This is a personal reflection of how our attempt at organizing fared, written with some input from my comrades. It is meant to be helpful to people engaged in labour organizing in the tech sector. It is a personal account and in no way expresses the views of TWC as a whole, and many TWC members would dispute my conclusions. I use the term “we” to vaguely refer to everyone active in the project at any given time. Throughout, I use language from Organizing for Power, in short “O4P” (in particular “Wall Chart”, “Structure Test”, and “Majority Petition”) as well as from Secrets of a Successful Organizer (in particular the Bullseye model on page 14).


We work for a German tech startup with about €100MM in funding, 180 employees in total, of which 80 are in Germany. The company is remote-friendly, with about half the German employees working remotely.

In the middle of 2023, 15% of staff are laid off. In the German office, these are mutual termination agreements (Aufhebungsverträge). In response, a signal group for people interested in the general idea of founding a works council emerges. After some initial experience gathering, we heed advice we received from people who’ve done this at other companies and found a second group to which only people who have attended at least one call are invited. In this group, we decide to use the invitation for the forming of a works council as a majority petition that we want to have co-signed by at least 60% of the workers in the German office, equivalent to about 45 signatures.

With 5 people in the small “Core” Signal group and 15 in the larger “Activist” group, we believe we can quickly get results. We create a spreadsheet (an O4P “Wall Chart”) of all the German employees along with our assessment of their likely support and who shall talk to each. We have one self-training for 1-1 conversations in the group with 4 people participating. However, our initial message soliciting signatures in the Activist group is largely ignored. Small administrative tasks in the Core group do not get done as expected. We miss our own deadline of end of September. We quickly get signatures from people who are politically supportive or good friends with the organizers but fail to go beyond that. The process becomes more arduous. People in the Core become less responsive and do not follow through on commitments to talk to specific people. At the end of October, we have 11 signatures and 17 people who were spoken to and seem like likely signers. Throughout November we flounder. Some Supporters leave the company. In December, we make a push to reach clarity by the end of the Christmas party, which we want to use as a last chance to talk to people in person. We end up with 26 signatures, about 35% of workers, missing our goal of 60%.

Because many non-signers indicated that they support the forming of a works council, we decide at the end of December send the invite anyway, but without the signatures. We bumble around with the administrative prep for a few weeks. In mid-February, in light of the lack of engagement and the lack of candidates for a works council, we decide to abandon the project altogether, about 7 months after we conceived it.


While the original spark for the project were layoffs, we encountered multiple grievances from workers:

Bad Communication: In the regular company-wide meetings, leadership would give conflicting information, often seeming more like the week’s hot takes rather than reflected conclusions. This was especially jarring when talking about layoffs or performance, with blame shifting from the overall economic climate to poor C-level decisions to underperformance of the workers as a whole.

Increased Hostility: People reported more yelling, insults, and other un-cooperative behavior.

Increased Pressure: As part of a shift to more monitoring, weekly individual goals and regular conversations with managers were introduced.

Compensation: As part of budget-cuts, a general pay-freeze was indicated. When asking for inflation-adjustments, people were given conflicting answers, often being told that they just missed the time for receiving a raise or that no compensation increases were possible at all.

Strategic Ideas

We wanted to use the strategies from Organizing for Power in our organizing. We believed that founding a works council was achievable but that one founded by a group of conspiratorial leftists would be ineffectual and might hurt further organizing efforts. We figured that a works council backed by a majority of the employees would be stronger and that the initial win of founding one would be a great success to then build more power from. And so we wanted to make the invitation for the works council, which formally only requires 3 signatures, a majority petition as outlined in O4P. We wanted to use it as a structure test so that by the time it came to electing the council and doing work in it, we’d have a good idea of who our organic leaders are, who can get people to take action, and what our power base is.

We also considered it a benefit that we’d be limited to the German workers, figuring it’d be simpler to get a majority in a subset of the company and then expand to the rest of it rather than trying to start with the whole company all at once.

What we should have done differently

We made various mistakes on our journey. I will use the Bullseye framework from Secrets of a Succesful Organizer to organize these. In the Bullseye model, you sort people into “Core” (people who are constantly organizing), “Activists” (people you can rely on to do a particular task, like collect some signatures), “Supporters” (people who take action to support you but are otherwise inactive), “Disengaged” (people who don’t care or don’t think organizing can help them), and “Hostile” (people who are actively opposed to the organizing).

Organizing should always try to move people to the next-more engaged ring of the Bullseye circle. In our case, we considered a “Supporter” anyone who’d signed the petition. I think we made mistakes at all of these rings and some problems fit into more than one.

Disengaged to Supporter

Friends vs Colleagues: We did not anticipate the unique challenges Activists have in talking to their friends rather than their colleagues. Many of our Core and Activist members struggled with making requests of their friends; while they had an easy time broaching the topic, they were bad at posing the hard question of whether their friend would sign or not and were bad at requesting and following up on commitments. If doing this again, I’d recommend assigning the worker’ non-friends for the 1-1s and asking their friends for prep in the conversation.

Lack of Urgency: We failed at creating urgency, giving Disengaged workers an easy out for delaying their decision or waiting to see developments. Since layoffs had already happened and there was no natural upcoming problem, it was hard to convey that now was the time to sign an invitation and start something. Since a works council is very preventative and bureaucratic, any possible benefits from it would only be seen much later. The lack of urgency was also a problem among the Activists. Looking back, we think the existence of big group chats is actually a net-negative since the perceived lack of engagement dilutes the sense of urgency any participant might feel. The next time, we’ll have clear rules of engagement for them. It could also be helpful to have an outreach Telegram broadcast channel where not everybody can post, only posting when there’s occasion for discussion and then letting that discussion happen in the post comments.

Weak Campaign Goal: Ultimately, the forming of a works council was a weak campaign goal that did not get people fired up. We failed at reliably communicating how it would help address the grievances people had. It also limited us to only target German employees and made us ignore those with different contracts, even when they would have been valuable assets. We did not appreciate enough how the German/non-German distinction did not make sense to departments outside of Engineering. I don’t know what better campaign goal for a majority petition there would have been though. Other ideas are transparency about stock awards and pushing back against hustle culture.

Supporter to Activist

We failed at converting organic leaders. We did an alright job at identifying them early on but then did not put adequate focus on converting them. Part of our problem was that – because they were organic leaders – they had good connections to multiple Activists and so the responsibility for engagement was split and thereby diluted. I believe that occasionally being canvassed by multiple people also created more confusion and hesitation in them. By assigning the organic leaders to people who had a social connection to them we also ran into the issue again that Activists find it difficult to make requests of people they consider friends rather than colleagues.

We did not do a good job at list work. While we created a good wall chart of all employees we’d have to talk to, including an early assessment of their stance, we became too concerned about the sensitivity of the data in it and gave it in care of just two people. This decreased visibility, likely made people feel less involved, and created a dynamic of power and nagging from the people with access to the sheet. It also meant that in most conversations with Supporters, no list work in the form of getting their input on their colleagues was happening. I am, however, not sure what a wall chart would look like that would satisfy the privacy and security expectations of a German tech crowd. The next time I’d try to use a pseudonomized spreadsheet that protects identities from being easily found out while still allowing for a central, shared document.

Activist to Core

We did not distinguish within the Core group enough. Early on, we considered every person who’d shown up to a video call to be Core. In reality, there were four people who were committed enough to regularly show up and do the work. I believe we would have been more effective if we’d accepted those people as core; in fact I’d recommend making participation in trainings the defining characteristic of who’s Core and who’s an Activist. E.g. when we were only 4 in our 6-step conversation training, I felt that we had a weak Core consisting of 7 people; instead I should have assumed that we had a good Core consisting of 4 people.

Grieving/Personal Reflections

This section is more personal and primarily meant to help me move forward from this failure productively. I don’t think you can learn much about what to do from it, but it may be helpful for somebody thinking about strategy and emotional management.

Overall, I’ve understood more that my ambitions for labour organizing in tech are different from those of others, often quite radically. The first is that I do not actually care about working conditions of tech workers in their own right – the workers are by and large doing fine and if I wanted to spend my time making people’s lives better, there are classes of people (care workers, gig workers) who’d benefit significantly more from it. Instead, I care about tech organizing for instrumental reasons, as a way to achieve changes in ownership structures and move towards socialism.

I believe that the tech world is useful in this respect because there is a lot of unrealized labor power and the startup world especially has structures that make it interesting for labour action. In my ideal world, tech workers would use the crunch of funding rounds to organize for enshrining employee co-ownership (and that includes co-determination, not just options with cash value), possibly by transferring parts of the company to a co-op that all employees are members of. We would spearhead a move to 4-day work week and become a vanguard for better working conditions across the economy, not unlike the unionized workers who got us the 8-hour day. We would demonstrate that exploitation is an opportunity; that the chunk of value that the owners of our workplaces take from us is high, especially in the tech world, and that we can and should appropriate it.

The dominant force in German tech startup labour organizing is the TWC, which focuses on forming works councils in as many companies as possible. I no longer think that this is a good strategy; in fact, I think the forming of works councils is a red herring. They’re good primarily for defensive purposes, they are extremely unexciting, and they’re so bureaucratic that they make involvement of one of the big, state-recognized unions a necessity. The unions, however, do not understand the tech startup world and are more interested in gatekeping, growing their membership numbers but not their power, and occasionally winning pay rises, rather than tackling the political-economical status quo. My understanding of the works councils that have been formed in tech companies is that they struggle to build connections with their base. Works councils also enshrine the idea that organizing work is done by others (i.e. the councillors), rather than the collective of workers as a whole.

On an emotional level, I’m struggling to not fall into a mood of blaming my colleagues and fellow organizers for the project failure. It is easy to say, as one of the workers I had an organizing conversation with did, that “some people just do not deserve better”. I’ve certainly been struck by how invested some of my colleagues are into the idea that they’re powerful and can defend their own interests while watching those very interests be trampled by our bosses. I find it hard to understand how you can let yourself be insulted by your manager, roll over when the pressure goes up, be scared of being fired for being seen organizing, and still say that you’ll just switch jobs if it gets too bad. And so I find it hard to connect to people in organizing conversations because the grievances they express, and the actions they’ve taken to address them, seem to just not make sense. I now believe that for a startup tech context, an emotion-first conversation about power and solidarity would be more beneficial rather than going for expressed grievances. At any rate, the model of 1:1, 6-step organizing conversations from O4P and Labor Notes need to be adapted.

I don’t know to what extent I’ll be continuing with labour organizing in the tech space. I feel like most people I talk to are actually just interested in working conditions or interested in doing something that’s loosely good, not dissimilar to how they might choose to work on FLOSS. I’m interested in building power and destroying private ownership of tech companies and I don’t know where to find the people that share this ambition.

Get involved! @TechWorkersBER