> less than $1.00 per user / month for a typical Reddit third-party app

This types of generalization irks me everytime. "typical user send 20 emails every month, surely if you want to send more than 20 emails you will happily pay for it".

I don't think that's the expectation here, as that would imply that users who aren't generating much traffic per user would be free, while you'd only charge users who are generating more traffic. Which would be terribly broken in this case, as Reddit is a long-tail service where the users who each make few requests, add up to the supermajority of the traffic served.

I think their expectation on Reddit's end is that these third-party apps "should" be charging a subscription fee to every single user that uses them, of something like $12/year, and passing most of that on to Reddit; where the heavy users will be negative-margin to retain, but where this doesn't matter because the users who don't use Reddit much will be subsidizing them. Like insurance.

Which is also terribly broken, for classical insurance-economics reasons: just like healthy people generally don't care about getting health insurance unless they're forced to do so, people who don't use Reddit much certainly won't be willing to pay $12/year for it. So it'll only be those power users interested in paying; and so the average usage among that group will be much higher; and so you'll need to charge them quite a bit more than $12/year... and they probably won't be willing to pay that.

This AMA is going as well as predicted.

Did I parse his usage statement correctly? Is he saying a typical reddit user only consumes 5000 API calls per _month_??? I find that VERY hard to believe.

Power-law usage distribution. The median Reddit user probably opens the app for five minutes a day. People who even bother to make accounts / join non-default subreddits are a minority. People who leave comments on Reddit are basically bigfoot.

Reddit is a social media app, you can’t consume anything significant in that time. So again, hard to believe.

And if this is somehow true then it doesn’t really paint a positive picture of Reddit’s business.

And this is about API calls, not user engagement. I can’t imagine that the difference between people who post and those who don’t, in terms of API usage, to be significant based only on how many API calls it takes to post a comment.

Why do you think the median user of any social-media app "consumes anything significant" through the app?

The median user of Facebook, Twitter, Instagram, TikTok, or whatever else, uses the app for on average ~five minutes a day. They just "check" it to see if there's anything interesting; they find that there's not; and then they close it. Or, they just decide to open one such app when waiting for a bus, and scroll until their bus gets there; then they open a different app while waiting for their lunch order to be made, then close that one when they get their lunch; etc.

Most people who "use" social media, are not addicted to social media, or even engaged with social media. They just barely use social media.

This group of people is why every single MAU figure is considered to be heavily inflated. All you have to do is open the app for a minute once a month and suddenly you're an "active user" for that month. Every social network has far, far fewer "actually active" users than they try to make it seem.

> And this is about API calls, not user engagement.

If you aren't aware: engagement drives linger time. The same people who create accounts, etc. are the people who spend hours on the app. People who just "poke their nose in" but don't engage almost always spend less time on the app. Degree of engagement with an app, can be used as a very reliable proxy (by third-party analysis firms who don't have access to actual usage-time data) for likelihood to spend more time using a given app.

No, I don't think you did...it would be 4000 API calls per month.