Hacker News Reader

Hacker News

Comment by abhibeckert | Hacker News Reader

Parent Full thread

abhibeckert

The cost is something like $500 per user per year. People just aren't going to pay that much.

View on HN

I know you're not saying this, but just to point out how absurd this price point is:

> Reddit generated $350 million in 2021, primarily from its advertising business

> Reddit has 52 million daily active users and approximately 430 million users who use it once a month

Are we really supposed to believe that Reddit is losing $500 * 400 million per year? Their total cost per user is probably something like $1/year. Twitch.tv which streams 1080p video prolly has costs of like $50/user/year. Insane pricing decision by Reddit.

100/minute are free. So roughly $0 per year which most would probably pay. But getting an API key, and the TOS are issues.

This is not accurate.

From the data presented, in aggregate, >90% fall well within the free tier of usage (100 API calls per minute for OAuth logged in client). The majority of the remaining fall into a $1/month tier [1].

Certainly there are some outliers that exceed those ranges and unfortunately the majority of the spotlight seems to be on the smallest of the populace.

It is also worth noting that app developers can (and should) better optimize their apps to reduce calls to API (bundling, caching, etc) which will help to reduce the number of calls required overall to take that cost down. Many apps do this already.

[1]https://www.reddit.com/r/reddit/comments/145bram/addressing_...

Why would you take the word of reddit management on this issue? That seems like the least trustworthy source here

Because they have gone on public record with those numbers and as of yet, no one has refuted them.

Do you have an alternate source of user level data that breaks down average vs. outliers and that ratio that you feel is more accurate?

The free tier is 100 queries per minute for an OAuth logged in client per the public statement, and that number has also been stated by a few 3rd party app developers.

That is a lot for the average user, especially for a well optimized app.

The main issue here is that the number isn't relevant: it's per API key, not per user. So unless reddit starts handing out API keys to every user then it doesn't make the slightest difference to third party apps.

It is relevant. Developers have the ability to throttle and/or monetize users based on usage to pass on those costs to power user outliers, while not impacting the majority of their users that fall into that free tier.

The point is "user" as used in this post is not a single reddit account or a single individual. A "user" is an app which supports few or many users. I would say if you measure it the way most people would mean user, in fact 90% or more of users are impacted by this price increase, because while each individual user will fall well below the usage threshold, they will in aggregate quickly exceed it because it's all being attributed to the app they are using. On the free tier you could support maybe 10 average users before you wind up with a bad experience as their requests start to bounce off the limits (it's not that their average usage would hit the limits, but the spikes would quickly result in requests become very delayed)

How many request per minute do you think most apps make when casually browsing reddit?

OK, so if each user were to use their own API key, reddit would still have the same costs but nowhere near the revenue, which speaks to a pricing structure aimed at the status quo where the API key is per developer instead. What's the betting that if apps switch to allowing users to specify their own API key that reddit actually allows most users to do this? The '90%' number is meaningless: apps which use the reddit API (or any API with a free tier) have an extremely long tail of niche uses which effectively no-one cares about.

The comment about API optimization is inane, apps already try to do this and bump up against the problem that reddit's API is atrocious (try actually getting all the comments in a comment chain once they get collapsed! if you actually want to succeed, even on a chain with only ~100 comments, you're gonna need around 30 API calls. Most apps don't bother to try, and the 'more comments' button is invariably broken).

I'm not aware that reddit ever presented user level API keys as an option, they'd likely instead base it on Reddit premium tiers tied to the OAuth account. But again, that's not currently a viable option on the table and if users started enmass requesting API keys intended for developers, yes that would get throttled quickly because that's not what they're intended for.

>90% of users is not a meaningless number by any measure. I'd suggest that punishing the average users due to the power use of the outliers is a terrible value proposition.

Developers hold the ability to throttle and/or price based on that.

I am a developer that has built clients and tools around the reddit API. If most apps don't bother to optimize, then that's a choice they can make.

Reply on news.ycombinator.com