And then there's Uber self-driving. Uber AI; Uber electric airplanes. Uber freight, Uber restaurant delivery, Uber grocery delivery, Uber this and Uber that. Oh, and Uber scooters.
Uber eats? Scooter integration? Mass transit support? Scheduled rides? Commuter cards? If you were building the app with 30 developers you'd simply not bother with those features.
There's legitimately a reasonable argument that Uber rides has a worse business case than Uber Eats. If I were in Uber's shoes I would be clinging onto both.
Logistics is a tough business so I’m bearish on Uber eats.
I reaaaaly wouldn’t assume its a profitable business unless you get some data from a few “normal” (i.e., not pandemic impacted) quarters.
Delivery is the first thing people will cut down on if we get a recession as well
At least that’s what Uber recruiters would tell me when trying to recruit.
Uber has 3x the headcount as eBay.
The eBay business requires no boots on the ground, Uber Eats does because they have to equip riders. And eBay has almost no market-specific laws/regulations which change every couple of months to worry about.
I think eBay doesn't process payments, that is outsourced to PayPal.
It looks like eBay is in about 25 markets, Uber is available in 85 countries.
Serious question: Why does eBay have so many employees? It's just a search engine for a user generated product catalog where users can place bids on the items. My guess is that most Uber employees aren't high-paid developers, it is more likely that they are in support roles.
Uber has to market themselves which needs local expertise, if nothing else to liaise with a local PR firm. Then they need local legal expertise to actually operate in the country (eBay transactions happen online, and the transport agency hired by the seller figures out how to get the package to its destination). Uber then has to have maps for every country it operates in, as well as change their standards to match local expectations.
I ignored the other stuff because I don't know WTF "pickup special cases" or "on-trip experience business logic" or "growth features" are. So I'm not informed enough to guarantee they aren't part of the core product offering - although you can probably guess my intuition on the matter.
The software that runs the world and gets actual work done day-in, day-out behind the scenes, is riddled with edge case handling. In really mature codebases impacting many stakeholders (not just direct users), the product team can categorize 1% or less of the stakeholder population by a tiny fraction of a commonly-used feature set they use. On that basis, the coding and maintenance effort for the edge cases can sometimes outweigh the sliver of features used by that sub-population of stakeholders.
We aren't talking about a web scraper or run of the mill DevOps here. Anytime you work with lots of business rules in multiple jurisdictions impacting the same processes, the edge case counts go up rapidly, and combinations of processes that you never thought would intersect but are forced to by specific jurisdictions also appear more frequently.
Here is how it works in most of my clients. Management is at the top of the heap. Business/sales/marketing comes next. Developers/engineering after them. Way down at the bottom of the heap, sits Operations. They're the "grunts".
The default dynamic is developers code up a system and toss it over the wall to Operations. When informed by operations of some edge case that comes up frequently on the sad path, developers roll their metaphorical eyes and tells Operations to "just" develop and follow some "SOP".
Cue "One Decade Later..." in French accent meme...
After enough of those edge cases over enough time, it isn't hard to see why all those manual intervention edge cases result in an operations-oriented staff bloat. These are especially challenging to address, because everyone is looking for a silver bullet and very few people in charge of writing checks will accept the reality that this is every bit as much technical debt as the code that runs on the systems, and working off that debt with interest is a slog of expensive unwinding.
Uber gets hit on multiple fronts with this dynamic. More jurisdictions, in an incumbent industry that has accreted more baroque government interfacing, that has simultaneously famously resisted making any of its relevant data available in any form whatsoever beyond paper and microfiche. Sometimes it will take more people than you can imagine.
[1]https://www.wired.com/2015/09/whatsapp-serves-900-million-us...
That's after shitting on compliance for years as part of your business model.
McDonald's restaurants have handicap accessible ramps but they don't staff ESL speakers. Apps and web services are no different.
Clubhouse, for example, isn't accessible to the hearing impaired and that's ok.
Uber specifically would definitely fall under a "public transportation" service - accessibility is non-optional if someone decides to sue. [2]
[1] https://www.ada.gov/2010ADAstandards_index.htm
[2] https://www.businessnewsdaily.com/10900-ada-website-requirem...
Firstly that is just plain incorrect in a lot of areas (of business) and jurisdictions. Secondly your outlook on accessibility is just ... SMH and walks away from keyboard.
Do you think complexity of Uber is comparable to that of Whatsapp?
There is a big difference here -- revenue. Real transactions. Financial reporting. Country-by-country regulations and reg reporting.
- iOS rider app
- android rider app
- iOS driver app
- android driver app
- ride/driver matching
- routing/supply
- security/compliance
- fraud detection/prevention
- backend rides services
- backend user services
^ 10 teams of at least 10 people off the top of my head. Amazing how underestimated engineering resourcing needs are.
Wow that was easy.
That is because its very well supported on both platforms, and one of the few language ecosystems that can actually create native looking/acting applications on both.
Yes, the initial development may be a bit slower, but that is common with C++ because what the smart people are doing is usually creating an application specific "language" out of C++. Then once that core bit is done the actual development would probably outpace many of the other choices. I've done this a couple times with native development toolkits, the "customer" in one case was really questioning how much work was actually going on when 1/2 way through the contract it barely had a single "screen" in a data collection/reporting application working. But, then another few weeks went by, and literally in the space of about 2 days the application went from looking like it had just been started, to being basically feature complete. That is because the C++ engine was complete and it took ~12 hours to fill out the few 10s of thousands lines of boilerplate ui descriptions that actually formed the UI (in that case it was a custom textish/declaritive application description language, most of the c++ code was parsing it and doing layout/drawing in response. Super happy customer too, once they understood that their inhouse "IT" people could update/change the app with little more than a text editor against some fairly simple to understand rules).
And many of the functions the GP was asking about are the kinds of things one could contract out. Aka, why not just use google maps API for the routing/etc.
Frankly, given what I saw a couple years ago when I went poking around in the iphone app store, it seems just about every region taxi company had reimplemented large parts of the uber interface in their own apps. Maybe not always as slick, but the core parts were in many of them, and I doubt random regional taxi companies can afford the engineering effort uber apparently is spending on.
If you're looking to trim fat - surprisingly - there might be better opportunities outside of engineering.
This comment is insane for any real app in the real world transacting in real dollars. There are a hundred countries with 100 regulations. Accounting/P&L/bookkeeping for the United States ALONE would take 30 developers for a revenue base this size.
Revenue 6854
Cost of Revenue 4026
Operations and Support 574
Sales and Marketing 1263
Research and Development 587
General and Administrative 632
Depreciation & Amortization 254
Total Costs 7336I would love to see the breakdown of drivers and their immigration status, because from what I can gather, what is happening here is not any different than in the past of America’s history where the whole business model relies on the exploitation of “immigrant” labor that knows no better and is easily exploited, aka their unrealized labor value is converted into profit, or better states, benefits and riches for the executives.
Rideshare expanded the reach of taxi-like services to more regions, people, and use cases. It's now viable to travel without renting a car in many US cities - you can see this by how rideshare affected rental car companies. That's what makes the rideshare business hard: the places where they didn't have existing competition, because that's where the matching is harder and the economics not as easy.
It's a similar story for food delivery: GrubHub/Seamless operated in core areas of cities like NYC for many years (since the 90s in NYC, I think). Uber Eats brings that service to far more places, and that's where the challenge comes from in that business.
You don't need rideshare for delivering service to large areas.
In my country we have a taxi app which can be used by any taxi driver. If you request a drive, your call can be seen by all drivers which are at a certain distance from you. That distance can be se by each driver.
One or more drivers will reply to your request and you get to pick the driver.
And you can either pay directly to the driver using cash or card or you can be charged through the app.
The fee for the drivers is very small so almost all drivers use it.
And the money are not subject to tax evasion, they stay in the local economy. Taxi drivers have wages, social insurance, health insurance and pension funds. They also have a valid license to transport people and are regularly checked to be able to hold that license.
No Uber needed.
In fact I’ve had a lot of conversations with Uber and taxi drivers who started as taxi drivers, switched to Uber when the bonuses were lucrative, and then some of them switched back because they liked the predictability of a fixed shift and not being ordered around by a machine. Others felt exactly the opposite.
It sounds like "developing the core product" falls entirely under this bucket, so with what others are saying about what falls under other buckets is right, then the one that looks most cuttable to me is actually "Research and Development". That sounds like the "experimental new stuff that may go nowhere" bucket, and if it's eliminated it would also put them just above break-even. Maybe they could focus on improving profits for their core product for a while before bringing that back.
Maybe there is necessary stuff included in it though, which I guess means that wouldn't be an option.
[0] https://www.investopedia.com/terms/c/cost-of-revenue.asp
A quick look at Ubers annual report states they do not capitalise r&d costs and instead expense it as incurred.
If you look at their financials, they show gross bookings, which include both the full billed values for food and delivery, and transportation of people. Revenues only show their share of that total.
In regards to incentives, it looks like a complicated question, I found an interesting outline[2]
[1] https://seekingalpha.com/article/4293755-insurance-primary-b... [2] https://news.bloombergtax.com/financial-accounting/rideshari...
Cost of Revenue goes a step further, and includes the next layer of costs that are necessary to cause sales to happen for a given product/service line in a reporting period. That includes things like sales and marketing, and distribution. Basically anything that would cause the revenue to stop coming in fairly immediately if it wasn't done.
Neither includes R&D to create the product/service in the first place, or general overhead.
It's a clever way to disguise their unit economics to look better.
Their marketing budget is mostly going to pay drivers.
Put another way - you could be saying - why not pay drivers EVEN less? Well, they're paying them the least they can already. You can be sure of that.
If they did actually paid the drivers more and not disguise it as marketing - then their unit economics wouldn't look good - and when the business as a whole doesn't look good either - that's not a good look.
They only need some occasional reminders that might be targeted at the few people that haven't used or heard of friends or family using an Uber before.
Conversely, I'm not going to take an Uber ride tomorrow because they showed me an ad today. When I need to get somewhere I'll look at my options and choose the best. So long as Uber meets the minimum level of me being aware of it then theyre good.
Marketing is as much about keeping your customers from going to rival providers as it is about finding new customers. If you don't want people to leave you need to remind them why you're better than the exciting new company that's spending VC money to take your market share.
Take rate ~40% = $6
- payments -0.7 (correct)
- servers -0.2 (correct)
- refunds -1 (driver is indemnified, not in cabs)
- shared rides -1 (on avg, not in cabs)
- insurance -1 (drivers and riders get generous insurance, not in cabs)
now we have $2.5 per ride to pay fixed costs, not including marketing. in reality, they also give out generous marketing coupons to riders which cabs did not do. $1 per ride on avg discount drops us to $1.5 per ride.
it's probably true that 100 engineers could sustain uber in the US without any new features. 100 * 500k payroll cost = $50M/yr = 33M rides per year.
Lyft alone did >300M rides per year. Uber and lyft combined would be a solid company able to sustain and be profitable. This company would provide a way higher quality service (and safer, and more insured) than it replaced, without monopoly profits being captured by rich medallion holding families.
However, uber and lyft are still pouring money into product development. Shared rides as an example, still hold the potential to improve everything in cities massively. Lower cost than taxi, faster than bus, less congesting than private car.
Both also seem to be expanding outside of ride share, which should be viewed like amazon expanding to aws 15 years ago. if it works, there could be hugely positive impact. if it doesn't, then it's a massive waste of shareholder capital.
¯\_(ツ)_/¯
Then we'd have a(n) (inter)national tech monopoly. Soooo much better.
I agree and am not quite sure where it's going (other than software).
Lyft only has 4500 employees
Best source I could find: https://www.themuse.com/profiles/uber/team/engineering
Ultimately, I wonder if Uber is prime to be disrupted if drivers got together and funded a few engineers to build a city-scale service for the hailing and payment aspect.
Apparently a whole bunch of folks are trying to do just that.
I was going to provide just one example, but a web search[0] shows a whole bunch of these efforts in a variety of locales. As such, I just provided the web search results here.
[0] https://html.duckduckgo.com/html?q=ride%20share%20cooperativ...
NYC used to have Juno, but it went bankrupt in 2019. I feel like if you can't run a single-city rideshare app in NYC, you're gonna have a hard time doing it anywhere else.
I haven't used it (I always pay cash, as the drivers are charged ~3-5% per fare for "CC processing" when a card is used), but every NYC taxi has a feature where you can pair with the onboard system and pay with your phone.
Not sure what "pairing" requires, perhaps someone else has used this feature and can comment.
For most of Europe that means country level, for the US and other federal states, that probably means at the state level.
Searching and installing a random app introduces a slight amount of friction, but I don't think payment necessarily has to be that much of a hassle. Just use apple/google pay.
Obviously this all could be easy, but it's amazing how many apps fail to make things easy to onboard. When I've just gotten done traveling for hours and just want to get to some place to relax, the last thing I want is to wrangle new account creation in some app where they're trying to be cheap and haven't hired any UX designers to smooth out the process because "you can rebuild Uber with 3 smart devs" as everyone on HN says.
Last time I tried the Lyon app it was barebone and not really up to par with Uber by a large margin. But still.
And they're...okay. I think they all use the same white label technology to make their apps.
The kinds of decisions taken for granted by the hacker news thread/etc are the kinds of decisions that could sink a self funded company, starting with the fact that they are basically writing multiple copies of the application for each platform (ios vs android) and going from there. This is maybe the problem with VC funded companies (like government contracts) the money just keeps flowing independent of all the bad decisions being made. Also, having enough engineers that whole teams can be split off to rewrite the application for no appreciable benefit except to peoples resume's is itself an upper mgmt problem. If the goal was a single unified application I might see how something like that could be justified, but they choose technology stacks that are to native for that to work.
Don’t forget about lawyers, compliance, lobbying, lawsuits, etc., etc. It’s also likely your assumptions are very wrong. Just think about what will happen when people figure out you auto-refund everything below a certain point.
> Fuck marketing
Possibly the bravest thing ever said on HN.
How else would they know which laws in which locales are most profitable to skirt.
I am sure AirBNB has one for the same reason.
Competition. Part of Uber's biggest expenses is "driver and customer incentives" which is corporate speak for bribing people to stay on the app. We know from economics that competition drives down profits and rid hailing services are a dime a dozen.
Uber has an absurd amount of logging & analytics.
If the app was simply drop a pin, get a ride - it wouldn't be that crazy.
Uber has 3.9M drivers world wide. There's probably very rarely more than 1M drivers active at any time. Probably less than 300k people looking for a ride at the vast majority of times.
Assuming you can update the driver's location 1 time per minute - that's ~1.5B requests per day - less than 25k requests per second (including user bookings).
That's like ~2TB of bandwidth per day. That's less than $200 per day. Almost everyone spends more than 5% of their cloud bill on bandwidth. Meaning, the rest of a drastically simplified (but nearly equally useful) Uber could run for ~$4000 per day in server expenses.
That's a $1.4M / year data center. Uber has revenues of >$11B.
They could be making a lot of money. They just aren't because they're spending AT LEAST 50x more on servers and product engineering than they NEED to.
They paid for growth for a long time. They have a monopoly now. There's not a lot of growth left to get. At some point the axe will come down.
Lyft is even worse. They're ~1/3rd the size.
Cost of revenue in just terms of infrastucture was measured in $/ride and minimizing it reliability is difficult, especially if you're a fast growing startup and on cloud providers. Unfortunately, if you're a fast growing startup you also don't usually run bare metal (even though I'm a fan of colo/OVH/Hetzner).
Don't forget that money used to free, with the way interest rates and investors were. It's far more difficult cutting down, than it is to throw money around, obviously.
You can have a fixed rate. Fraud detection is a problem that thousands of apps requiring in app payments solved it somehow.
You don't need to have 1 million features, you only need to have the features demanded by the people who will make you most of the money.
Most people will say they’re going to get an Uber, even if they end up having to use Lyft, no? Ubers, as well as others’, expressed strategy has long been not only first mover, but also monopolization of all aspects of their space, expressly anti-competitive. Part of that is not only being the leader, first to mind, but also draining the enemy/competitor’s resources and undermining their efforts to even challenge you. It’s a total market domination strategy that shouldn’t even be allowed, but they’ve also paid off politicians and captured government in other ways too, so don’t expect anything from there either. There used to be other ride sharing services, I don’t even know if they exist anymore, but even before the Great Monopolization, aka COVID, they were barely scraping by on crumbs in a few local markets while the likes of Uber worked in basically every market, especially in the high spending business travel and entertainment segments.
Fortunately it haven't worked. But could it work assuming the VC investors would be able and willing to pour much more money into it?
100x, bringing their opex to hundreds of millions on billions of revenue. They should still be able to cut the fat and actually turn a handy profit, but they won't.
Riders never see the driver's location before booking.
Uber doesn't need the exact location to get a decent match.
And anyway - they send the ride to several drivers and the lowest / first bidder wins.
So I'm curious if you have any knowledge, or if what you are saying is "I don't need realtime", or perhaps more charitably, "I can't imagine realtime being valuable to users". I push back on the 2nd, and I strong push back on the idea that someone with can reach the conclusion that Uber is wasting money on things that don't drive user value.
Unless, of course, you work/worked there and worked on these projects, and saw firsthand that Uber decided to waste a bunch of money internally.
And it's clear they didn't employ anything like Uber massive work force to do that.
They're in the business of saving gas. Not cruising around idly while they wait for a ride.
Sure - some drivers are finishing a trip nearby. But you know the route they are taking...
You also don't know for-sure what route they're taking. Almost every driver I've had was using Waze or some other mapping app that usually took a different reasonable route than whatever Lyft was showing when I booked it.
Multiply by 3 for redundancy & availability. Multiply by 3 for other endpoints. You're not even 10x...
And this was not Uber as it exists. This is a simplified version of Uber that gives the user a nearly equal experience.
If someone does a mobile or desktop chat app (think Whatsapp or teams) he will use web technologies on the backend. But IRC is very old and still works like a charm without using web technologies.
Likewise, if email was invented today, it would be just calling some REST api instead of using SMTP.
When I worked as a game developer, the server side of our games was a slow and ineficient PHP crap becouse the founders were friends with a PHP programmer. I had to jump through many hoops to mask huge latencies from web server or even the web server not replying at all.
I think using web for anything can add a massive overhead.
I can see the advantage of the web, though. You don't need to implement a server, you just think in terms of HTTP requests and the framework will transform those for you in data and transform the data back into replies.
But if you need something real-time you have to use web sockets, and at that time the simplicity goes away and you can use OS sockets just as good.
I am arguing that using web frameworks for an app which won't have a web fronted adds almost the same kind of overhead as using Javascript for mobile and desktop apps. It's doable, yes. It's the best usage of resources, no.
Normally I should be the last one to complain since I architect and develop web apps for living.
But I do like efficiency, I think that sometimes many layers of virtulization, abstraction, indirection, protocol encapsulation are hurting both the performance and the speed of the development.
And I also know how to develop a server side app without using an HTTP server and a web framework.
I wonder if the root of the problem is that apart from the the people doing mobile apps, the rest are mainly learning a frontend or backend framework and they can't do or are not willing to do anything besides that. Or, even worse, new developers are just learning dynamic laguages such as Javascript or Python or PHP which catters mostly to the Web.
If you only have a hammer, everything looks like a nail.
When I had to design a tool to masquerade the real people data in our microservice based app, everyone from my team was amazed I wrote a console app instead of a Web based application. But since there it wasn't a need for that tool to be called through an API, why should I have made it a web app? Just to use curl instead of command line arguments?
Uber is a monopoly that already has all the drivers and spent $10B+ to aquire them - plus all the riders - they're already trained to open that app - already have it downloaded. No one can compete with that without spending $1B+ - at which point - you'd need a 50% margin for years just to return the value. Maybe I'm uncreative, but I cannot think of a worse business to enter.
It's equally amusing to me that you seem to assume Uber is efficient just because it's a public company.
Sure, GE and IBM don't waste money either.
A smaller and leaner model could work at small scale by focusing on via drivers focusing on their returns with word of mouth marketing.
If you use Uber a bunch, then it is a great way for good drivers to get repeat business privately. A simple platform which is driver-friendly which focused on the whale customers (like myself when I was spending $500/mo+ for commuting) could squeeze the Uber even more. When I started working with a few drivers via text, I was able to save money whilst the driver made more money
Yes, it means you won't get all the shiny quality of life services offered by cloud providers, but you're in a _margin sensitive business_. Deal with it. Optimize every cost.
Also, you don't need expensive engineers re-inventing the most basic things (I know Uber had a huge not-invented-here syndrome). Use the boring tools for the job. Only reinvent what is necessary. You don't need engineers practicing resume-driven-development.
Now this is not about self driving or whatever else, it's about a ride hailing app. The point is there is no reason ride hailing can't be profitable.
Now, using any company to prop up r&d and investor hype for a moonshot, that's a whole other idea...
I think if the tech investment bubble bursts (if it is a bubble) then this makes it easier for Uber in terms of competition as ride prices will tend towards the original "Taxi" prices and people will have to get used to that. And then Uber can make their profit (and also the drivers can make a reasonable living).
A big chunk of that $15 goes to driver incentive promotions. Not the steady state cost, but the sign bonuses and what not. Given high driver churn, this is always a big line item.
Google first result says just 3% of drivers drive for them for more than a year.
And no, you can’t just automate support. Refunds are not the only issue.
Most of their markets are terrible markets.
You flipped the values. The driver is making a bunch, and that's per ride, not per hour
Do you have an MBA or other similar business experience? Because you can’t “fuck marketing.” Without marketing you have no business. How are people going to find out about your product?
I think it was because they provided an unique service and word of mouth was enough.
I remember when Gmail was in beta. There was no marketing. I was fortunate to receive an invite and once I get 5 invites, friends were begging me to send them one.
You can write a hugely popular app with great UX, and super scalable backend but if you don't align your marketing parameters you won't make ends meet.
And you are wondering that people who are in business to teach you something tell you how vital is what they are teaching you? All salesmen sell things that are vital for your success, health, well being if you trust what they are telling.
~25% take rate on a ride ($15 average): $3.75 take
Payment processing: 2.5% + 30c = $0.68
Servers / datacenters: $0.20 (for a margin-sensitive business, you should be colo'ing your own servers, or using cheap alternatives like OVH/Hertzner)
Customer support: Automate as much as possible (auto refunds up to a certain point; for lost items, connect directly to driver); assume 1 in 50 rides require manual human support with a $3 cost = $0.06 support cost per ride
Fraud/refunds: Assume a 2% fraud rate that cannot be reclaimed; thus $0.30 cost for fraud. Refunds for things like driver purposefully took a longer route can be clawed from the driver.
Gross COGS: $1.24
Gross profit: $2.51
What am I missing?? Marketing? Fuck marketing when you can't turn a profit. Everyone knows about Uber or Lyft already, you need to turn a profit, not waste $30 per CAC.