When you think of a tech worker, the image in your mind is probably a 20-something making a six-figure income. But the industry is not at all homogenous. Plenty of software developers are making significantly less, often because their coding school credentials are not seen as prestigious. What’s more, only a fraction of workers at a given tech company are actually software developers. Most employees of large companies work in other productive roles like design or data science, or in supporting roles like public relations, recruiting, or community management. Finally, tech companies of all sizes hire vendors to serve food, clean offices, or drive for private transportation networks. Viewed at the scale of the whole tech industry, highly-paid software engineers are only a small portion of the tech workforce.

This provides some opportunity for organizing the tech industry bottom-up. Each role in the tech industry brings with it different areas for agitation. Dividing roles into categories will allow us to create a strategy that targets each role individually and thus build up organizing strength over time.

Service workers

Service workers are generally hired through vendors. They maintain tech company campuses and serve the needs of other tech workers.

The process works as follows: a vendor wins a contract with a tech company to provide a service, such as cleaning, food service, or security. The vendor itself often barely exists as an independent company, with all of their workers, aside from a small HR department, physically working at the tech company’s campus. What’s more, the vendor rarely brings any expertise in the actual service being provided. Instead, the vendor’s expertise is in management, which is often just a euphemism for worker intimidation. They are an HR department for hire, separate from the tech company’s HR department so that the tech company can benefit from service workers’ labor without extending them the pay, perks, and job security of full-time employment. For example, if the worker is wrongfully terminated and wishes to file a legal complaint, the defendant in any suit will be the vendor, not the tech company, whose pocketbooks are protected by the liability shield of having hired a vendor.

There is a leverage point for agitating around deficiencies in service jobs: the vendor’s customer is the company, which is itself made of workers further up the hierarchy. If those workers demand that a vendor clean an area better or provide a different kind of food, the vendor is generally eager to satisfy their customer, since they know they can be easily replaced by another vendor with relatively little disruption. If a tech company’s direct employees can formulate a clear and powerful demand to a vendor, the vendor can be made to comply.

In terms of unionization, service workers tend to have more organized labor participation than other roles in the industry. Drivers, janitors, builders, maintenance staff, and security personnel have traditionally been unionized in some regions, and this extends to roles at tech companies as well. New frontiers in service worker unions are being opened in tech, as well, as food service workers at tech companies are quickly unionizing and winning contracts.

Contingent workers

Contingent workers are hired “on contract.” They are generally hired through a staffing agency. Roles may include Quality Assurance Testers, Content Strategists, Designers (on trial), Policy Enforcement Specialists, workers who train AI algorithms, and many more.

A contingent worker is often barely distinguishable from a full-time employee (FTE), often working on the same products and in similar roles as FTEs. So why the need for this separate category of worker? Sometimes, corporations choose to hire contingent workers because of a temporary need for more staffing, where it wouldn’t make sense to create a new full-time role. For instance, contract designers or engineers hired to create content for an upcoming conference are generally let go, or returned to the staffing agency’s pool, once the project is over.

More commonly, however, tech companies supplement their FTE staff with contingent workers for other, less sympathetic reasons. “Headcount”, the number of slots available for hiring FTEs, is heavily fought over amongst managers at tech companies, but contingent workers do not count against a manager’s headcount. Because contingent workers do not receive the same benefits as FTEs, they are cheaper, even if their base wages are often comparable. And while full-time wages and benefits show up on a firm’s balance sheet as permanent, ongoing expenses, contingent worker expenses appear as one-time costs that are less off-putting to shareholders who want to see recurring costs lowered while revenues rise. Management is thus incentivized to use contingent workers as an “escape hatch” for hiring within the confines of the headcount system.

Contingent workers receive substantially reduced benefits compared to FTEs, partly in order to to conform to federal laws requiring that workers classified as contractors have materially different job descriptions than full-time employees. They are not eligible for equity grants, and their health insurance options are usually provided through their staffing agency, which are always much worse than the options offered by the tech company. Furthermore, contingent workers are not eligible for arbitrary sets of perks extended to FTE workers, including even invitations to company holiday parties. This separation can breed resentment among contingent workers toward FTEs.

Some contingent workers are hired on a trial basis with the promise of converting to FTE at the end of their three- or six-month contract. This practice is popular with many roles, including some of the higher-prestige roles like UI designer. These contracts are often reupped with multiple additional trial periods. The constant threat of losing precarious employment—with the promise of full-time employment always just beyond one’s grasp—has deleterious effects on mental health. In many states, there are legal limits placed on the length of contract work, precisely in order to prevent employers from paying contingent worker prices for full-time labor. However, companies routinely skirt these laws by bouncing contingent workers between staffing agencies, or by moving them to a vendor without changing anything about their work or workplace conditions.

Non-productive employees

Non-productive roles keep the company operating but do not directly create the product the company produces for revenue. Roles may include Receptionists, Admins, Salespersons, Facilities managers, and others.

Non-productive full-time employees receive essentially all of the perks available to tech workers. They are generally on the same leveling hierarchy as other FTEs, and their wages are tied to this leveling system. They receive equity grants, have the same benefits as productive employees, and are given access to campus perks like food, gyms, and transit.

Depending on the role, wages for non-productive FTEs can be much lower than productive FTEs, and they can be calculated very differently. The starkest example of this is in sales departments, which tend to follow the tradition of sales departments in other industries. Sales workers are paid a base salary then are eligible for commissions depending on how much revenue they bring in, with aggressive revenue targets that are generally set from above. Competitive and hostile sales cultures are common, with some companies opting to fire the bottom 10% of performers every sales cycle.

While non-productive FTEs are closer in benefits to productive FTEs than service workers or contingent workers, companies create structural, physical, and cultural barriers to mixing between non-productive and productive employees. Sales workers are in a completely different department than productive FTEs, and usually the first common manager in an engineer’s and a sales worker’s reporting chain is the CEO. This structurally limits how much a sales worker and an productive worker can engage with each other. Physically, sales workers are almost always sequestered in a different building and often on a different campus from productive workers. At some large tech companies, a cultural norm is enforced that productive workers never interface with sales workers. If there are productive employees working on sales products, such as engineers building ad targeting systems, it becomes an unwritten rule that those engineers never transfer into product engineering roles, and vice-versa. This cultural split allows product engineers to minimize their psychological engagement with the fact that their paychecks are paid by advertising, which, anecdotally, most productive workers in the tech industry like to forget.

Productive employees

Productive employees build whatever a tech company produces to directly produce revenue. These workers have the best wages and benefits and highest prestige at a tech company.

From the 1990s, software development teams were organized into a triad of productive roles: Software Engineer, UI Designer, and Product Manager. Software engineers write programs that are sold or generate revenue through some other means like advertising. UI Designers define the visual look and functionality of a product. Product managers are sometimes referred to as the “CEO of a product,” which could be considered a non-productive managerial role, but within the tech industry PMs are uniformly considered part of the productive workforce. Contemporary tech companies often add new roles to the triad, including User Researcher, Data Scientist, Product Specialist, or Community Manager. These roles are not yet fully standardized across companies and the distinction as to whether these roles are treated as productive or non-productive is not yet clear.

While productive FTEs certainly have it the best in the tech industry, there are still points to agitate around. Productive FTEs at large companies can see the top echelons of the industry, but their decision-making power is quite limited in practice, especially when it comes to any decisions around revenue and core metrics that feed into revenue, like active users or time spent. Productive employees are often extremely motivated by impact and creating value for users. In tech, this motivation is encouraged by management and tech workers are indoctrinated at elite institutions like Stanford into a neoliberal mindset wherein profit and value are inextricably linked. When a productive worker’s desire to create user value runs up against company top-line metrics, the metrics win out everytime. Productive workers are forced to build and support products that they do not believe are valuable, like evermore tracking or surveillance software that drives advertising revenue, or attention-abusing products like low-content push notifications or misleading notification badges on new features. These products drive the other metrics that are a proxy for revenue. Despite the cultural barriers erected to keep most productive workers from having to acknowledge they work at a profit-driven company, productive workers generally do bump up against profit-driven constraints and, anecdotally, these conflicts lead to burnout.

The relatively small number of productive employees offers an additional leverage point. Productive FTEs are few in number compared to the user base of successful products. During their growth phases, large companies will often tout their users-per-engineer ratio to inspire staff and potential staff with the promise of power. A million users for each engineer has often been seen as a ratio worthy of bragging about.

As productive employees have the best benefits and usually have very high wages, agitating around benefits and pay is difficult. Productive employees are also very encouraged to identify with the company and with a company’s executives, who, at the largest companies, are given public relations staffs of their own solely to generate internal propaganda. Productive employees also receive equity grants in the forms of stock or stock options, which financially incentivize them to identify their success with the success of the company. Realistically, any marginal hire at a large company receives such a tiny quantity of stock that it should be seen as compensation, not equity. Still, the psychological encouragement to identify with the executives, board, and shareholders of a company is powerful when reinforced with the background noise of a libertarian or Californian Ideology mindset among well-compensated tech employees.

On the other hand, productive workers are aware of the pay inequality along racial and gender lines found among their ranks, and efforts to bring awareness and alleviation of these inequalities has originated from productive workers. A software engineer working at Pinterest in 2013 proposed that companies publish statistics on the gender distribution of their engineering staff, and the movement has taken hold, with medium and large companies publishing annual diversity reports. Salary transparency movements have also taken hold at large companies, organized entirely from below and against the wishes of management. While company management often tries to pass the buck on discrimination in the tech industry by blaming the “pipeline problem,” the lack of diverse candidates coming out of the education system, the reality is that structural discrimination is practiced in the tech industry. This structural discrimination, which suppresses wages for non-white, non-male workers, can serve as a point of agitation for productive workers, as it lays bare the fact that workers’ interests do not align with management’s interests.

To date, large companies have never publicly shown fear that their productive employees would organize. Company executives have little or no experience in fighting union drives. Productive employees are free to publicly express a wide range of political and economic opinions, including support for unions both in general and on tech campuses. However, at the small startup Lanetix where a bargaining unit of software engineers announced their intention to hold a union election, management retaliated by firing its entire engineering staff. Large tech companies have nearly bottomless budgets and every incentive to fight unionization of their productive workers, so we should expect hard fights and significant retaliation as union campaigns take hold on tech campuses.

An Outline of a Plan

I propose to start union campaigns from the bottom of the job role hierarchy and move to the top. At each step, workers further up the hierarchy will offer solidarity downwards and gain experience organizing along the way. Workers at the bottom of the hierarchy, service workers, are suffering most from the ill effects of the tech boom in the form of gentrification-driven displacement, and organizing there is both most established and most beneficial.

Employees further up the hierarchy may be less motivated by the promise of personal gain through organizing. Still, they are fully capable of understanding the injustice of tech companies making record profits while their own workers are displaced from their communities by the wealth created by tech and hoarded by executives and owners. Tech companies sell not only products but also a utopian vision, of automation and convenience and expanded human potential. This same vision is sold internally, and it is hard to believe this vision should be brought to the world when a tech company can’t even take care of its own.

While there is organized labor participation on tech campuses today, the unions involved work strictly on a craft, rather than industrial, model of unionism. Under the craft unionism model, separate unions represent each group of workers in units divided along job role lines. One union will represent drivers, another janitors, and another cafeteria workers. Under an industrial unionism model, all workers recursively employed within an industry would be part of one union. While we should throw our support behind any labor activity on tech campuses, we should keep in mind the radical goal of organizing the tech industry in its entirety. An industrial union containing high-wage, high-prestige tech workers will be far more powerful than a smattering of craft unions supporting service workers, even if they have the informal support of tech workers. Industrial unions in tech also give workers a wider range of issues to fight and win over. For instance, imagine if a tech company is found to be collaborating with ICE in building deportation machinery, as several companies have been, while the company’s own service workers are immigrants living under the threat of deportation. An industrial union could strike against such a company’s collaboration and potentially force the employer to halt.

I have seen the beginning of this strategy in effect at my workplace. Service workers on my company’s campuses have organized and won union contracts, and workers further up the hierarchy have actively supported these efforts. This support has taken the form of agitating inside the company, circulating petitions, and attending meetings and actions organized by the service workers. Service workers and professional union organizers are happy to leverage support from high-prestige tech employees, and tech employees gain first-hand experience working on an organizing campaign.

In my experience, workers from different roles quickly come to identify together when engaged in struggle. Said another way, no comrade in an actual struggle has stopped to ask, “Are tech workers really working-class?”. Solidarity is built in struggle.

Editor’s note: The author of this piece originally wished to remain anonymous, but asked to be de-anonymised as of December 1, 2018.