Chapter 3: Getting Your Swarm Organized: Herding Cats

From United States Pirate Party
Jump to: navigation, search

Back to Swarmwise Contents Back to Library

Previous: Chapter 2: Launching Your Swarm

If the last chapter was about the first six to eight days of the swarm’s lifecycle, this chapter is about the first six to eight weeks.

While the effective swarm consists almost entirely of loosely-knit activists, there is a core of people – a scaffolding for the swarm – that requires a more formal organization. It is important to construct this scaffolding carefully, paying attention to known facts about how people work in social groups. Without it, the swarm has no focal point around which it can… well, swarm.

In building this scaffolding of go-to people, of the swarm’s officers, it is your responsibility to be aware of limits to group sizes that prevent further growth once reached, and break up the groups that reach these sizes into smaller subgroups when that happens.

You also need to be aware that any organization copies the methods and culture of its founder. This means that the swarm will do exactly as you do, regardless of persistent attempts to teach them good manners. The only way to have the swarm behave well is to behave well yourself. We’ll be returning to this observation later in this chapter.


The few people upholding the scaffolding of the swarm will resemble a traditional hierarchical organization. However, it is important to understand that the role of this scaffolding is not directing and controlling the masses, as it would be in a corporation or other traditional organization. Rather, its role and value is in supporting the other 95% of the organization – the swarm – which makes its own decisions based on the values you communicate, and looks to the scaffolding only when assistance, support, or resources are needed.

Nevertheless, to build an efficient scaffolding, we must understand the human psyche as it comes to optimal group sizes and organization theory.

It can be easily observed in any organization that working groups larger than seven people fragment into two smaller groups. There are several theories as to why this happens, but the prevailing theory has to do with the amount of effort we need to spend upholding and caring for relationships within a working group. Let’s illustrate with an example.

In a group of two people, there is just one relationship that the group needs to care for.

In a group of three people, there are three relationships.

In a group of five, there are suddenly 4+3+2+1 = ten relationships. And if we up the group size to the critical seven people, there are twenty-one relationships between people that the group needs to maintain in order to function as a working group.

As we can see in this math, the social complexity of the group increases much faster than the group size. At some point, the group becomes inefficient, having to spend so much effort just on cohering the group that it gets very little or no actual work done.

When we add an eighth member to a group, the number of relations to maintain climb from 21 to 28. So while adding an eighth member to the group adds 14% work capacity to the group compared to seven people, it also requires the group to spend 33% more of their combined work capacity on the task of maintaining the group itself, on maintaining 28 relations instead of 21. At this point, or sometimes at a ninth member (36 relations), the group falls apart.

What we learn from this is that the scaffolding needs to be constructed so that no more than seven people work closely to one another in a given tight context.

We do this in the classical way, by constructing the scaffolding’s organization chart so that no person has more than six people working with him or her in a given context. This means that, for a given geography (like any state, country, city, et cetera) in the organization chart, that geography must subdivide into at most six smaller geographies which has other people responsible for those smaller geographies.

For now, we call this type of officer a geography leader for the swarm. They could be state leaders, city leaders, circuit leaders – any size of geography, but their duties will basically be the same.

(You will recall that we kickstarted the swarm by subdividing it by geography and letting geography leaders emerge through self-organization.)

Also, for every geography, we will probably have four function officers and one or two deputies in addition to the geography leader. (We’ll be returning to these terms and a sample organization chart later in this chapter.) This, again, makes a group of at most seven in total.

So the key message here is that no geography leader should have more than six people working directly with him or her in a given context. This means that we construct a number of organizational mini-pyramids from the top down in the scaffolding, each with (at most) seven people in it, where each geography leader is both at the bottom of one pyramid and at the top of another, the one immediately below, as shown here:

So the smallest of the three magic social group sizes is 7.

The largest is 150.

There is no relationship between these numbers. The number seven appears to come from a practical limit to the effort spent on maintaining a group, as previously explained. The more elusive number 150 appears to be a limit hardwired into our brains.

The number 150 appears in tons of places through human organizational history. It is our maximum tribe size. In a given context, we have the capacity to know this many people by name and have the loosest of bonds with them.

Anthropologists, looking at the size of our neocortex part of the brain and comparing it to those of other primates and their tribe sizes, tend to regard this number as a biological limit.

This limit is also known as Dunbar’s Number, or the Dunbar Limit, from British anthropologist Robin Dunbar who first wrote about it.

If you are working at a company which has less than 150 employees, odds are that you know them all by name – or at least, you have the capacity to do so. Beyond that size, you start referring to anonymous people by their function rather than referring to people you know by their name. You’ll go see ”somebody in Support”, rather than ”having a talk with Maria or Dave”.

The most successful companies, organizations, and cultures are keenly aware of this human limit. To take the Amish as one example, as their settlements approach 150 in size, they split the settlement into two. The company Gore and Associates – more known as the makers of the Gore-Tex fabric – never puts more than 150 employees in a single plant. There are many more examples.

The effect on your organization-building is the same as in every other successful organization: you need to know that groups above 150 people in size will lose the social bonding required for efficiency and, well, the fun.

However, you probably won’t have any formal group of this size. Rather, it is the informal groups that inevitably form that you need to pay attention to, and how they – once they hit this limit – can prevent further growth of the swarm.

In particular, you need to pay attention to the initial and horizontal team of people that will gather in a chat channel or similar, probably titled “chat channel about everything related to the swarm”. This organically-formed group will have a glass ceiling of 150 people in size, and unless you are aware of these mechanisms, that glass ceiling won’t be noticed. When this happens, further necessary growth of the swarm will be prevented, as more people can’t be socially integrated into that initial chat channel.

Therefore, it is your task to make sure that there are social subswarms everywhere that can attract and retain new people, and not just one centrally located chat channel. These subswarms, too, will have that social maximum size of 150.

Finally, the third magic group size is 30. This is a group which falls between our tight working group and those we know by name, but not much more: we are capable of knowing more than just their names in the group of 30, we know a couple of interests and curious facts about the others in this group, but we can’t work tightly with all of them. It can be thought of as an extended family.

You will probably have a couple of formal groups that are about 30 people in size, like the assembled group of all officers and leaders for a certain function or geography, but in general, you should strive for the seven-person group. When looking at how several of these groups cooperate on a daily basis, if you notice that some groups cooperate closer than others, you should be aware of the 30-person group size limit. For example, if the people coordinating all the aspects of the work in a particular city starts approaching 35 people, then that group is blocking further growth of the swarm and should be divided into two, allowing for more growth: divided into two groups handling the north and south parts of the city, for example.

After reviewing this, we also realize why we divided the swarm by geography in chapter two, and tried to have not more than 30 geographies. There’s you, who founded the swarm, and you communicate directly with the (at most) 30 geography leaders.

If you did this, then three to four weeks into the swarm’s lifecycle, it is suitable to insert a layer of officers between you and those 30, so that you communicate directly with five or six newly-inserted geography leaders, and they in turn communicate with five or six each of the original geography leaders.

So to summarize the important part of this: keep formal working groups in the scaffolding to about seven people. When several groups are working together, try to keep the size at or below 30. Finally, pay close attention to when informal swarm groups approach 150 people in size. When that happens, take steps to break them up in smaller subgroups.

(I first learned of the different dynamics of these three group sizes – 7, 30, and 150 – as part of my Army officer’s training in my early 20s. It is no coincidence that they correspond to squad, platoon, and company size, respectively. Since then, these group sizes have reappeared in almost every leadership training and management workshop, from one aspect or another. More importantly, all my experience with building swarms confirm their importance.)


All this talk about leaders and formal structure sounds very… conventional, doesn’t it? We’re building this thing called a scaffolding, but it sounds very much like a traditional, hierarchical, boring organization. So what is new?

The new part is the entire swarm around the scaffolding, and the role that these officers – these geographical and functional leaders – must take in order to support it.

One key insight is that the responsibility of the swarm leaders is not so much managerial, as it is janitorial. Nobody answers to them, and their task is to make sure that the swarm has everything it needs to self-organize and work its miracles.

Remember, leadership in a swarm is received through inspiring others: standing up, doing without asking permission, and leading by example. In this task, the various officers and leaders have no organizational advantage over other people in the swarm: those who inspire others in a swarm, cause things to happen.

Put another way, the leaders and officers are not somebody’s boss just because they have some responsibility.

The first time you see people self-organize, it feels like magic. What you need to do is to communicate very clearly what you want to see happen and why. If people agree with you, they will make that happen, without you telling a single person what to do further. They will self-organize, and people interested in making it happen will gravitate by themselves to a subtask where they can help deliver the desired result. Each person will do this in their own way according to their own skill set, with no assignment or microsupervision necessary, causing the whole of the task to happen.

This is also a key mechanism in swarm organizations. You cannot and should not try to tell anybody in the swarm what to do; rather, your role is to set goals and ambitions, ambitions that don’t stop short of changing the entire world for the better.

We have seen something similar happen already, when the first onslaught of activists happened in chapter one, and several hundred people were waiting for instructions. You told them to self-organize by geography and choose leaders for the geographies. That was a form of self-organization, albeit a rudimentary one.

In a swarm, working groups will form by themselves left and right to accomplish subtasks of your overall vision, subtasks you haven’t even identified. This is part of how a swarm works and why it can be so effective.

So once the scaffolding of officers is in place, with its responsibility to support the swarm, groups and activities will form all over without any central planning – and importantly, without any central control.

Your passion for the swarm’s mission is going to be key in making this happen. You need to constantly show your passion for the end goal, and those who see and pick up on your passion will seek out things they can do to further it – all on their own.

Your role in this is to lead by example. People will copy you, in good weather and bad. Therefore, make sure you’re being seen in good weather. More on this later.

Another thing you will notice as the self-organization starts to happen, is that it doesn’t necessarily follow geographical boundaries. This is fundamentally good; you will have groups that form around accomplishing specific tasks that are geographically unbound, as well as groups that form around tasks that are bound to a specific area by nature. The task of producing a press center isn’t tied to a city, but the task of handing out flyers is. When people self-organize, this is taken care of by itself.


There are three key concepts the swarm organization is optimized for: speed, trust, and scalability. When building the Swedish Pirate Party, this was a deliberate decision from the start, and it proved very successful.

We can optimize for speed by removing all conceivable bottlenecks. A swarm is typically starved of money, so it must compete on other grounds. Its reaction speed and reaction weight are more than enough to offset the lack of funds.

We can optimize for trust by keeping the swarm transparent and giving everybody a very far-reaching mandate to act on their own. We would establish this mandate by very clearly communicating that different people drive the swarm’s goals in different ways, and that we all trust one another to do what they believe is best, even if we don’t understand it ourselves. The three-activist rule, which we will discuss shortly, is a very efficient way to achieve this.

We can optimize for scalability by constructing the entire scaffolding at its finished size at the swarm’s get-go, providing space in the organization chart for geography leaders down to neighborhood level. However, we would leave upwards of 99% of the roles in the scaffolding empty for now – below the original 30 geography leaders, nothing has been appointed yet, despite us having another six or seven layers of empty boxes in the scaffolding’s organization chart. This means that these geography leaders can and will grow the organization downwards as activists volunteer to become new geography leaders at lower levels in the scaffolding. Then, those leaders will grow the organization in turn, and so on.

A swarm optimizes for speed, trust, and scalability.

The first time you notice that somebody you’ve never heard of has been appointed to formal responsibility, it feels like magic, and it shows that the scaling-out is working.

A swarm grows by people who are talking to people at the individual activist level. You don’t have the luxury of putting out ads, but your passion and desire to change the world for the better (along with a complete denial of what other people would call the impossibility of the task) makes people talk among one another. This is how your swarm grows: one conversation at a time, one person at a time.

This is how the Swedish Pirate Party grew to 50,000 members and 18,000 activists. One conversation at a time between passionate activists and potential new passionate activists.

In general, we can divide the people of the swarm into three groups by activity level: Officers, activists, and passive supporters. The officers are the people in the scaffolding, people who have taken on formal responsibility of upholding the swarm. Activists are the actual swarm, the people that make things happen on a huge scale. The passive supporters are people who agree with the goals as such, but haven’t taken any action beyond possibly signing up for a mailing list or membership. (The passive supporters may sound less useful to the swarm, but that’s not the case: they are the primary recruiting base for the next wave of activists. We’ll discuss this more in chapter eight, as we look at the Activation Ladder.)

So let’s take a look at what officers would typically be needed to support a swarm. In other words, let’s look at a template organization chart.

Let’s take a typical geography as an example. It could be a county, it could be a city, it could be a state, doesn’t matter. From the experience with the Swedish Pirate Party, we know that a particular geography works best when there is not just one geography leader, but a leader and a deputy that divide the work between them and who cover for one another. These people become go-to people for everything that happens in the area. The advantage of having two people is that people can drop out for a while from time to time. We can change jobs, we can fall madly in love, we can get sick, or lose interest in activism briefly for a myriad of other reasons. This is human, and always okay. If there are two people sharing the workload, the activity doesn’t stop when one drops out for a while. Most geographies had one deputy geography leader, some had two.

“If you feel you need to take a break from activism, that is always the right thing to do. It’s always better to get rested and come back, than to burn out and get bitter. There will always be something to do when you come back: you don’t have to worry about the world running out of evil while you’re away.” — Christian Engström, Member of European Parliament

Over and above this, drawing from experience if designing an activist swarm today, I would have four function leaders at every geography in addition to the overall leaders: one function leader each for PR/media, for activism, for swarmcare, and for web, information, and infrastructure. (These are roughly in order from most extroverted to most introverted.) All of these could – and maybe should – have their deputy in turn.

The person responsible for PR/media would be responsible for interfacing the swarm to legacy media at his or her particular geography. That includes sending press releases, making sure press kits with information are available, and other things related to serving oldmedia with information about the swarm and its activities. (We’ll be returning to exactly what this is in chapter nine.)

The activist leader would not lead activism as such, but rather, support it (as is the case with all of these roles). Whenever activists decide swarmwise that they want to stage a rally, hand out flyers, put up posters or do some other form of visible activism, this is the person responsible for the practical details such as PA equipment, permits, and other details on the ground to make things happen.

The person responsible for swarmcare would welcome new activists into the swarm, and continually measure the overall health of it. A typical task would be to call new activists just to make them feel welcome, and tell them when the next events – social as well as operational – take place. This is more than enough for one person to chew.

Finally, the information-and-web guy is the person who maintains the infrastructure of a blog or other web page that summarizes the relevant information of the swarm in this particular geography. (This person also communicates internally when events, such as rallies, happen. The swarm decides when and if they happen; it is the job of this person to communicate the consensus.)

Of course, your needs may vary. Consider this a template that you can use as a starting point. In any case, these boxes are all empty to begin with; organic growth is crucial.

People should not be appointed to these positions just because it’s fun to have a title; rather, the organization chart should lag slightly behind the observed reality. When somebody has already taken on the de-facto role of fixing all the practical stuff for rallies, for example, and everybody already knows that that person is the one to call to sync things up to get the PA – that’s when the org chart should be updated to reflect that. The person who should update the formal roles is the geography leader, who is responsible for keeping the swarm at optimal conditions in this particular geography.

One person should have one role in the scaffolding, with any kind of multi-role person being a temporary measure. In this, watch out for people who start advertising many titles in their signature or similar places – that’s a sign they’re more after the titles themselves than a single responsibility to do well.

Empty boxes in the scaffolding’s organization chart are not bad. They can and will fill up as time passes and groups fill up to the magic size limits, and need to break out into subswarms. Don’t unnecessarily appoint people to roles because you think empty boxes look bad: an occupied box will block somebody else from filling that role, and so, may be preventing the overall growth of the swarm if the person originally appointed to the box wasn’t really interested.

Do not be afraid of empty boxes in the organization chart.

So do not be afraid of empty boxes in the organization chart. They provide opportunity for somebody to step up to the plate informally, at which point the chart can be updated to reflect reality. It can help to think of the organization chart as the map rather than the terrain – when there’s a conflict between the two, the terrain wins every time. The organization chart is an estimate, at best, of what the organization actually looks like.

(This does not apply to military maps. When those have misprints, the military modifies the terrain to match the map, which happened at least once during my Army term.)


In a typical office setting, people keep in touch about day-to-day operations in quite natural ways – by bumping into each other in the corridor, over coffee, but also in formal meetings. When working with a swarm, almost all of the cooperation happens over a distance – so you must find ways to compensate for the lack of eye contact and subtle body language that otherwise keeps a team jelled.

One of the easiest ways to do this is to have regular meetings over the phone or over a chat line where you just synchronize what’s going on and where people are with their respective work items (volunteered work items) to make that happen. The purpose isn’t for you to check up on what’s going on – the purpose is for everybody to know the state of the whole.

These meetings should be limited to 7 people if on the phone, or 30 people if in a chat channel. Otherwise, they can quickly turn into noise. You should have such regular heartbeat meetings once a week or once every other week with the people closest to you in the swarm’s scaffolding, and those people in turn should ideally have heartbeat meetings with their nearest crew, as well.

Some swarms or subswarms have preferred physical meetings. While such meetings provide for a lot higher bandwidth and opportunities to sync up, prevent conflicts and brainstorm ideas, their timing and location can also serve to lock out activists from engaging in the swarm – often inadvertently. For example, if you have a subswarm in a city that meets every Sunday afternoon, you can get lots of students engaged in the swarm – but the choice of Sunday afternoons will make sure that no working parents will ever show up to the meeting, as this is prime family time. Such factors need to be considered, and it is easy to be blind for limitations outside of your own demographic that prevent people in a certain stage of life from attending.

Once method I used to make it easier for people to attend the party management meetings when I was party leader was to limit the meeting to a strict time frame. We would start the meeting at 8pm on Tuesdays, and the meeting would end at 9pm, no matter whether everybody thought we were finished or not. That made sure that two things happened: it let people know that they could plan things with their family from 9pm in the Tuesday evening, and it forced people to take the important things first, as the meeting cutoff would happen whether they were done or not.

In short, the simple rule of having a hard meeting cutoff time made sure that people (including me) didn’t waste other people’s time.


Speaking of wasting other people’s time, some activists will tend to take meetings a little too seriously. It is important that you maintain meetings as a necessary evil, because people who are eager to be part of the swarm can easily see meetings as the purpose of the swarm – they will tend to see meetings as work itself, rather than the short timeframe where you report and synchronize the actual work that you do between the meetings.

Bureaucracy and administration will very easily swell to become self-justifying, even in a swarm of activists. Do not let this happen. Keep reminding people that meetings are there for the purpose of synchronizing the work done to forward external purpose of the swarm, and that every minute spent with each other is a minute not spent changing the world.

In particular, activists in a subswarm dealing with oldmedia (newspapers, television, etc) can easily become self-absorbed in their own titles: “I attend the media meetings, therefore, I work with media, and thus, I am really cool”. We’ll return to that particular problem in chapter nine.


As the swarm’s founder, you must be aware of human psychology of leadership. People will do as you do, exactly as you do, even if and when you are having one of the worst days of your life.

If you show yourself in a thoroughly wretchy mood to a swarm of 50,000 people, they will all emulate your behavior from that day, down to the most minute of details. This is not what you want.

So, ironically, one of the most important parts in founding and leading a swarm is to take good care of yourself. Sleep well, eat well, work out, allow yourself time and space to breathe. This is for the good of the swarm, and has the nice side effect of being good for you, too. If you feel aggressive, short-tempered and frustrated one day, you should probably refrain from all interacting with the swarm until that passes; if you don’t, those moods will become core organizational values.

If you’re not taking care of yourself, you’re not taking care of your swarm.

On the flip side of that coin, understanding, patience, collegiality, and passion are values that you want to show. Be aware of your own mood, and know that the swarm copies from you – whether you are behaving in ways that sets the swarm up for long-term success or catastrophic infighting, the swarm copies your behavior in more detail than you can notice consciously.

One value that you must absolutely communicate for the swarm to work is trust. You need to trust in people in the swarm to further the swarm’s goals, even if they choose a different way of doing so than you would have chosen, and even if you can’t see how it could possibly work.

You also need to communicate that everybody must trust each other in this regard. Leading by doing is necessary here, but not sufficient; you need to periodically repeat that one of the core values of the swarm is that we trust each other to work for the swarm in the ways that we can do so as individuals.

It turns out that one thing that makes swarms so outstanding in efficiency is their diversity. People come from all walks of life, and once they realize they have a full mandate to work for the swarm in the ways that they can, they will just do so.

In the Swedish Pirate Party, we had manifested this through a three-pirate rule, which can easily be translated into a three-activist rule for any swarm. It went like this: if three activists agree that something is good for the organization, they have a green light to act in the organization’s name. It’s not that they don’t need to ask permission – it goes stronger than that. Rather, they should never ask permission if three activists agree that something is good.

Asking permission, after all, is asking somebody else to take responsibility for your actions. But a swarm doesn’t work like that. Also, the person who would have given that permission would probably be in a worse situation to determine if this action would work in the context the original activists had in mind.

Of course, many balk at this. Letting activists run loose like this? Trusting them with your name and resources to this extent? I heard frequently that it would be a recipe for disaster.

In the five years I led the Swedish Pirate Party, peaking at 50,000 members in this time, this was not abused once. Not once.

It turns out, that when you look people in the eyes and say ”I trust you”, and give them the keys to the castle, many are so overwhelmed by the trust that they don’t hesitate a second to accept that mantle of responsibility.

It’s also important that this was only a mechanism for self-empowerment, and never a mechanism that allowed three activists to tell somebody else what to do or not do.

As a final note on trust, the part about trusting people to act for the best of the swarm is crucial. This means that there is never a blame game; if something goes wrong, the swarm deals with it after the fact, and never spends time worrying what may go wrong in advance.

If something doesn’t go as intended, the swarm learns from it and moves on. On the other hand, if something is wildly successful, it gets copied and remixed across the swarm to new variants to get even better. This happens organically, without you needing to interfere, as long as activists can publish their successes.

In the next chapter, we’ll take a closer look at how activists of the swarm interact with the outside world, learn from mistakes, and remix the successes to evolve and improve.

Next: Chapter 4: Control The Vision, But Never The Message