Yeah, that seems like the right comparison? Revenue is a proxy for value produced, so if you are arguing about whether something is worth funding philanthropically, revenue seems like the better comparison than costs. Though you can also look at costs, which I expect to not be more than a factor 2 off.
Isn’t a lot of FB’s revenue generated by owning a cookie, and cooperating with other websites to track you across pages? I don’t think it’s fair to count that revenue as generated by the social platform, for these purposes.
Your argument also feels slippery to me in general. Registering that now in case you have a good answer to my specific criticism and the general motte-and-bailey feeling sticks around.
I am not sure what you mean by the first. Facebook makes almost all of its revenue with ads. It also does some stuff to do better ad-targeting, for which it uses cookies and does some cross-site tracking, which I do think drives up profit, though my guess is that isn’t responsible for a large fraction of the revenue (though I might be wrong here).
But that doesn’t feel super relevant here. The primary reason why I brought up FB is to establish a rough order-of-magnitude reference class for what normal costs and revenue numbers are associated with internet platforms for a primarily western educated audience.
My best guess is the EA Forum could probably also finance itself with subscriptions, ads and other monetization strategies at its current burn rate, based on these number, though I would be very surprised if that’s a good idea.
I think the more relevant order of magnitude reference class would be the amount per user Facebook spent on core platform maintenance and moderation (and Facebook has a lot more scaling challenges to solve as well as users to spread costs over, so a better comparator would be the running expenses of a small professional forum)
I don’t think FB revenues are remotely relevant to how much value the forum creates, which may be significantly more per user than Facebook if it positively influences decisions people make about employment, founding charities and allocating large chunks of money to effective causes. But the effectiveness of the use of the forum budget isn’t whether the total value created is more than the total costs of running, it’s decided at the margin by whether going the extra mile with the software and curation actually adds more value.
Or put another way, would people engage differently if the forum was run on stock software by a single sysadmin and some regular posters granted volunteer mod privileges?
Or put another way, would people engage differently if the forum was run on stock software by a single sysadmin and some regular posters granted volunteer mod privileges?
Well, I mean it isn’t a perfect comparison, but we know roughly what that world looks like because we have both the LessWrong and OG EA Forum datapoints, and both point towards “the Forum gets on the order of 1/5th the usage” and in the case of LessWrong to “the Forum dies completely”.
I do think it goes better if you have at least one well-paid sysadmin, though I definitely wouldn’t remotely be able to do the job on my own.
As to costs, I’d have to dig further but looking at the net profit margin for Meta as a whole suggests a fairly significant adjustment. Looking at the ratio between cost of revenue and revenue suggests an even larger adjustment, but is probably too aggressive of an adjustment.
If Meta actually spent $200 per user to achieve the revenue associated with Facebook, that would be a poor return on investment indeed (i.e., 0%). So I think comparing its revenue per user figure to the Forum’s cost per user creates too easy of a test for the Forum in assessing the value proposition of its expenditures.
Yep, totally, it’s a pretty bad proxy. I think the obvious analogy at least for the EA Forum would be that the organizations who are hiring people from the EA Forum are in a comparable position to advertisers, but it’s not amazing.
Yeah, that seems like the right comparison? Revenue is a proxy for value produced, so if you are arguing about whether something is worth funding philanthropically, revenue seems like the better comparison than costs. Though you can also look at costs, which I expect to not be more than a factor 2 off.
Isn’t a lot of FB’s revenue generated by owning a cookie, and cooperating with other websites to track you across pages? I don’t think it’s fair to count that revenue as generated by the social platform, for these purposes.
Your argument also feels slippery to me in general. Registering that now in case you have a good answer to my specific criticism and the general motte-and-bailey feeling sticks around.
I am not sure what you mean by the first. Facebook makes almost all of its revenue with ads. It also does some stuff to do better ad-targeting, for which it uses cookies and does some cross-site tracking, which I do think drives up profit, though my guess is that isn’t responsible for a large fraction of the revenue (though I might be wrong here).
But that doesn’t feel super relevant here. The primary reason why I brought up FB is to establish a rough order-of-magnitude reference class for what normal costs and revenue numbers are associated with internet platforms for a primarily western educated audience.
My best guess is the EA Forum could probably also finance itself with subscriptions, ads and other monetization strategies at its current burn rate, based on these number, though I would be very surprised if that’s a good idea.
I think the more relevant order of magnitude reference class would be the amount per user Facebook spent on core platform maintenance and moderation (and Facebook has a lot more scaling challenges to solve as well as users to spread costs over, so a better comparator would be the running expenses of a small professional forum)
I don’t think FB revenues are remotely relevant to how much value the forum creates, which may be significantly more per user than Facebook if it positively influences decisions people make about employment, founding charities and allocating large chunks of money to effective causes. But the effectiveness of the use of the forum budget isn’t whether the total value created is more than the total costs of running, it’s decided at the margin by whether going the extra mile with the software and curation actually adds more value.
Or put another way, would people engage differently if the forum was run on stock software by a single sysadmin and some regular posters granted volunteer mod privileges?
Well, I mean it isn’t a perfect comparison, but we know roughly what that world looks like because we have both the LessWrong and OG EA Forum datapoints, and both point towards “the Forum gets on the order of 1/5th the usage” and in the case of LessWrong to “the Forum dies completely”.
I do think it goes better if you have at least one well-paid sysadmin, though I definitely wouldn’t remotely be able to do the job on my own.
As to costs, I’d have to dig further but looking at the net profit margin for Meta as a whole suggests a fairly significant adjustment. Looking at the ratio between cost of revenue and revenue suggests an even larger adjustment, but is probably too aggressive of an adjustment.
If Meta actually spent $200 per user to achieve the revenue associated with Facebook, that would be a poor return on investment indeed (i.e., 0%). So I think comparing its revenue per user figure to the Forum’s cost per user creates too easy of a test for the Forum in assessing the value proposition of its expenditures.
No it fucking isn’t—that’s an approximation of the value it generates advertisers, not users.
Yep, totally, it’s a pretty bad proxy. I think the obvious analogy at least for the EA Forum would be that the organizations who are hiring people from the EA Forum are in a comparable position to advertisers, but it’s not amazing.