I think the cost per hour of engagement is a good intuitive metric to assess the cost-effectiveness of running the EA Forum. From footnote 1, the daily cost to run the EA Forum is 1.48 k$ (= 1.3*10^6*2.5/6/365.25). There have been around 240 hours of engagement per day over the last 6 months or so. So the cost per hour of engagement has been roughly 6.17 $ (= 1.48*10^3/240). I suspect the engagement time would drop significantly if users had to pay 6.17 $ per hour they spend on the EA Forum. I believe this suggests the marginal cost-effectiveness of running the EA Forum is negative (trusting my guess for the users' revealed preferences), and that there should be a reduction in the time spent running the EA Forum.
The iteration of September 2023 of The Introductory EA Program had 1.58 attendances per hour spent running the program. For 30 $ per hour spent running the program, and 4 h of engagement per attendance (3 h of preparation, plus 1 h of discussion), there would be 6.32 h (= 1.58*4) of engagement per hour spent running the program, and the cost per hour of engagement would be 4.75 $ (= 30/6.32), 77.0 % (= 4.75/6.17) as much as for the EA Forum. So, considering uncertainty, it looks like the iteration of September 2023 of The Introductory EA Program had a cost per hour of engagement similar to that of the EA Forum nowadays.
You can try to make comparisons with other programs, or estimate which work you are doing which generates the most engagement time per $ (although figuring out the additional engagement time a given activity generated is tricky; for example, running an EA Forum event will of course cause engagement with the posts and comments related to the event, but will tend to decrease the engagement with other posts and comments).
Hmm, I think thereās some sense to your calculation (and thus I appreciate you doing+showing this calculation), but the $6.17 conclusionāspecifically, āengagement time would drop significantly if users had to pay 6.17 $ per hour they spend on the EA Forum, which suggests the marginal cost-effectiveness of running the EA Forum is negativeāāstrikes me as incorrect.
What matters is by how much engaging with the Forum raises altruistic impact, which, insofar as this impact can be quantified in dollars, is far,Ā far higher than what one would be willing and able to pay out of oneās own pocket to use the Forum. @NunoSempere once estimated the (altruistic) value of the average EA project to be between 10 and 500 million dollars (see cell C4 ofĀ this spreadsheet;Ā hereās the accompanying post). That is far higher than the actual dollar cost of running the average project. (Indeed, if one is funded by EA money, then oneās generation of altruistic dollars needs to outpace oneās consumption of actual dollarsāand by a large multiplier, if one is to meet the funding bar.)
Going back to NuƱoās spreadsheet: If I make the arrogant assumption that Iām within an order of magnitude of Ben Todd, impact-wise, then that means my lifetime impact is at least 10 million dollars. Assuming linearity (which isnāt a great assumption, but letās go with it for now) and a career length of 40 years, this means my impact over the past 4 years has been ā„1 million dollars.[1]Ā In that time, Iāve spent maybe 500 hours on the EA Forum.[2]Ā Meanwhile, Iād say that the Forum has contributed greatly to my intellectual development, i.e., added at least 20% to my impact. (The true percentage may in fact be much higher, because ofĀ crucial considerations that the Forum has helped me orient toward, but letās lowball things at 20%, for now.) This would imply that my impact has been amplified by at least $200,000/(500 hours) =Ā $400 per hour spent on the Forum. ([Insert usual caveats about there being large error bars.]) Contrast with your $6.17.
(I did this calculation on myself not because Iām special, but because I know what the numbers are for myself. Iād guess that the per-hour bottom line for other Forum users would be ~similar.)
We can now go one step further, and estimate the Forumās āaltruistic dollar generated per actual dollar spentā multiplier to be at least 400/6.17 ā 65. Embarrassingly, I donāt know how this compares against todayās funding bar,[3]Ā but seems very plausible to me that itās above.
(Nonetheless, people may still not pay $6.17/hour to use the Forum because $6.17/hour is a non-trivial cost considering peopleās actual incomes. Additionally, people are just used to being able to browse the internet for free, and so I suspect many wouldnāt do the expected value calculations and reach the ārationalā conclusion that they should in fact pay.)
That is, 500 hours consuming the Forumās content. Iāve also spent time writing on the Forum, but if we model the Forum as a two-way market, with writers and consumers, and say that itās the consumers who benefit from being here, then it doesnāt make sense to include my writing time. (Alsoāand perhaps more relevantlyāI donāt think writing time gets counted by the Forumās analytics engine as engagement time if itās spent mostly in a Google doc.)
Further detail: What really matters is what the multiplier is on the margin (i.e., what it is for the last dollar being spent on a project), rather than what it is for the project as a whole.
Thanks for sharing your thoughts, Will! I agree what matters is the additional altruistic impact caused by engaging with the Forum. However, I think my point holds as long as people have accurate views about how to maximise their altruistic impact.
For example, if you believed "factual impact of your marginal hour on the Forum" - "counterfactual impact of this hour" < "impact of donating 100 $[1]Ā to the organisation or project you consider the most cost-effective", and using the Forum costed 100 $/h, I think you would have a greater altruistic impact by your own lights by spending less time on the Forum, and donating the savings. Do you agree?
Analogously, if the user spending the marginal hour on the Forum believed "factual impact of their marginal hour on the Forum" - "counterfactual impact of this hour" < "impact of donating 6.17 $ to the organisations or projects they consider the most cost-effective", and using the Forum costed 6.17 $, I think they would have a greater altruistic impact by their own lights by spending less time on the Forum, and donating the savings. In this case, if the marginal user-hour costed 6.17 $ to the Forum team[2], I believe they would also increase altruistic impact in the eyes of the user of the marginal hour by spending less time generating engagement on the Forum, and donating the savings to the organisations or projects that user considers the most cost-effective.
(Toby's title change was basically just the two of us trying to figure out how to better communicate his role to external people, it's not related to his level. He has a fair amount of autonomy in his role, and I thought "Content Manager" didn't properly reflect that. Also I personally could never remember which of "Content Manager" and "Content Specialist" was the more senior title... š )
As a reminder, you can view our teamās half-quarterly OKRs viaĀ this public doc that I keep updated. I recently added our Q2.2 plans (May 20 - July 1).
Neutral votes (i.e. adding a reaction) no longer count toward the vote total. Previously, this was leading to people assuming they were downvoted more than in reality, like if you saw that a comment had 3 votes but only 1 karma. Oftentimes people will react (such as agree or disagree) without voting on the karma.
Thanks for this helpful update. I assume it applies to both comments and posts.
Since then, our usage metrics have been pretty stable, nice!
Nitpick. You included a graph of the monthly active users at the start of the post, but I tend to think the daily hours of engagement are a better metric. It accounts for both the number of active users, and the hours of engagement per active user. In any case, the daily hours of engagement have also stabilised since the middle of October.
The "Copy image" feature below is not working for me. It copies to the clipboard the link of the dashboard instead of the image.
I used AI to fix transcription errors, rerrarange the ideas, and suggest tweaks to the title and some sentences.
Three of the most exciting projects to come out of EA in recent years are, in a vague sense, CEA spinouts:
* Kairos is directly a spinout of CEA and now handles most support for university AI safety groups. Basically everyone I've found who knows them is really excited about what they do
* NEST is an opinionated ideas-fi...
TLDR; To help the effective animal advocacy movement cost-effectively absorb greater amounts of funding in the near future, we are seeking expressions of interest from people who could found a new organization focused on:
* Highly neglected animals: insects, wild animals, shrimp, fish, etc, or
* AI and animals: AI alignment and governance for animal welfare, strategic actions considering transformative AI, AI for wild animals, etc.
* ...
In January, IĀ wrote about the EA Forum Teamās shift in direction. Since then, our usage metrics have been pretty stable, nice! š (Compare to 2024 where the metrics were going steadily downward the whole year.)
When I wrote that post, my role was focused on the EA Forum, and I had intended to spend the majority of my effort on the EA Forum. However, my role and my team are in flux right now, primarily in order to better align withĀ CEAās strategy. Iām currently running the CEA Online Team, and so far in 2025 we've broadly been spending a smaller portion of our resources on the EA Forum over time[1]. I think we need to be careful, because I still believe strongly in what I wrote inĀ my previous post, and I have a worry that reducing the resources we spend on the EA Forum does risk losing most of its value[2]. But there are always tradeoffs, and hopefully we can figure out the right balance between working on the EA Forum and working on other projects[3]Ā in service of CEAās overall strategy.
I wanted to include this here because, if we do end up spending less capacity on the EA Forum, you may see fewer updates from us. As always, Iām happy to hear feedback and thoughts from the community.
Site improvements
Updated notifications
To summarize WillāsĀ post from March, weāve updated a lot of the default behavior for Forum notifications (both on-site and email), primarily to better match user expectations from using other sites.
We encourage you to check yourĀ account settings and update your notification preferences! š
Draft comments
As Will wrote about last week, we've implemented a simple feature to let users save comments and quick takes as a draft.
Often, Forum comments and quick takes are quite long and substantive, and can be even more valuable than posts. So we hope that this makes them easier to write, especially in the case when you start one on your phone and want to add the finishing touches on your laptop later (as happens to me).
Users can run their own polls
We've made our debate week poll widget available for all users! You canĀ run your own poll by adding it to a post, comment, or quick take.
Tweaks to voting
We made a couple small but hopefully useful changes:
Neutral votes (i.e. adding a reaction) no longer count toward the vote total. Previously, this was leading to people assuming they were downvoted more than in reality, like if you saw that a comment had 3 votes but only 1 karma. Oftentimes people will react (such as agree or disagree) without voting on the karma.
WeĀ changed the text describing karma voting from āHow much do you like this overall?ā to āIs this a valuable contribution?ā.
Redesigned user menu
WeāveĀ updated the design of the user menu in our site header:
New āMade me laughā react
We added a new react to the family, by popular demand. š
Forum events
Forum events calendar
Toby set up a Google Calendar to help you keep track of Forum events:
The Forum tends to focus on pressing problems in the world, and it's nice to highlight some good news every once in a while. In January, we celebrated some of the ways the world is getting better.
Draft Amnesty Week
We ran another Draft Amnesty Week at the end of February, which is an event where we encourage people to publish posts that are sitting in their drafts and may never get finished. This time we got more posts than before, and more posts >50 karma. :) Check out the posts here!
@Toby Tremlettš¹ās title has changed from āContent Managerā to āContent Strategistā, to better reflect the work that he does on our team.
Other updates
Weāve set up aĀ Substack mirror for our weeklyĀ EA Forum Digest. We hope that this helps more people find and engage with valuable Forum posts!
Weāve launched a redesigned version ofĀ effectivealtruism.org! š Our primary goal was to improve visitorsā understanding of EA, broadly as part ofĀ CEAās goal of improving the EA brand. Weāve also aimed to humanize EA more than the previous design did, and to more clearly highlight how visitors canĀ take action.
We are continuing to run theĀ EA Opportunity Board and the associated bi-weeklyĀ newsletter. Please share it with anyone who might benefit! š
We built Forethoughtās newĀ website (designed by AndāNow) as part of an external partnership.
As a reminder, you can view our teamās half-quarterly OKRs viaĀ this public doc that I keep updated. I recently added our Q2.2 plans (May 20 - July 1).
Want to help the EA Forum?
Weāre looking for volunteers to help us keep the Forum running smoothly, as Forum facilitators! š If you already like being on the site, this is a way to give back to the community.
Processing new users, approving non-spam accounts and messaging users who seem confused
Generally being helpful and welcoming to new users
Helping us properly categorize new posts
Assisting the Forum Team and moderators by flagging relevant issues to us
This is a vital role to ensure this community space continues to run smoothly. Youāll join our moderation Slack workspace, and thereās the potential to take on more responsibility if youāre interested.Ā This post from two years ago is still a pretty good description of what the moderator vs facilitator roles are like.
The hours and the tasks are both flexible, and we can work with you to find a good fit. :)
We currently don't have any official facilitation capacity over the weekend, so sometimes new users don't get processed for multiple days. We'd be especially grateful for people who are willing to spend 30 min over the weekend clearing the short queue.
Share your feedback
Thanks for reading! :) Let us know if you have feedback or questions about these changes. You can comment on this post orĀ reach out to us another way. You can also share feature requests in theĀ feature suggestion thread.
@Toby Tremlettš¹ will be running a Forum workshop and office hours at the upcomingĀ EAG London ā feel free to connect with him there!
Iād currently estimate that this is 2.5 FTE: ~0.5 FTE for running the team, ~0.75 FTE of content, ~1 FTE of engineering, and ~0.25 FTE of product design. Our budget for 2025 is ~$1.3m, and most of that was for our staff of 6 FTE, so you can estimate the current cost of the Forum team to be ~$600k per year, plus some additional overhead that CEA pays on top of this (such as flying people to retreats).
For example, it's likely we will put capacity towards improving the online presence of the EA Opportunity Board and EA Funds in the second half of 2025.
I think the cost per hour of engagement is a good intuitive metric to assess the cost-effectiveness of running the EA Forum. From footnote 1, the daily cost to run the EA Forum is 1.48 k$ (= 1.3*10^6*2.5/6/365.25). There have been around 240 hours of engagement per day over the last 6 months or so. So the cost per hour of engagement has been roughly 6.17 $ (= 1.48*10^3/240). I suspect the engagement time would drop significantly if users had to pay 6.17 $ per hour they spend on the EA Forum. I believe this suggests the marginal cost-effectiveness of running the EA Forum is negative (trusting my guess for the users' revealed preferences), and that there should be a reduction in the time spent running the EA Forum.
The iteration of September 2023 of The Introductory EA Program had 1.58 attendances per hour spent running the program. For 30 $ per hour spent running the program, and 4 h of engagement per attendance (3 h of preparation, plus 1 h of discussion), there would be 6.32 h (= 1.58*4) of engagement per hour spent running the program, and the cost per hour of engagement would be 4.75 $ (= 30/6.32), 77.0 % (= 4.75/6.17) as much as for the EA Forum. So, considering uncertainty, it looks like the iteration of September 2023 of The Introductory EA Program had a cost per hour of engagement similar to that of the EA Forum nowadays.
You can try to make comparisons with other programs, or estimate which work you are doing which generates the most engagement time per $ (although figuring out the additional engagement time a given activity generated is tricky; for example, running an EA Forum event will of course cause engagement with the posts and comments related to the event, but will tend to decrease the engagement with other posts and comments).
@Toby Tremlettš¹, you may be interested in this comment.
Hmm, I think thereās some sense to your calculation (and thus I appreciate you doing+showing this calculation), but the $6.17 conclusionāspecifically, āengagement time would drop significantly if users had to pay 6.17 $ per hour they spend on the EA Forum, which suggests the marginal cost-effectiveness of running the EA Forum is negativeāāstrikes me as incorrect.
What matters is by how much engaging with the Forum raises altruistic impact, which, insofar as this impact can be quantified in dollars, is far,Ā far higher than what one would be willing and able to pay out of oneās own pocket to use the Forum. @NunoSempere once estimated the (altruistic) value of the average EA project to be between 10 and 500 million dollars (see cell C4 ofĀ this spreadsheet;Ā hereās the accompanying post). That is far higher than the actual dollar cost of running the average project. (Indeed, if one is funded by EA money, then oneās generation of altruistic dollars needs to outpace oneās consumption of actual dollarsāand by a large multiplier, if one is to meet the funding bar.)
Going back to NuƱoās spreadsheet: If I make the arrogant assumption that Iām within an order of magnitude of Ben Todd, impact-wise, then that means my lifetime impact is at least 10 million dollars. Assuming linearity (which isnāt a great assumption, but letās go with it for now) and a career length of 40 years, this means my impact over the past 4 years has been ā„1 million dollars.[1]Ā In that time, Iāve spent maybe 500 hours on the EA Forum.[2]Ā Meanwhile, Iād say that the Forum has contributed greatly to my intellectual development, i.e., added at least 20% to my impact. (The true percentage may in fact be much higher, because ofĀ crucial considerations that the Forum has helped me orient toward, but letās lowball things at 20%, for now.) This would imply that my impact has been amplified by at least $200,000/(500 hours) =Ā $400 per hour spent on the Forum. ([Insert usual caveats about there being large error bars.]) Contrast with your $6.17.
(I did this calculation on myself not because Iām special, but because I know what the numbers are for myself. Iād guess that the per-hour bottom line for other Forum users would be ~similar.)
We can now go one step further, and estimate the Forumās āaltruistic dollar generated per actual dollar spentā multiplier to be at least 400/6.17 ā 65. Embarrassingly, I donāt know how this compares against todayās funding bar,[3]Ā but seems very plausible to me that itās above.
(Nonetheless, people may still not pay $6.17/hour to use the Forum because $6.17/hour is a non-trivial cost considering peopleās actual incomes. Additionally, people are just used to being able to browse the internet for free, and so I suspect many wouldnāt do the expected value calculations and reach the ārationalā conclusion that they should in fact pay.)
Sanity check: 80,000 HoursĀ says that impactful roles generate millions of dollars worth of altruistic impact per year.
That is, 500 hours consuming the Forumās content. Iāve also spent time writing on the Forum, but if we model the Forum as a two-way market, with writers and consumers, and say that itās the consumers who benefit from being here, then it doesnāt make sense to include my writing time. (Alsoāand perhaps more relevantlyāI donāt think writing time gets counted by the Forumās analytics engine as engagement time if itās spent mostly in a Google doc.)
Further detail: What really matters is what the multiplier is on the margin (i.e., what it is for the last dollar being spent on a project), rather than what it is for the project as a whole.
Thanks for sharing your thoughts, Will! I agree what matters is the additional altruistic impact caused by engaging with the Forum. However, I think my point holds as long as people have accurate views about how to maximise their altruistic impact.
For example, if you believed "factual impact of your marginal hour on the Forum" - "counterfactual impact of this hour" < "impact of donating 100 $[1]Ā to the organisation or project you consider the most cost-effective", and using the Forum costed 100 $/h, I think you would have a greater altruistic impact by your own lights by spending less time on the Forum, and donating the savings. Do you agree?
Analogously, if the user spending the marginal hour on the Forum believed "factual impact of their marginal hour on the Forum" - "counterfactual impact of this hour" < "impact of donating 6.17 $ to the organisations or projects they consider the most cost-effective", and using the Forum costed 6.17 $, I think they would have a greater altruistic impact by their own lights by spending less time on the Forum, and donating the savings. In this case, if the marginal user-hour costed 6.17 $ to the Forum team[2], I believe they would also increase altruistic impact in the eyes of the user of the marginal hour by spending less time generating engagement on the Forum, and donating the savings to the organisations or projects that user considers the most cost-effective.
Implying the marginal cost-effectiveness of your time on the Forum is 25 % (= 400/100) of the past cost-effectiveness.
I guess it costs more due to increasing user-hours becoming more difficult as user-hours increase.