What proportion of people with Dropbox or Google Docs or Hotmail are paying customers?
Dropbox is nice enough to list it.
They have 700m users, and 18m of them pay for the service, so about double of OpenAI. Completely unlike OpenAI, however, they make quite a bit of profit, having a revenue of 2.55b and 1.63b in operating costs. OpenAI subscribers can’t even cover their own cost of inference.
Thanks for the numbers, you were saying the subscriber percentage was embarrassing so I was curious about that rather than their fairly infamous losses.
You’ve said OpenAI have about half the subscriber percentage of Dropbox, but if Dropbox is that profitable then that seems like they are doing particularly well and perhaps that subscriber percent is above average?
Peope don’t usually compare OpenAI with Dropbox, but Dropbox isn’t particularly great with the free-to-paid converion rating. 2.6% is pretty bad, but they STILL manage to make money because what they do is pretty cheap on a per-user basis. They just host data, and most of that data isn’t really used much. Also, I don’t know if Dropbox is “that” profitable. All I could find is that their revenue exceeds their operating costs, but I don’t know if that covers R&D or marketing, which they probably spend a LOT of money on.
Comparisons with YouTube and Spotify get thrown around a lot more, which convert around 5% and a whoppingly insane 36%, compared to OpenAI’s measly 1%. And both YouTube and Spotify actually make a LOT of money on their “free” users, via ads. OpenAI has no monetisation beyond subscribers, and they’re very bad at getting people to pay for their stuff.
Pretty sure they didn’t for a while. It’s the same approach as always. Operate at a loss, gain (nonpaying) users, maybe sell or use their data and slowly turn up whatever you are doing to make money (ads, fee).
Glad you’re enjoying it, that’s awesome. I used it a decade and more ago, but Spotify never worked out for me for a few reasons:
I discovered hi-fi music, have multiple good setups and for years Spotify was lower quality. It took them close to a decade to do FLAC quality.
I listen to a large variety of stuff from multiple countries, and Spotify would often lose rights to tracks I liked. At one point something like 10-15% of the music on my list disappeared.
I prefer to own my music in high quality, have bought a ton of albums and have around 1.2 TB of FLAC on my NAS.
I don’t like how artists get completely screwed on pay by Spotify.
My home setup has multiple budget hi-fi setups with passive speakers and mini-amps, and a 5.1 system in my living room. I’ve got a mini-PC server that runs Lyrion with Bliss mix plugin, Plex (for PlexAmp), and AssetUPnP. I still have Qobuz for streaming and discovery (and it integrates in Lyrion to mix with my local collection), but prefer local music overall.
Imagine having 79 unpaid users for every subscriber. Most companies would die from embarrassment.
It’s like Google, Facebook, and Microsoft don’t exist.
Let’s say it together, kids: “If the service is free, then you’re the…?”
penis? No wait, that doesn’t sound right…
The companies are stroking it to get the valuable juice out, and then steal your orgasm too. All you’re left with drinking more water to sustain that.
What’s a normal amount? What proportion of people with Dropbox or Google Docs or Hotmail are paying customers?
Having a little over 1% doesn’t seem that bad, I am faar more surprised that over 1% of users pay for ChatGPT (if your numbers are accurate).
Dropbox is nice enough to list it.
They have 700m users, and 18m of them pay for the service, so about double of OpenAI. Completely unlike OpenAI, however, they make quite a bit of profit, having a revenue of 2.55b and 1.63b in operating costs. OpenAI subscribers can’t even cover their own cost of inference.
Thanks for the numbers, you were saying the subscriber percentage was embarrassing so I was curious about that rather than their fairly infamous losses.
You’ve said OpenAI have about half the subscriber percentage of Dropbox, but if Dropbox is that profitable then that seems like they are doing particularly well and perhaps that subscriber percent is above average?
Peope don’t usually compare OpenAI with Dropbox, but Dropbox isn’t particularly great with the free-to-paid converion rating. 2.6% is pretty bad, but they STILL manage to make money because what they do is pretty cheap on a per-user basis. They just host data, and most of that data isn’t really used much. Also, I don’t know if Dropbox is “that” profitable. All I could find is that their revenue exceeds their operating costs, but I don’t know if that covers R&D or marketing, which they probably spend a LOT of money on.
Comparisons with YouTube and Spotify get thrown around a lot more, which convert around 5% and a whoppingly insane 36%, compared to OpenAI’s measly 1%. And both YouTube and Spotify actually make a LOT of money on their “free” users, via ads. OpenAI has no monetisation beyond subscribers, and they’re very bad at getting people to pay for their stuff.
Yeah that’s a fair point. I don’t think Dropbox does ads but the others I mentioned and the ones you mentioned all show/play ads for the free tier.
I guess OpenAI will be pretty keen to get ads into their free tier too, once they run out of investors’ money.
Pretty sure they didn’t for a while. It’s the same approach as always. Operate at a loss, gain (nonpaying) users, maybe sell or use their data and slowly turn up whatever you are doing to make money (ads, fee).
Sure, it took dropbox 9 years, Amazon took 7.
OpenAI just turned 10, with profit nowhere in sight, and a path to profitability completely invisible.
This is called moving the goal post
What is? Pointing out that dropbox has a conversion more than twice as high as OpenAI?
The paid subscribers subsidize the unpaid ones. Sam Altman is a staunch socialist. Trump too, that communist-loving Mamdani fan. /s
Even for paying customers, inference alone costs OpenAI several times more than revenue.
Honestly, having ~1.3% of paying users is fine considering they are using all the other users as free learning material for their models.
It’s really not, since all their models cost progressively more to run, and don’t bring in any more money.
Remember, all those paying subscribers COST them money every month.
Which is really pathetic since we are paying their utility bills.
I’d rather not imagine being Spotify. Feels gross.
Spotify has 280m paid users, out of 780m monthly users. Roughly 26 times better conversion than OpenAI
https://investors.spotify.com/financials/default.aspx#quarterly-results
TIL. Holy shit, that’s wild.
tbf Spotify is easily the best subscription I’ve got, according to my wrapped I was listening to the equivalent of 82 days straight in the last year
Glad you’re enjoying it, that’s awesome. I used it a decade and more ago, but Spotify never worked out for me for a few reasons:
I discovered hi-fi music, have multiple good setups and for years Spotify was lower quality. It took them close to a decade to do FLAC quality.
I listen to a large variety of stuff from multiple countries, and Spotify would often lose rights to tracks I liked. At one point something like 10-15% of the music on my list disappeared.
I prefer to own my music in high quality, have bought a ton of albums and have around 1.2 TB of FLAC on my NAS.
I don’t like how artists get completely screwed on pay by Spotify.
My home setup has multiple budget hi-fi setups with passive speakers and mini-amps, and a 5.1 system in my living room. I’ve got a mini-PC server that runs Lyrion with Bliss mix plugin, Plex (for PlexAmp), and AssetUPnP. I still have Qobuz for streaming and discovery (and it integrates in Lyrion to mix with my local collection), but prefer local music overall.
And the subscribers are underpaying by such a huge amount that they’re losing even more money from them than they are from the unpaid users.