4 ms·

How the fuck was uber's pricing unsustainable? They aren't manufacturing anything physical. The core business doesn't need tons of employees. They're aren't pay

How the fuck was uber's pricing unsustainable? They aren't manufacturing anything physical. The core business doesn't need tons of employees. They're aren't paying extra for drivers' gas or mileage. The app is conceptually simple. Where is all the money going? Why is it WhatsApp can serve most of the fucking world with something like 100 people but Uber needs a huge staff?

twelve40 3y ago
well when they started, the whole concept of ridesharing was unproven. They didn't just set out to replace taxis, the idea was to offer a service initially so cheap that people who otherwise wouldn't even think about getting a car, would start doing it on a regular basis (same with Eats). So yes, i'm pretty sure they have been subsidizing rides, and in certain markets probably still are. So now the question is: can this market - where some people even grew accustomed to commuting to work on Uber - sustain itself, or is it too expensive and people will just revert to the old mode where (outside of NY) getting a cab was a rare event. tldr - i don't think it's the staff that's the expensive part. Also, they do a bunch of non-trivial things with maps, etc, so not quite a simple crud app.
- pdntspa 3y ago
  Of course... the specifics of routing so many rides at scale is certainly non-trivial. But in 2019 they had 26k employees, 32k in 2022 according to https://www.macrotrends.net/stocks/charts/UBER/uber-technologies/number-of-employees https://www.macrotrends.net/stocks/charts/UBER/uber-technolo...
  - twelve40 3y ago
    dunno... that's probably a lot, but unlike say twitter, they do a lot more stuff. With Eats, you have to coordinate with restaurants locally, etc etc. So while it sounds a bit bloated, i think there are worse examples elsewhere in tech. But I don't think it matters very much at all. If people are the problem, they can be quickly cut. But if their market fundamentally shrinks or is unable to pay for itself, you can't really fix that.
bruce511 3y ago
It's a good question. There are enough local clones to suggest it's not the software that eats all the money. Bandwidth plays a role. Updating driver positions constantly, at scale, chews bandwidth. As does broad-casting trips, acceptances, and so on. I'm guessing Lawyers take a huge chunk. Uber is forever being sued especially in years past and they must have a pretty terrible legal bill. It was always part of the model to operate outside the law, so legal issues were inevitable. Plus lobbying - at every level. Changing the model (which they have done) takes a lot of lobbying. More lawyers, at lawyer rates. Marketing of course takes a chunk. That's to be expected, especially as competitors enter the market (riding on the back of their legal coat-tails). By comparison WhatsApp has no legal bill, no marketing bill, no lobbying bill, and so on.
- piva00 3y ago
  > Bandwidth plays a role. Updating driver positions constantly, at scale, chews bandwidth. As does broad-casting trips, acceptances, and so on. How could this chew bandwidth? It's just some bytes for the positioning, over some socket (probably a websocket), it really doesn't sound like it'd chew bandwidth to send an ID + coordinates every second or so if your connection is already established. Maybe a few more bytes for a signature to avoid tampering/spoofing when updating the position.

twelve40 3y ago

well when they started, the whole concept of ridesharing was unproven. They didn't just set out to replace taxis, the idea was to offer a service initially so cheap that people who otherwise wouldn't even think about getting a car, would start doing it on a regular basis (same with Eats). So yes, i'm pretty sure they have been subsidizing rides, and in certain markets probably still are. So now the question is: can this market - where some people even grew accustomed to commuting to work on Uber - sustain itself, or is it too expensive and people will just revert to the old mode where (outside of NY) getting a cab was a rare event. tldr - i don't think it's the staff that's the expensive part. Also, they do a bunch of non-trivial things with maps, etc, so not quite a simple crud app.

pdntspa 3y ago
Of course... the specifics of routing so many rides at scale is certainly non-trivial. But in 2019 they had 26k employees, 32k in 2022 according to https://www.macrotrends.net/stocks/charts/UBER/uber-technologies/number-of-employees https://www.macrotrends.net/stocks/charts/UBER/uber-technolo...
- twelve40 3y ago
  dunno... that's probably a lot, but unlike say twitter, they do a lot more stuff. With Eats, you have to coordinate with restaurants locally, etc etc. So while it sounds a bit bloated, i think there are worse examples elsewhere in tech. But I don't think it matters very much at all. If people are the problem, they can be quickly cut. But if their market fundamentally shrinks or is unable to pay for itself, you can't really fix that.

bruce511 3y ago

It's a good question. There are enough local clones to suggest it's not the software that eats all the money. Bandwidth plays a role. Updating driver positions constantly, at scale, chews bandwidth. As does broad-casting trips, acceptances, and so on. I'm guessing Lawyers take a huge chunk. Uber is forever being sued especially in years past and they must have a pretty terrible legal bill. It was always part of the model to operate outside the law, so legal issues were inevitable. Plus lobbying - at every level. Changing the model (which they have done) takes a lot of lobbying. More lawyers, at lawyer rates. Marketing of course takes a chunk. That's to be expected, especially as competitors enter the market (riding on the back of their legal coat-tails). By comparison WhatsApp has no legal bill, no marketing bill, no lobbying bill, and so on.

piva00 3y ago
> Bandwidth plays a role. Updating driver positions constantly, at scale, chews bandwidth. As does broad-casting trips, acceptances, and so on. How could this chew bandwidth? It's just some bytes for the positioning, over some socket (probably a websocket), it really doesn't sound like it'd chew bandwidth to send an ID + coordinates every second or so if your connection is already established. Maybe a few more bytes for a signature to avoid tampering/spoofing when updating the position.