Review of EA Global 2016

My name is Amy Labenz, and I was Ex­ec­u­tive Pro­ducer and Cu­ra­tor of EA Global 2016. On be­half of the team, I’m post­ing to provide the EA com­mu­nity with in­for­ma­tion about our:

  1. Goals for EA Global 2016

  2. De­ci­sions and lessons

  3. Eval­u­a­tion of EA Global 2016

  4. Plans for EA Global 2017

Summary

What was it? EA Global 2016 was the largest gath­er­ing of the effec­tive al­tru­ism com­mu­nity to date. Just over 1,000 at­ten­dees joined us for a two and a half day con­fer­ence at the Univer­sity of Cal­ifor­nia, Berkeley where we had 3 stages, 3 work­shop rooms, and 5 dis­cus­sion rooms. You can see the com­plete sched­ule here.

What were the goals? We set three goals for EA Global 2016: (1) to cre­ate con­nec­tions and fa­cil­i­tate co­or­di­na­tion among peo­ple in the effec­tive al­tru­ism com­mu­nity, (2) to cre­ate an op­por­tu­nity for peo­ple to im­merse them­selves in EA con­cepts and effec­tively com­mu­ni­cate com­plex ideas, and (3) to en­courage the de­vel­op­ment of the EA move­ment as an in­tel­lec­tual com­mu­nity.

How did it go? Over­all, we think EA Global 2016 was a suc­cess. In par­tic­u­lar, we were pleased with our abil­ity to en­courage new con­nec­tions be­tween our 1,000 at­ten­dees as well as the pos­i­tive feed­back that we re­ceived about the pro­gram, theme, and pro­duc­tion. We did, how­ever, make mis­takes, and re­ceived some nega­tive feed­back as a re­sult, par­tic­u­larly con­cern­ing per­sis­tent mar­ket­ing emails, late an­nounce­ment of the pro­gram and dates, late ad­mis­sions de­ci­sions, and last-minute pro­gram changes.

What will we change? Go­ing for­ward, I will tran­si­tion to man­ag­ing a ded­i­cated Events Team full-time and will start prepar­ing for EA Global 2017 much ear­lier. We in­tend to en­gage more with the com­mu­nity through­out the plan­ning pro­cess to get in­put on pro­gram­ming de­ci­sions and to help po­ten­tial at­ten­dees eval­u­ate the value of at­tend­ing. In ad­di­tion, as part of the CEA merger, we have re­al­lo­cated staff and have be­gun mak­ing im­prove­ments for next year—for ex­am­ple, with the new ea­global.org.

Goals for EA Global 2016

We se­lected three main goals, based on an anal­y­sis of how EA Global, EAGx, and re­lated events have cre­ated value in the past:

  1. To cre­ate con­nec­tions and fa­cil­i­tate co­or­di­na­tion among peo­ple in the effec­tive al­tru­ism com­mu­nity,

  2. To cre­ate an op­por­tu­nity for peo­ple to im­merse them­selves in EA con­cepts and effec­tively com­mu­ni­cate com­plex ideas, and

  3. To en­courage the de­vel­op­ment of the EA move­ment as an in­tel­lec­tual com­mu­nity.

The first two goals were easy to agree upon, but we were more cau­tious in de­cid­ing the third. Goal #3 re­flects our view that the EA move­ment should cul­ti­vate an at­mo­sphere of in­tel­lec­tual mod­esty and self-skep­ti­cism. Our rea­sons for this view are: (1) our ex­pec­ta­tion that new high-pri­or­ity cause ar­eas will con­tinue to emerge and that a com­mu­nity-wide fo­cus on cause pri­ori­ti­za­tion re­mains im­por­tant, and (2) feed­back from com­mu­nity mem­bers that pre­vi­ous events had been too fo­cused on spe­cific causes.

De­ci­sions and Lessons

Our goals in­fluenced our de­ci­sion-mak­ing pro­cess, par­tic­u­larly on strate­gic ques­tions.

Programming

To help build con­nec­tions be­tween at­ten­dees, the sched­ule had many long breaks to al­low peo­ple to min­gle, as well as re­cep­tions and an af­ter­party to fa­cil­i­tate net­work­ing. To pro­mote im­mer­sive in­for­ma­tion ex­change, we had a large num­ber of work­shops and dis­cus­sion groups on par­tic­u­lar topic ar­eas. We recorded all of the talks on the two main stages so that at­ten­dees were able to par­ti­ci­pate in work­shops and dis­cus­sions and did not feel rushed to leave valuable con­ver­sa­tions. In ad­di­tion, our talks and pan­els in­ves­ti­gated re­search challenges and po­ten­tial break­throughs rather than par­tic­u­lar or­ga­ni­za­tions. We se­lected top­ics to at­tract ex­perts who could dis­cuss is­sues such as how EAs can in­fluence policy and how to nav­i­gate in­tel­lec­tual dis­agree­ments. Fi­nally, we in­tro­duced the Effec­tive Altru­ism Global Re­search Meet­ing, which in­cluded light­ning talks and aca­demic poster pre­sen­ta­tions to pro­mote in­tel­lec­tual ex­change and a col­lab­o­ra­tive at­mo­sphere.

Our main pro­gram­ming mis­steps were: (1) we did not post the pro­gram on the web­site early enough and failed to ad­e­quately com­mu­ni­cate sched­ule and pro­gram­ming changes at the event, (2) the large num­ber of con­cur­rent ac­tivi­ties limited the num­ber of shared ex­pe­riences at the con­fer­ence, (3) the work­shop rooms were not large enough, so some peo­ple who were ex­cited to par­ti­ci­pate in the most pop­u­lar work­shops were dis­ap­pointed, and (4) a num­ber of the recorded talks and pan­els had poor at­ten­dance, which was un­com­fortable for the speak­ers. We may want to con­sider hav­ing less com­plex­ity in the fu­ture, with fewer stages or fewer talks and more work­shops with larger work­shop spaces. We should also bet­ter com­mu­ni­cate which talks will be recorded, and may con­sider record­ing ad­di­tional or more varied con­tent.

Pro­duc­tion and Logistics

Based on feed­back from pre­vi­ous EA Global events, we tried to im­prove the pro­duc­tion and lo­gis­tics for EA Global 2016 so that peo­ple could bet­ter fo­cus on the con­tent of the event. This year, we spent more time work­ing on room lay­out de­sign, sig­nage to di­rect peo­ple, cater­ing es­ti­mates to en­sure that there would be suffi­cient food, and vol­un­teer re­cruit­ment, role de­scrip­tions, and train­ing so that we could an­ti­ci­pate and re­solve is­sues. In gen­eral, we made large im­prove­ments over pre­vi­ous years, but we still had is­sues with poor on-site com­mu­ni­ca­tion of sched­ule changes that re­sulted in con­fu­sion and frus­tra­tion for at­ten­dees.

Date and Venue

We de­cided to hold EA Global in Au­gust be­cause: (1) it is a con­ve­nient time for stu­dents to travel, (2) uni­ver­sity fa­cil­ities and hous­ing are more read­ily available, and (3) ven­dors lower prices to com­pete for busi­ness dur­ing their off-sea­son. Given our goal of fos­ter­ing an in­tel­lec­tual com­mu­nity, we thought that these benefits out­weighed the fact that some high-value at­ten­dees, such as Bay Area VCs, might be less likely to at­tend at this time. Peo­ple were gen­er­ally happy with the venue and re­ported that its phys­i­cal lay­out was one of the things they would like to see again. This lay­out, which was a large im­prove­ment over the sprawl­ing Google venue from 2015, was one of the other main con­sid­er­a­tions that led us to se­lect UC Berkeley. In ad­di­tion, work­ing with the EAs of Berkeley al­lowed us to get large dis­counts. One challenge of work­ing with the stu­dent or­ga­ni­za­tion was that it in­tro­duced ma­jor bu­reau­cratic hur­dles, par­tic­u­larly with fi­nal­iz­ing con­tracts with the uni­ver­sity and re­serv­ing rooms. This caused us to de­lay our an­nounce­ment of the event for months, which prob­a­bly pre­vented some peo­ple from at­tend­ing, and re­sulted in dis­rup­tive room changes, such as the last-minute changes to the work­shop rooms on Fri­day.

Diversity

We’d like EA Global to help fur­ther di­ver­sity within the move­ment by at­tract­ing both at­ten­dees and speak­ers from un­der­rep­re­sented groups, par­tic­u­larly in terms of na­tion­al­ity, race, and gen­der. To re­cruit speak­ers, we con­sulted with the CEA team and lead­ers in the com­mu­nity to cre­ate a mas­ter list, which we eval­u­ated on a num­ber of di­men­sions. We op­ti­mized for fit with the pro­gram, se­lected ex­perts in fields that we be­lieve to be high value, and pri­ori­tized speak­ers based on their fa­mil­iar­ity with effec­tive al­tru­ism, their name recog­ni­tion, and our per­sonal con­nec­tion. As part of that pro­cess, we made spe­cific efforts to in­vite peo­ple from tra­di­tion­ally un­der­rep­re­sented groups. In ad­di­tion, when build­ing the pro­gram we made sure there were no all-male pan­els (we did have one panel en­tirely com­posed of women—this was not by de­sign, but we no­ticed it with time to make ad­just­ments and de­cided not to). We made sig­nifi­cant im­prove­ments to our pro­cess over EA Global 2015; how­ever, we’re not satis­fied with our progress here and would like to do bet­ter in the fu­ture. We hope that start­ing the speaker se­lec­tion pro­cess ear­lier will give us more op­por­tu­nity to find speak­ers who bet­ter rep­re­sent the di­ver­sity that we’d like to see in the move­ment.

Speakers

Over­all, we did not op­ti­mize for at­tract­ing “big name” speak­ers, though we did in­vite some in that cat­e­gory. We found in the past that head­line-mak­ing speak­ers can re­sult in in­ac­cu­rate per­cep­tions out­side of the com­mu­nity. For ex­am­ple, at EA Global 2015, Elon Musk par­ti­ci­pated in an AI side event, but be­cause of his sta­tus and the team’s de­ci­sion to fea­ture him promi­nently in some event mar­ket­ing, he was seen as the face of EA Global 2015. This con­tributed heav­ily to the im­pres­sion that EA Global 2105 was fo­cussed on tech­nol­ogy and AI risk, even though this was not the in­ten­tion of the or­ga­niz­ing team. With more lead time and care­ful pre­sen­ta­tion of fea­tured speak­ers and top­ics, we think it will be pos­si­ble to se­lect for “big name” speak­ers that also fit well with con­fer­ence themes. This year, a num­ber of speak­ers in this cat­e­gory were in­ter­ested but had schedul­ing con­flicts. Next year we will start the in­vi­ta­tion pro­cess ear­lier. If you would like to nom­i­nate a speaker, please do so here.

This year we had ~80 speak­ers. Ali­son Wood­man, our Speaker Li­ai­son Team Leader, man­aged speaker com­mu­ni­ca­tions be­fore the event and helped train vol­un­teers on the Speaker Li­ai­son Team. Each speaker had a Speaker Li­ai­son to es­cort them to their stages and to make sure they were com­fortable and un­der­stood the event lay­out. We also hosted a speaker re­cep­tion where we con­nected speak­ers with one an­other and with EA or­ga­ni­za­tions, and we ex­pect to see friend­ships and col­lab­o­ra­tions as a re­sult. Gen­er­ally, our speaker in­ter­ac­tions went well, and we have re­ceived pos­i­tive feed­back from many speak­ers. Un­for­tu­nately, we had com­mu­ni­ca­tion de­lays, par­tic­u­larly about panel de­scrip­tions and light­ning talk ac­cep­tances, which were likely stress­ful for the speak­ers. This was largely caused by a pro­cess mis­take where we had only one em­ployee draft­ing the de­scrip­tions and mak­ing de­ci­sions about light­ning talks. Once we in­vited ad­di­tional team­mates to ap­prove talk ap­pli­ca­tions and write panel de­scrip­tions, things went more smoothly, but be­cause of de­lays in mak­ing that call we prob­a­bly ap­peared less pro­fes­sional than we would have liked.

Ad­mis­sions and Marketing

We have an ad­mis­sions pro­cess be­cause we want to se­lect for peo­ple that will get the most value out of the con­fer­ence and who will con­tribute to the pos­i­tive ex­pe­riences of other at­ten­dees. We look for al­tru­is­tic peo­ple who are cu­ri­ous and have an an­a­lytic mind­set and try to screen out peo­ple who have his­to­ries of dis­rup­tive be­hav­ior. This year we re­ceived 2,152 ap­pli­ca­tions for a tar­get con­fer­ence size of 1,000 peo­ple, which we suc­cess­fully achieved. The ap­pli­ca­tion pro­cess, how­ever, did have sev­eral prob­lems: (1) con­ver­sion on our ap­pli­ca­tions was lower than ex­pected, and as a re­sult, we did a push for ad­di­tional ap­pli­ca­tions that was seen as overly per­sis­tent or mis­lead­ing (See Kerry Vaughan’s Mar­ket­ing post), (2) spel­ling mis­takes made some of the pre-event com­mu­ni­ca­tions ap­pear un­pro­fes­sional, (3) some ques­tions in the ap­pli­ca­tion may have been off-putting or dis­cour­ag­ing, and (4) pro­mo­tions, such as al­low­ing some at­ten­dees to bring a friend, cre­ated in­con­sis­tent ad­mis­sions stan­dards. We plan to ad­dress these is­sues by open­ing ap­pli­ca­tions ear­lier and mak­ing more use of wait­list­ing next year. This will al­low for re­con­sid­er­a­tion of those who ex­pressed early in­ter­est and will give us more time to build pro­mo­tions that are con­sis­tent with our ad­mis­sions stan­dards. We also plan to have a fo­cus group led by our Com­mu­nity Direc­tor, Ju­lia Wise, in or­der to check the ap­pli­ca­tion for tone. Ju­lia will also re­view the pre-event com­mu­ni­ca­tions and mar­ket­ing plan to make sure they do not clash with the in­tel­lec­tual and pro­fes­sional goals of the event.

Location

Our biggest con­sid­er­a­tions when de­ter­min­ing where to hold EA Global 2016 were the con­ve­nience for and cost to the com­mu­nity. We con­sid­ered (1) where most EAs that would like to at­tend are lo­cated (based on the map of Effec­tive Altru­ists on the EA Hub, and the re­sults of the most re­cent EA sur­vey), (2) the strength of the ex­ist­ing com­mu­ni­ties in the var­i­ous lo­ca­tions, (3) the cost of flights for the EA com­mu­nity, and (4) the lo­ca­tion of EA or­ga­ni­za­tions. Based on our anal­y­sis, we de­cided on the Bay Area. Our de­layed re­lease of the lo­ca­tion likely pre­vented a num­ber of Europe-based EAs from mak­ing the trip, and we plan to make a de­ci­sion much ear­lier this year. Please let us know your lo­ca­tion prefer­ences here.

Food

Food was the largest sin­gle bud­get item for the con­fer­ence; we spent ap­prox­i­mately $100 /​ per­son for the week­end. We de­cided to have food on-site for the open­ing and clos­ing re­cep­tions and for lunches be­cause we wanted to cre­ate op­por­tu­ni­ties for in­ter­ac­tion be­tween con­fer­ence go­ers. After dis­cus­sions in­ter­nally and on Face­book, we chose a veg­e­tar­ian and ve­gan menu. We did this ac­knowl­edg­ing that the com­mu­nity con­tains many who do not be­lieve that per­sonal diet changes are an effec­tive way to re­duce suffer­ing, and also many who be­lieve that a meatless menu is a ba­sic ges­ture of re­spect to an­i­mal ad­vo­cates. We in­cluded a restau­rant guide in the pro­gram with the hope that nearby restau­rants could provide fur­ther op­tions if peo­ple found that our cater­ing didn’t meet their needs. Food is always a touchy sub­ject at con­fer­ences, and our sur­vey re­sults showed mixed re­sponses on the meat ques­tion but trended pos­i­tive in terms of over­all food satis­fac­tion.

Finances

CEA lead­er­ship set three high-level fi­nan­cial goals for the event: (1) use re­sources wisely, in keep­ing with our com­mu­nity val­ues, while en­sur­ing that at­ten­dees could fo­cus on the con­tent in a pleas­ant, stress-free en­vi­ron­ment, (2) keep tick­ets as af­ford­able as pos­si­ble while mak­ing schol­ar­ships available to ap­pli­cants who could not have at­tended with­out as­sis­tance, and (3) roughly break even on the event. In the end, we were quite suc­cess­ful in achiev­ing these goals. The to­tal bud­get for the event was ~$250K (ex­clud­ing the cost of staff time and travel). We es­ti­mate that CEA will spend $20K on costs not cov­ered by ticket pur­chases and spon­sor­ships (~$230K). We are still wait­ing on a full ac­count­ing from our book­keeper and will provide an up­date when we have the fi­nal num­bers. Here is a rough sum­mary:

Note: fi­nan­cial aid and schol­ar­ships were our largest source of fore­gone rev­enue, how­ever, we did not ac­count for them as a “cost” in our spend­ing overview. Stu­dents were given au­to­matic dis­counts of $100 off, and ev­ery­one who ap­plied for fi­nan­cial aid re­ceived some fi­nan­cial as­sis­tance. We heard sub­stan­tial nega­tive feed­back about the posted ticket price, which sug­gests that we could have bet­ter com­mu­ni­cated our fi­nan­cial aid policy. We in­sti­tuted a pric­ing strat­egy where the ticket price was con­sid­er­ably higher than the price we ex­pected the av­er­age at­tendee to pay, based on our es­ti­mates of the num­ber of peo­ple who would need schol­ar­ships. It seems that the cur­rent pric­ing strat­egy might dis­cour­age value-al­igned peo­ple who need fi­nan­cial as­sis­tance but are hes­i­tant to ask for it. We plan to con­sider al­ter­na­tive pric­ing strate­gies and are open to sug­ges­tions. We will also con­sider adding more fi­nan­cial aid to our bud­get for next year.

Eval­u­a­tion of EA Global 2016

EA Global takes thou­sands of hours of EA time, so it is im­por­tant that we earnestly as­sess how we did. We col­lected data from two sur­veys: (1) Clos­ing Sur­vey: 255 peo­ple re­sponded to a 5-minute sur­vey af­ter the clos­ing talk, (2) Fol­low-up Sur­vey: 92 peo­ple re­sponded to a sur­vey that we sent a week af­ter the event. The sur­veys were a mix of free-form re­sponses and nu­mer­i­cal rat­ings. (See the links for sum­maries and Oliver Habryka’s up­com­ing post for more de­tail).

On our Clos­ing Sur­vey, at­tendee satis­fac­tion with the con­fer­ence over­all got the high­est rat­ing, with a score of 6.1 out of 7 (dis­played as 2.1 be­low). Similarly, our Fol­low-up Sur­vey re­sults showed a 4.1 out of 5 over­all satis­fac­tion rat­ing.

In our Fol­low-up Sur­vey, we asked at­ten­dees to rate how much differ­ent parts of the event were is­sues for them. The mean was 1.5 on a 4 point scale, sug­gest­ing that none of these were ter­ribly both­er­some on av­er­age; how­ever, of the is­sues listed, the clar­ity of the pro­gram, lo­gis­tics com­mu­ni­ca­tion, and the ticket price ap­pear to have been the biggest con­cerns for at­ten­dees. This matches the feed­back that we have re­ceived in-per­son and on­line.

In our Clos­ing Sur­vey, we saw a sig­nifi­cant im­pact as mea­sured by the num­ber of at­ten­dees who re­ported hav­ing changed their minds on par­tic­u­lar top­ics. A large ma­jor­ity of peo­ple sur­veyed listed at least one area where they changed their minds (though this re­sult may have been a con­se­quence of the sur­vey de­sign).

In ad­di­tion, in re­sponse to our Fol­low-up Sur­vey, 14% of the ~90 re­spon­dents said that they changed their life plans sig­nifi­cantly as a re­sult of EA Global, 27% said that EA Global changed how much they in­tend to give to effec­tive char­i­ties, and 5% of peo­ple re­ported that they took the Giv­ing What We Can Pledge as a re­sult of EA Global. We can­not sim­ply ex­trap­o­late this pro­por­tion to all at­ten­dees of the event since there are likely se­lec­tion effects at work (those re­spond­ing to the Fol­low-up Sur­vey may have been more moved by the event). It ap­pears that 23 at­ten­dees signed the GWWC Pledge at EA Global or in the weeks af­ter the event and 8 more signed up for Try Giv­ing (only 3 cited the con­fer­ence as how they first heard about GWWC, though this re­quired a write-in an­swer, which may have re­sulted in un­der­re­port­ing).

We ex­pect to see ad­di­tional sources of value in the months to come, but based on at­tendee and EA or­ga­ni­za­tion feed­back, as well as our own con­ver­sa­tions with po­ten­tial col­lab­o­ra­tors, po­ten­tial hires, and new donors, we think that EA Global 2016 was a valuable use of CEA team time. On the mar­gin, we might have saved time by out­sourc­ing more of the work, or by limit­ing the com­plex­ity and/​or size of the event. Given that we re­ceived feed­back sug­gest­ing that peo­ple would like to see more fa­cil­i­tated in­ter­ac­tions & net­work­ing and rel­a­tively fewer pan­els & in­ter­views, CEA might be able to plan a ver­sion of EA Global with less cu­rated con­tent and more work­shops, which would de­mand less staff time while bet­ter satis­fy­ing at­tendee prefer­ences.

The Com­mu­ni­ca­tions and Outreach Team is cur­rently work­ing on im­proved meth­ods for eval­u­at­ing CEA pro­jects against one an­other, but this is be­yond the scope of this par­tic­u­lar post.

Plans for EA Global 2017

For EA Global 2017, we will start the plan­ning pro­cess ear­lier and in­vite more com­mu­nity par­ti­ci­pa­tion through­out. We aim to start plan­ning EA Global 2017 this month, with a goal of se­lect­ing a lo­ca­tion and iden­ti­fy­ing keynote speak­ers in the next month, and work­ing from there to se­lect a date. If you are in­ter­ested in con­tribut­ing con­tent or in hav­ing your or­ga­ni­za­tion rep­re­sented at the event, please use our pre­limi­nary work­shop sub­mis­sion form here.

To build on the suc­cess of EA Global 2016 and shore up our weak spots, I am tran­si­tion­ing to run­ning the Events Team full-time, which will in­clude EA Global, EAGx, and ad­di­tional in-per­son out­reach. The team will in­clude Rox­anne He­ston, Ju­lia Wise, and an­other team­mate (de­scrip­tion here, ap­ply here), and will re­ceive sup­port from the rest of the Com­mu­ni­ca­tions and Outreach Team, led by Tara Mac Au­lay. We will bring on a new team­mate to man­age CEA US Oper­a­tions (de­scrip­tion here, ap­ply here). As a re­sult of the new team struc­ture, we have already started mak­ing im­prove­ments to our pro­cess for next year. For ex­am­ple, we have a new web­site, thanks to Sam Deere. With the new sys­tem, many staff mem­bers can edit the web­site and push changes quickly. This will al­low us to share in­for­ma­tion about EA Global, in­clud­ing in­for­ma­tion about the pro­gram, speak­ers, and sched­ule, with­out hav­ing a sin­gle bot­tle­neck.

Please visit the new web­site to see the videos and pic­tures from this year, and to get in­for­ma­tion about EAGx events and EA Global 2017. I ap­pre­ci­ate all of the feed­back that we have got­ten so far and I look for­ward to EA Global 2017! If you have feed­back but haven’t had a chance to share it yet, please do so in the com­ments or email us at hello@ea­global.org.

Thank you!

Thank you to the at­ten­dees, speak­ers, staff, and vol­un­teers who helped make EA Global 2016 pos­si­ble. Spe­cial thanks to: Oliver Habryka (Direc­tor of Pro­gram­ming) and Ju­lia Wise (Com­mu­nity Direc­tor) for help with this post. Thanks also to the rest of the EA Global team: Ali­son Wood­man (Speaker Li­ai­son Team Leader), Peter Buck­ley (As­sis­tant Pro­ducer), Kerry Vaughan (Direc­tor of Mar­ket­ing), Larissa Rowe (Mar­ket­ing and So­cial Me­dia), Tara Mac Au­lay (COO of CEA), Will MacAskill (CEO of CEA), An­jali Gopal (Vol­un­teer Team Leader), and our 50 vol­un­teers. Also, thank you to our MCs: Nathan Labenz, Rox­anne He­ston, and Jonathan Court­ney, and to Tessa Alex­a­nian for helping me and oth­ers with our slides, An­drew Lap­in­ski-Barker for cre­at­ing the signs for the event, Kather­ine Xiang for de­sign­ing the pro­grams, Michelle Hutch­in­son for or­ga­niz­ing the Re­search Meet­ing, and the many Pareto Fel­lows who joined in to help.