through the collective bargaining agreement, and is enforced by a group of members (aka union stewards and officers) combined with paid staff (of both the union and the company).


As the rate of US unionization has declined, employers have done more and more offloading of their employment to third parties. Many workers who seem identical to full-time employees on their face, are ineligible for traditional unionization because they are employed by temp agencies or other kinds of labor brokers.

For those of us who are interested in seeding new types of worker organization, or evolving some part of our current organization to be more platform-like, the next post in this series will look at how we can design them using best practices from platforms.

2 thoughts on “Designing 21st century platform unions—part 1

  1. Pingback: “The companies hiring minimum wage workers…act as if they are managing herds.” |

Leave a Reply

Your email address will not be published. Required fields are marked *