Oct 10, 2017
EDIT Sept 2018: I am no longer convinced that the model of community building presented here is a particularly useful model for considering EA community building as opposed to any other model that could be conceived. I hope to write a future article on how we use models to design and develop strategy as community builders.
That said I still think most of the advice here is good advice to consider when community building.
For the past year I have been funded by the EA community in London to grow, run and support the community. When setting out to write up my findings from the last year I decide to split it into a few parts:
This document aims to capture my intuitions and general views on EA community building. Including this last year I have nearly 7 years experience on EA community building work and I feel I am starting to build up ideas of what works and wanted to get the models in my head onto paper.
A further document [link pending] will cover the hard data on event attendance, what worked in London and specific advice on growing a large local (non-student) EA group.
There is also an Impact assessment (and a document on future plans [Link pending].)
In general advice on building a community, especially armchair theorising such as this piece, should be taken with a big pinch of salt. It is difficult to generalise and there is often large disagreements on even basic assumptions. See the Annex at the end of this document on: what can we usefully say about community building?
In this document I suggest some models of EA community work in order to attempt to capture my intuitions. I aiming for usefulness over veracity. Some key conclusions of these are:
We can consider EA as: EA ideas, the global EA movement, and smaller EA community groups.
EA communities have an impact primarily through growth (spreading and reinforcing EA ideas), retention (keeping EAs engaged) and networks (making connections between EAs).
I suggest model I find useful for thinking about the benefits of community growth:
Impact of Growth = Access x (Propensity for EA x Offer) (Impact Potential x Offer)
It is important to consider the relevance and quality of the Offer, (ie. the advice given on how to do good). What you offer determines both how interested people will be in your message and how much more impact they will have as a result.
Considering the Propensity for EA and what you have to Offer of a potential audience of EA outreach work can suggest a strategy. As set out here:
Signal interest. Build connections with individuals. Eg. finance.
Start senior & cascade down. Eg Civil servants. OR Go for mass outreach. Eg. Students
No clear offer
Reconsider outreach in this case.
Work with small groups and invite people to develop an offer with you. Eg. Quakers,
I also provide further advice on other aspects of community building. Including:
Community builders should be cautious and risk averse. The rate the EA movement grows is less important than the eventual shape of the movement and ideas.
The easiest way to start a community is to build upon existing connections within the groups you want to outreach to.
You may want to consider gatekeeping your community or your events. There are many different ways to do this and it does not necessarily lead to lower audiences.
I also provide thoughts on other topics such as maintaining momentum, being welcoming (diversity, values, etc), freeriders and strategy
I propose we split actions to build up EA into the following categories:
EA ideas. Intellectual. People adopt EA ideas or assent to one or more of EA’s core claims (Eg charities effectiveness matters). They may act on these ideas.
EA movement. Tribal. The global EA community. People adopt an EA identity. They are likely to want to associate with other EA people, explore EA ideas and take action.
EA community groups. Normally Local, sometimes digital. Communities meet and connect around EA topic. Attendees get to know one another
Eg: local EA communities, student groups, Facebook groups, an EA academics forum.
Usefulness: I found this breakdown useful to consider when writing this document
I would break down the impacts of building and maintaining a community as follows:
Impact Total = Impact Growth + Impact Retention + Impact Networks + Impact Other - Risks - Costs
Impact of growth.
New people join the community, adopt EA ideas, and the EA movement grows. People already aware of EA learn more and get a better grasp of EA ideas. This spreading of EA ideas creates a positive impact for the world as people will go out and be better at doing good. (The value of spreading EA has been written about already, for example here and here and here.)
Impact of retention.
Without a community people may drift away from EA and do less good (some EA London members, including ‘core EAs’, said this could have happened to them). It also seems plausible that people who engage with EA at university drift away after graduating (EA London has picked up such people). It is also possible that the EA movement is shrinking. More research into retention would be useful.
Impact of networks.
A regularly meeting group of people can collaborate better, learn from one another, offer welfare support to each other and can thus achieve more than they could alone.
Case study: One attendee at EA London policy events used the contacts she made to found an All Party Parliamentary Group (a collection of MPs and Lords) to look at how policy impacts future generations.
Local communities can have a bunch of indirect impacts including:
Support. The community organiser can further support community members by providing advice or directly helping them in their attempts to do good.
Learning. Lessons learned may be applicable to other EA community projects.
Personal benefits. Running a community could be good for the organisers career capital.
Up-skilling (of non EA ideas). If your community involves informative events such as talks or workshops this could boost the general skills and knowledge of community members.
Running a community takes money and time that could potentially be used better elsewhere. (For an example see EA London’s impact assessment.
The global EA movement will hopefully achieve great things. So with any community building activity we should be especially careful about ways we could accidentally harm the movement or decrease its potential. See section below on caution.
Usefulness: I found this breakdown useful to consider when considering the various impacts and costs of my community work, both to assess work done or develop strategy for future work.
I suggest thinking about the expected benefit of community growth in the following way:
Impact Growth = A x (P x O) x (I x O)
A = Access – how well you can reach and talk to a group.
P = Propensity to take onboard EA ideas.
O = Offer – how can you help this person do good
I = Impact potential, eg. wealth or time
• (P x O) tells you the interest - how attracted a particular group will be to effective altruism.
• (I x O) tells you how much impact people in this community could have if engaging with EA.
Usefulness: Thinking through this model has been useful for me in planning how to outreach to different groups. In particular it:
Reinforces the importance of having a good offer
Suggests a strategy growing a fledgling EA community, by considering P and O
Breaks down the things to consider when outreaching EA through a community.
Access is how well you can reach and talk to a group. We can break down into:
Available potential community members. People you already have contacts to who would be willing to be part of this community. This is the biggest factor to getting access. (See section below on Getting Started).
Eliteness. How elite / high status people are inversely affects how easily they can be reached. Gatekeeping may be needed for elite communities (See section below on Gatekeeping).
Skill. The persuasiveness and skill of the person doing the community outreach work.
Other factors. Existing channels for communicating to your audience on mass. Willingness of people in target group to talk to their colleagues about doing good. Etc. Etc.
P is the propensity to which people who you are trying to reach with your community are likely to take on board EA ideas. Roughly it is the percentage of people who are:
Altruistic enough to care about having an impact,
Interested in effectiveness enough to care about doing the most good
Flexible enough to change any existing altruistic plans they have. (Eg Founders Pledge reaches startup founders who are flexible about how they spend their future money exit.)
Remember, people already voluntarily engaging with EA are easy to access and there is clear evidence they have a propensity to accept EA ideas. It can be valuable to focus community building efforts on such people. Eg GWWC puts effort to help people who have pledged donate effectively.
EA provides useful tools to support people to do good.
Sometimes EA has a strong offer, for example if you are a student making a career decision EA has a lot of advice that can multiply your impact. EA has produced valuable research on early career choice, cause selection, charity evaluation, existential risk, and animal ethics.
Often EA has a very weak offer, (at times I have tended to forget this, assuming that because EA was useful to me, others will find it helpful). We currently have little to offer to people who are not utilitarian, not at the start of their career, interested in activism, want to create systemic changes, already experts in global poverty, have communities encouraging them to do good, have limited capacity (time/money) for doing good, etc, etc.
Case study: when I was running THINK lots of universities were successfully starting EA chapters. Duke however, was struggling because it already had an abundance of do-gooder-y societies offering high quality support on how to make change. The EA society had less additional benefit to offer.
In the formula above, O, offer, comes up twice. It is both relevant for making people want to engage with your community, and a key part of having an impact.
Therefore, the question of what can EA offer this group should be at the forefront of any EA community builder’s plans.
If you do not have an offer you could develop one. This could mean doing research yourself or could be part of your outreach to work with keen individuals to develop an offer.
Case study: I struggled to engage UK civil servants with EA. When there was interest I struggled to direct it (for example, the official in charge of £13bn of overseas aid said he was interested in EA but didn’t see how to apply it in his work). This became easier after working with 80000 Hours to develop research on how civil servants could best use their carers to have an impact. This both attracted people and provided them with direction leading to career plan changes that will hopefully have a long run impact.
If you are planning how to outreach EA ideas to a group you roughly want to consider P and O as laid out in the chart here:
Signal interest. Build connections with individuals. Eg. finance.
Start senior & cascade down. Eg Civil servants. OR Go for mass outreach. Eg. Students
No clear offer
Reconsider outreach in this case.
Work with small groups and invite people to develop an offer with you. Eg. Quakers,
High propensity and developed offer.
On the whole people will recognise what you have is useful and be willing to listen and share it. Mass outreach methods may work. Alternatively, the most influential people in a group may be willing to help spread your message (although this is variable and depend on the individuals involved).
Case study: Building interest in an EA student society at LSE was relatively easy. Students at freshers’ fair were keen to sign-up to learn how they could have an impact with their career. We worked with 80K to run and advertise (Facebook ads) a career workshop. Nearly 300 students signed up and 80 attended.
Case study: After HIPE felt it had the beginnings of a useful offer for civil servants we carefully approached one of the UKs most senior civil servants. He liked the idea of supporting civil servants to think about the impact of their careers and offered to support us and cascade our message throughout the civil service.
Low propensity and developed offer.
It may be difficult to get people to listen to your ideas in this situation, but those that do can find it very useful. Outreaching may mean signalling that you are altruistic, relying on word of mouth, and putting in effort to work closely with any interested individuals. You may also want to look for pockets of more interested people with the group you are engaging with.
Case study: Corporate outreach in finance. Initial attempts to outreach effective giving in a large finance company giving went poorly. A lot of effort went in with very little interest. We hypothesised that most people in this finance company assessed their self-worth by the amount they earned and had little interest in doing good. That said one wealthy senior person showed interest and took the GWWC pledge. In future we want to focus more on reaching staff at less prestigious quant trading firms.
High propensity and under-developed offer.
People in this category want to do good will give a cursory look at what EA has to say but often see minimal benefit it adopting EA ideas into how they try to do good. It can be useful to work with the keenest individuals to see exactly what EA has to offer them. If done well this can create new resources and ideas for the EA movement to support people globally to do good.
Case study: A number of people approached me saying they wanted to improve the world and that even upon reflection, they believed that justice and equality were part of their moral systems. I looked for EA writing on this but found very little. I started a working group to decide for people with justice/equality ethics to decide where to give £1000. This got new people involved with EA ideas and hopefully created research others around the world will find useful.
Low propensity and under-developed offer.
I cannot think of any successful EA community building actions that appear to full into this category. (Although perhaps a no true Scotsman fallacy).
I, impact potential, which is how much can the people in this community could increase the impact of the EA movement by becoming a part of it.
Power. For example, wealth, political influence, skill, free time. This is the biggest factor.
Future power. Sometimes it may be reasonable to assume a group contains people who will be powerful further down the line.
Support or hinders needs of the community or wider EA movement. For example recruiting a lot of rich white men may not lead to a vibrant growing movement. Alternatively someone may have a skill that is not rare but is of use to others in your community.
Number of people in potential audience. A larger target group means more trying to do good. For EA London size has very rarely been a limiting factor unless we’re reaching out to a very niche group (<1000).
More powerful groups are also likely to be smaller and more elite and therefore harder to access. Overall it is still likely useful to reach out to powerful people, especially if you have a good offer and some reason to think you can access them.
It would be possible to turn the above formula into a rubric to compare community building efforts. For example, by quantifying O, P, A and I.
Case study: I am reasonably convinced it is high impact to reaching out to UK civil servants as we have a decent offer (career support), they have a propensity to accept EA ideas, they are accessible (to me) and they have a high potential for impact.
There are a number of risk to rapid movement growth. Best put across here. For example:
Loss of key ideas, such as cause neutrality and changing one’s mind.
Internal upheaval: Eg. if too popular may attract difficult to handle individuals.
Dilution: Not enough EAs to lead to good conversations etc. (Like the Eternal Sept effect)
The rate a movement grows is less important than the eventual shape of a movement (as explained here). Does does not mean we should not be spreading EA, just that it is importance a community builder is cautious, and aware of how their actions could be impacting the wider EA movement and ideas.
A community builder should:
be honest about what EA can offer and how helpful it is. Be aware that how you market or sell an idea can change that idea.
Be intellectually honest about the weirder parts of EA (Eg. AI risk). Presenting weirder ideas well requires more evidence, and we should be wary of rushing into them straight-away or if time is limited. However I would recommend against hiding these ideas. See further advice on this here [link pending].
Be cautious about rapid growth
consider prioritising actions to build understanding of core EA ideas within your community, above actions to grow the community.
Be cautious about reaching out to individuals who could be toxic to the movement.
Case study: A senior UK political figure writes stuff that suggests he could be interested in EA. However we held off reaching out, as this figure expressing support for EA could damage the reputation and the non-partisan nature of the movement.
Consider what the ideal EA movement would look like and build towards that.
Be aware if your group is very un-diverse. See paragraph below on diversity.
If developing what EA has to offer, focus on areas that add to EA, where the EA movement would clearly want to solve the question at hand at some point. Avoid areas that detract from EA.
Case study: I was strongly against a suggestion for an EA London sub-group on effective arts charities. I believe very few people who take a cause neutral approach to giving would want to focus on arts, and so such a group would primarily detract from EA ideas and not produce research useful to other EAs. I was in favour of groups focused on: justice, climate change and systemic change. These are plausible candidates for priority areas, members of these groups could be cause neutral, and work in these areas would contribute useful research to the wider EA movement.
Address problems as soon as they arise.
Acting promptly can stop problems escalating.
Do not be scared to ask for help from others. Talk to CEA if you think any action you or anyone in your community might take could be damaging to the wider movement. If unsure you can contact me at any point on firstname.lastname@example.org
Case study: An EA London attendee legally trademarked “effective altruism”. This was addressed with support from CEA. Perhaps flagging the issue sooner could have helped.
Don’t be a dick. Not only should you be a role modelling of doing good for your community but you should avoid bringing disrepute to the community.
Be careful about inviting controversial speakers or arguably unethical people.
Be careful how you present poverty. See: Guidelines on Depicting Poverty
Case study: There was a backlash against a “poverty simulation” EA event at an EA society. This hit the national news.
Momentum keeps a community functioning well (see section below on Momentum)
There may be first mover benefits of taking risky actions that others may do anyway and could end up doing worse.
We should remain open to new ideas. In particular:
We should be wary of saying “no this new idea could damage or negatively change the EA movement” as an excuse to ignore potentially useful constructive criticisms.
Toxic individuals or controversial speakers may have very useful things to say.
Case study: an EA London attendee said: “I wish I had created the EA politics Facebook group. I think it is a terrible thing that descends quickly into highly unconstructive partisan discussions, wastes time and helps no one be better at doing good. Had I created it first I could at least have heavily moderated it.”
The best way to get started building a new community is to have the community already exist in some form. Often there is a body of people who would benefit from a community and it just needs a bit of action to bring it together.
When creating communities like this it is helpful to have someone who is already part of the target group who is willing to run a community.
For example when THINK seeded new student groups this had a high failure rate, but where there was existing keen student(s) willing to run a group success was more likely.
Case study: The London EA finance group was started by people working in finance in London. To kickstart the group we pulled together a list of existing EAs who worked in finance (thinking of contacts we knew and inviting people form the EA London email list to let us know if they worked in this area). This meant that the initial events already had a sizable body of interesting engaged EAs, making it easier to develop events good enough for people in finance to bring colleagues to. The group was run by an EA who worked in finance.
Even one or two people who of the type you are looking to outreach too is really useful.
For example REG already had contacts within professional poker. The Future Generations APPG of UK Lords and MPs benefited from having Lord Martin Rees on board.
It feels like obvious advice but it is so rarely stated. If you want to run a community full-time you should be generally good at running things:
Be business smart. understand strategy and how to set targets, make plans, reach goals, etc.
Find good mentors and advisors to support you. Other EAs will help you!
Take a lean approach, try to prioritise learning.
The default state of a community is decay - people attend when something is new gradually will return less and less frequently. Communities need to retain momentum to keep people engaged. This means fostering:
Change. A community that grows, changes and has a stream of new people and new ideas remains interesting.
Sub-communities that long term attendees can join (such as a career network or being a member of GWWC).
Friendships. Connect people to others they’d get on well with. For example you could facilitate and encourage EA flatsharing, and so on.
Case study: I decided to host a dinner at my flat for friends in my area. 18 people showed up and we had a great time. I decided to make it a regular event, every other Monday. As time went on the numbers dropped. People were less excited by it, knew they could put off coming until another week. The numbers went down to 15 then 12 and are now at around 10 people.
It is the view of the author that conversations challenging people on their moral intuitions are often:
Off-putting. No one likes to be moralised. See article here.
Unproductive. These conversations do not change minds easily. See article here.
Damaging to the EA community. We want a community where people with different moral views can work together to do more good than they could individually.
Arrogant. Why are your moral intuitions better than anyone else's?
As a community organiser you may want to be aware if such conversations are happening frequently and potentially look to minimise them.
That said it is of course possible to create time and space for constructive conversations on ethics, intuitions and values. Furthermore giving people the tools and resources to do good effectively does obviously include supporting them to assess and understand their own moral intuitions and values.
Sometimes EA communities lack diversity, which can be off-putting to people from the under-represented groups.
Case study: About 15 minutes into an EA social I realised the room contained 8 men and no women. The first women who arrived commented that they felt out of place.
This has not generally been a huge problem in London. Unfortunately, I have yet to find practical solutions to this. I have weak intuitions that:
Diversity in public facing roles helps. When a person of colour created an event on Facebook we saw a few more people from minorities at the event.
Gatekeeping can help. The career focused sub-communicates (eg policy and finance) seem to be reasonably representative of the careers backgrounds of attendees (except in age, the groups are still predominantly young).
A good piece to read on this is: Making EA groups more welcoming.
You should expect that any large long-running EA community group to collect free-riders, people who take more than they contribute. Often this is not a large problem, just something to be aware of.
Case study: The young Fabians are welcoming left wing political group. Apparently, they were so friendly they started to attract a small body of people (largely guys) with limited social skills. Drawn in by the welcoming atmosphere they felt at home. However, this made the group off-putting to other people.
I have continued to ensure the London community stays welcoming and friendly. I have handled this issue by asking people to volunteer or supporting people to improve how they approach others or talk about effective altruism.
Gatekept events or whole communities tend to be higher quality and people are more likely to attend an event if they feel it will be relevant to them. Furthermore gatekeeping does not necessarily lead to less attendance at events, a gatekept event may well get higher turnout. (Of course this has to be weighed against the fact that gatekept events may be elitist and off-putting).
Case study: Our first event for people in finance attracted a lot more interest than expected. Wealthy financiers who would not have come to a standard effective altruism event turned up. Attendance numbers we similar to our average public social events (although more effort was put into advertising)
Gatekeeping is not an either/or decision, but there is a broad scale of how tightly or loosely gatekept each event will be. Here are some examples of gatekeeping form the minimal to the most severe:
Reactive gatekeeping: Turn away disruptive or aggressive individuals
Eg. an individual turns up who drinks far too much and is rude to others so you tell them not to come again
Eg. discourage an attended who you think is looking just to plug their own product form turning up to an event by sending a personal email warning them that we do not like people to push their own products at events.
Passive gatekeeping: only market to certain channels
Eg. Post on Facebook so only the more engaged people in your community who follow the Facebook group see the post
Eg. Not run an event join with a particular group as you worry that your own community will be inundated by people from this group and become undiverse.
Off-putting: Make it an effort to find out about or attend events. How off-putting an event is can be varied drastically.
Eg. charging for events or hosting in locations that have a significant travel time cost.
Eg. On event signups: If you want to come to this event please say in 150 words what your aims are and what you want to get out of the event.
Eg. Saying: if you are interested in this topic please email the following address. Then only inviting the people who emailed to attend the event or join the community.
Incidental entry requirements
Eg. event is in a school so only students can attend
Eg. You must be a GWWC member to attend this event.
Eg. You must be a past or current UK Civil Servant to attend this event / be on this email list.
Eg. Question on a form for people interested in finance events: please state which area of finance you work in? Asking for details can prevent people from falsely attending.
Eg. Saying: If you are interested in founding a charity please let us know by filling out this short form. Then only telling people who gave good responses that you will host an event (no one is knowingly rejected).
Eg. Saying: to attend EA Global please apply by answering these questions. We will pick people we think are interesting.
Invitee only. Possibly no-one else even knows about event or community.
Eg. Inviting the local leaders in animal rights advocacy to a networking event
For specific advice on running a large non-student local group in a city please part 3 of 4 of this write up [Link pending].
This will likely include things that that could be relevant to other community builders too. I hope to cover:
Monitoring and evaluation (attendance, pledges, etc)
Increasing awareness (advertising, bring-a-friend, Meetup, Eventbrite, etc)
Increasing engagement (follow up, books giveaways, sign-up forms, etc)
Changing behaviour (pledge drive, career support, etc)
Keeping in contact (email newsletters, Facebook, etc)
Responding to individuals (supporting individuals, supporting projects, buddy scheme, etc)
Event design (
Talking about EA (
Strategy (strategy meetings, streamlining tech, fundraising, community buy-in, etc)
For advice on getting started running an effective altruism group in a city. It can be really easy and take almost no time and effort. See:
There are also an abundance of guides on how to build a student community. These are at:
We should generally be very sceptical of the usefulness of advice on community building.
Academic evidence on this topic is poor. Studies of communities and movements are often not replicated or even replicable.
It is hard to generalise. What works for one community or at one point in time will not necessarily work for another community or at another point in time.
We should have a prior that providing useful advice is really difficult, and where it exists it is well known (similar to seeing an article offering “a sure-fire way of making money”.
There is substantial disagreement. People in the animal rights movement cannot agree on topics ranging from the large (is civil disruption good or bad) to the small (whether leaflets should show pictures of animal suffering).
The author of this piece does not have an academic background in communities or movements, meaning this is largely ‘armchair theorising’.
On the other-hand
If you are working in this area then you will need to make best guesses on what to do. Being more informed, by reading articles like this, is likely to help guide your intuition and decisions.
The more specific advice is to a particular situation the more useful it is for guiding decisions. So advice focused on “EA community building” could have useful content.
The author of this piece has been working on EA community building for 7 years, and has developed enough intuitions about what works that he feels it is useful to try to share them.
Thanks to everyone who’s work on EA London led to so many case studies. Especially Kit Harris for running the EA London finance events; Tildy, Sophie, Dan, Alex and Alice for EA policy work; and Holly and David for general EA London work. Also thanks to David Moss for proofreading.