General lessons on how to build EA communities. Lessons from a full-time movement builder, part 2 of 4

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:

  1. We can con­sider EA as: EA ideas, the global EA move­ment, and smaller EA com­mu­nity groups.

  2. EA com­mu­ni­ties have an im­pact pri­mar­ily through growth (spread­ing and re­in­forc­ing EA ideas), re­ten­tion (keep­ing EAs en­gaged) and net­works (mak­ing con­nec­tions be­tween EAs).

  3. 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)

  4. 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.

  5. 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:

Devel­oped offer

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.
Do re­search to de­velop an offer.

Work with small groups and in­vite peo­ple to de­velop an offer with you. Eg. Quak­ers,

Low propensity

High propensity

I also provide
fur­ther ad­vice on other as­pects of com­mu­nity build­ing. In­clud­ing:

  1. 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.

  2. The eas­iest way to start a com­mu­nity is to build upon ex­ist­ing con­nec­tions within the groups you want to out­reach to.

  3. 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


Break­ing EA into: EA ideas EA move­ment and EA communities

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


The full im­pact of an EA community

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.

Other impacts

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.


The benefit of com­mu­nity growth – a model

The model

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.

A, Ac­cess – lev­er­age po­ten­tial com­mu­nity members

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, Propen­sity for EA

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.

O, offer – Re­in­forc­ing the im­por­tance of hav­ing a good offer

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.

Also see: What is Valuable About Effec­tive Altruism

P and O gives you a strat­egy to grow a fledgling EA community

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:

Devel­oped offer

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.
Do re­search to de­velop an offer.

Work with small groups and in­vite peo­ple to de­velop an offer with you. Eg. Quak­ers,

Low propensity

High propensity

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

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.

Im­pact Growth – Con­sid­er­ing to­tal impact

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.


Risks and caution


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.

Tips for the cau­tious move­ment builder

A com­mu­nity builder should:

  1. Be hon­est.

    • 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].

  1. 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.

  1. 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.

  1. 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 sam@ealon­

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.

  1. Be eth­i­cal.

    • 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.

Against caution

  • 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.


Fur­ther thoughts and advice

Get­ting started – con­sider who you already have on board.

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.

Be busi­ness smart

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.

Be­ing welcoming

Be wary of con­ver­sa­tions that try to change values

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.


For ex­am­ple read: Sup­port­ive Scep­ti­cism and 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

  • En­try requirements

    • 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.

  • Ap­pli­ca­tion form

    • 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


Fur­ther read­ing:

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:

How I or­ganise a grow­ing effec­tive al­tru­ism group in a big city in less than 30 min­utes a month.

There are also an abun­dance of guides on how to build a stu­dent com­mu­nity. Th­ese are at:

Fur­ther ad­vice on fresh­ers fairs/​tabling is here, here, here and here. There is also a List of EA Pre­sen­ta­tions, an EA Pitch Guide, Tips on talk­ing about EA, and more Re­sources and Guides here.


An­nex: what can we use­fully say about com­mu­nity build­ing?

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.