Hacker News Reader

Comment by harryh | Hacker News Reader

Uber passengers only pay 41% of the cost of trips, with investor capital making up the difference

FWIW this assertion (which isn't really core to the central thesis of the post, but still) is wrong. That number comes from

https://ftalphaville.ft.com/2016/12/01/2180647/the-taxi-unic...

but the author of that story misread the data. Uber only counts their cut as revenue not the full cost of the ride.

Despite this repetition (now corrected, thx!) of this incorrect data I find the overall thesis of the post compelling! As a disinterested bystander, it will be interesting to see how it all plays out.

EDIT: It turns out the original 41% statement comes from http://www.nakedcapitalism.com/2016/11/can-uber-ever-deliver... not from the Financial Times. It can be hard to trace these things back sometimes.

View on HN

You know what, I may get downvoted for this, but my take is you just can't stop these guys. They'll do whatever it takes to win. I probably spend around $2,000 USD on uber trips per month. And frankly, as much disturbing stuff I hear about these guys, Uber beats the heck out of renting a car in a snowy area (especially after having taken a red-eye to a client site). It also beats the crap out of taking a cab in most places, I usually have to navigate to where I want to go and the taxis aren't as clean or the drivers as friendly. They could double the price, and as long as my company reimbursed it, I'd still take it, because they've just got it down so well. Lyft and the other competitors suffer from availability issues, and the average time it takes for a driver to get to me on Uber is much less than the other ride shares. They've got their product down pat, it's so well designed. What I don't support however is their workplace attitude towards women (whom I applaud for coming out about it, takes guts.) It's made me rethink using the service a few times, but when you've got a product that good and such a demanding schedule, the (sad?) truth is that you'll take whatever little comforts you can that make your life that much easier.

I too was very reluctant to let go of uber for the same reasons. It's difficult to let go of what is clearly an outstanding product. I also have mixed feelings about Kalanick. I love how bold and aggressive he has been in going to war with ossified privilege and wish that more entrepreneurs had his balls. But when I read Susan’s blog, I thought to myself “this cannot go unpunished.” The toxic culture has to be corrected, not just at uber but everywhere it exists. Creating a world in which our daughters will not have to experience what Susan experienced is far more important than the convenience of a great app. Unfortunately, people do not genuinely change their deeply ingrained habits unless and until they reach a pain threshold which compels them to change their core values. Uber needs to feel that pain right now for the greater good of everyone including uber. So I deleted uber and switched to using Lyft and, so far, the experience has been surprisingly good. It seems that the engineers at Lyft have done a really great job as well. I have found that, as a general rule, making the decision to do what is right even when it is uncomfortable is always the decision that leads to greater happiness.

An outstanding product that seems unsustainable (I dont know, just reusing the average theory about Uber bubble) seems a lot like a drug. It accelerates everything, but it dries up it's own soil in the process.

I hear ya, but your assumptions are predicate on the quality and availability of Uber service remaining static if they increase their price, and I'm not sure if they can do that. (The current seemingly endless spigot of cost free tech investment capital is leading to some interesting unintended consequences).

I'm curious how one would spend $2000 on uber per month? For what group of people would that be typical?

I believe they mentioned business use. They are getting reimbursed by their employer.

At a guess: cross country sales travel. You could easily rack up 2 or 3 fairly long distance Uber trips / day doing that (Airport -> Hotel -> Client -> Airport to travel to the next client)

Consulting

"They could double the price, and as long as my company reimbursed it, I'd still take it"

Mmmkay

Right?

"As long as I don't have to pay, I don't care what the cost is! Wait, why is everyone looking at me all funny?"

The nakedcapitalism post that started the 41% myth actually has enough info to calculate the real number, at least for 1H2015 (http://www.nakedcapitalism.com/2016/11/can-uber-ever-deliver...)

Farebox ratio = fares / total cost of operations

Total cost of operations = fares - profits

Farebox ratio = fares / (fares - profits)

For 1H2015:

Farebox ratio = 3,661 / (3,661 - -987) = 3661 / 4648 = 78.8%

That is a somewhat reasonable answer. The reason that I don't consider it exactly correct is that the original claim was:

"Uber passengers were paying only 41% of the actual cost of their trips."

What you have calculated is that Uber passengers paid 78.8% of the cost of operating the company in 1H2015.

The cost of operating the company isn't the same thing as the cost of all their trips. The company does a lot more than provide trips.

Also, that number must be Uber's total loss distributed per ride, rather than the profitability of each ride?

I can't imagine that Uber loses money on each additional trip. If that was true, the boycotts are helping Uber!

The boycotts might help Uber's profits - by bringing them closer to zero - but that's irrelevant to Uber's goal of growing its user base.

That's probably not how it works. Most of the rides itself will make some profit for Uber. However, Uber has huge fixed costs via free rides, ads, salaries, etc. These will create the loss.

So if you have a coupon and use it, it's quite likely that Uber makes a loss. But for most other rides, the driver will usually get a bit less than what you pay. It's just not enough to cover the company costs.

> However, Uber has huge fixed costs via free rides, ads, salaries, etc. These will create the loss.

Would those be essential at this point? Seems like everyone who wants a ride-sharing app has one, so their marketing expense should be substantially down by now.

Startups typically keep the salary expenses controllable by compensating via equity (which does not have to be reported as expense unless one is doing GAAP), so it seems that it should be under control as well.

Personal experience, so take it with a grain of salt.

I've talked with 2-3 Uber drivers in LA area. They seem to be getting 65% to 75% of what passenger pays. I think many of the core Uber markets are fairly profitable. This would imply that majority of the burn is focused on growth in new markets.

Some potential problems with extrapolating the data:

* Small sample size.

* Uber rides cost more in California, observed 65% to 75% might not hold in other mature markets.

That is my suspicion as well though I will add that in addition to spending to grow in new markets they are also, quite obviously, spending a lot to develop self driving cars (possibly sometimes through nefarious means!). Unclear if you count that as "growth in new markets" or not.

I think you're misinterpreting what the post and the GP meant.

What's at issue here isn't that Uber drivers don't get to take home X% of what the passenger pays, but rather that the cost Uber charges passengers isn't enough to keep the business afloat without a significant amount of VC funding.

I think the point of the comment was as kind of an upper bound. For instance if drivers made 100% (or more) of what the passenger pays, then Uber would be obviously not breaking even. At 65-75% it's feasible that at a certain scale Uber's cut would make them profitable.

Feasible but not likely. From this 25% to 35% Uber will still do this

1. Run their operating business - Maintain data centers - Maintain a global web infrastructure - Maintain the app

2. Run campaigns to compete in markets with strong local competition - Europe (Hailo, Car2Go, etc) - Asia (Didi, Ola, Grab, etc)

3. Finance their ongoing operations - Equity financing expensive - Debt financing hard to get at this stage and cost money too

4. Hire and retain top talent 5. Legal fees and licensing 6. Rentals for global offices 7. Fund and maintain fleet businesses 8. Invest in R&D (self-driving technology & talent) 9. Entertain M&A

So it looks they have a lot of costs on their plate which their operating business can not cover.

A lot of that is about expansion. You don't need to factor it in to a question of whether they can run a viable business on 30%.

Everything in that list, aside from 2 (advertising and expansion), 7 (fleet operations), and 8 (self-driving R&D, which is honestly well outside of their current business) is dirt cheap in comparison to their operating costs - paying drivers, vehicle purchasing, insurance, and maintenance, and buying fuel.

Service and hardware businesses have completely different cost structures as compared to tech companies.

Think about it - standard deduction is about about 50 cents per mile for vehicle costs, and actually having a human in the seat probably costs more than 50 cents per mile ($ per hour divided by average speed), and the average drive is about 5 miles. We're talking well over $5 in wholesale operating costs for each ride. Times literally a million rides per day.

The cost of a few programmers and servers is tiny in comparison.

> vehicle purchasing, insurance, and maintenance, and buying fuel

These are all things Uber does not do. It comes out of the cut the driver takes home, not Uber's revenue.

> self-driving R&D, which is honestly well outside of their current business

According to the CEO, self-driving cars are "existential" to Uber's success. They may need something that disruptive to meet the company's sky high valuation.

as far as it goes for "world domination" I've read before that Uber's plan is to make everyone's (self-driving car) a money making machine for Uber and the car owner.

You go to work at 9am and while at work 9.01am to 4.59pm your car drives around making you both money. The same applies for when you return home 6.01pm to 8am the next day.

No drivers, only Auto-bots ;) and paying customers!

(makes me wonder how much Insurance companies will be making from this setup)(and if they are ready/preparing their numbers for this type of business)

I can only imagine how filthy your car would be when it gets back to you at 5pm.

I don't want to imagine what it'd be like at 8am after giving lifts to people coming home after a night out

It wouldn't be too hard for them to throw in a cleaning mechanism in there somehow.

If you find the car filthy, I'm sure a customer would too, so it's in their interest to fix it and I'm sure it's even easy to automate the cleaning process :-)

Uber seems to offer quite generous bonuses to drivers who complete milestones (e.g. $600 bonus for 100 trips in a week, for instance). Every time I've chatted with a driver in Oakland, it seems that the bonuses are generally in the $5-7 per trip range. I've also been alone for many Uber Pool rides where I am nowhere close to being an economically viable passenger for the driver ($6 for a ride that costs the driver at least half an hour of time and requires more than 10 miles of driving).

Pool is profitable on average, but low margin, so any particular ride is likely to be an expense. Much like how a casino loses money on 49% of slots sessions.

Sliding off topic but I think that's more like 49% of pulls. The amount of time that it takes a slot machine to drive your odds of walking away richer down to hopeless levels is pretty short.

I have only played slots for a few hours, but my lifetime average is up ~3x and a few hundred dollars vs. what I put into the machines. So, I think they are set fairly close to break even as that's from a few dozen minor wins vs a single jackpot. Honestly, if every single time someone played for 3+ hours they where always down then fewer people would come back.

That said, they probably set machines in very high traffic areas to terrible odds to catch as many suckers as possible.

It's likely that the RNG's in slots don't produce uniformly distributed wins and losses except over long timeframes.

A machine that's truly random will result in everyone losing money slowly and reliably after a large number of pulls. So not fun. A machine built to have the odds wander around over a period of time will be more unpredictable with a greater earnings spread

That's not how random works. If odd wandered like that, there would be long time frames where the slots were 'not fun' and would drive people away.

I don't think that slots can be stateful between plays - there are extremely stringent regulations for slot machines and odds. For instance, if you advertise a 99.95% slot giveback, they will test that.

Also, for video poker and card games, the wins and losses are distributed as if it were a real deck of cards shuffled between hands. There are strict regulations there too, and payout percentages based on the rules of the game and the payout values for each type of poker hand.

Well I know that at least progressive jackpots are allowed. https://en.wikipedia.org/wiki/Progressive_jackpot There of course, the state is not hidden.

Yes. I should have amended the statement in that slot machines, the odds of each outcome are stateless, and the payouts for each outcome must be clearly stated.

For progressive Jackpots, the chances of you hitting the Jackpot on two successive pulls are the same as you hitting the jackpot on two pulls far apart. The payouts are different, but as you said, that is not hidden information.

Slot machines are able to use psychological tricks, I believe, such as showing "almost hits" with a higher probability than them actually occurring. "Oh, if that one symbol had just changed, I'd be a millionaire right now!"

not allowed in Nevada. odds on a slot machine can be changed but

State law allows them to change the odds after a machine has been idle for four minutes, and then they must not allow anyone to play the machine for four more minutes. During that time, the screen must indicate a change is being made to the game's configuration, said Travis Foley, laboratory manager for the technology division of the Nevada State Gaming Control Board, who is overseeing the Treasure Island test.

You seem to misunderstand human nature.

(How many people do you suppose play lotteries week in week out never winning back a fraction of what they out in, but dreaming about and planning what they're gonna spend the $10,000,000 jackpot on?)

These machines have a fairly low max payout. Winning 500$ is nice, but hardly a life changing event. Further, the cost is higher, where a lotto might cost 2$ for some hope these machines can burn 20$ fairly quickly.

As to odds. If you have a 1% chance to win 99$ and play 90 times you are likely to be up 9$ and have a lower odds of being down 90$ And even lower odds of being up 108$. Over long enough time frames you still lose money, but it feels very different from losing 0.01 cents per pull.

They probably set those the highest so passerbys see people win most and sit down. That is what I seem to have heard but a case could be made either way

Thanks, I've updated the post.

Can you explain further what was wrong with the calculation? This number is widely reported and yours is the first refutation I've seen.

The number was determined by looking at gross revenues (which is just Uber's cut of rides not the total that customers pay) and comparing it to corporate expenses (which is not the same thing as what it costs to deliver the rides). It's an almost non sensical calculation.

The fact that it has been widely reported is a sad statement on the innumeracy of journalism.

^ to reiterate. This is a bit arbitrary since investor capital is not actually used this way, but the numbers would be:

- 41% of Uber's corporate expenses are paid for by the customer, 59% by investors

- 100% of non-corporate costs (driver, car, fuel, maintenance, etc) are paid for by the customer and 0% by investors.

But this instead somehow got interpreted as "the customer only pays for 41% of the ride cost". Which is completely false.

That is an excellent explanation. Thank you.

I will add that I suspect that some corporate expenses are going to the driver in the form of bonuses or minimum per hour earnings (Uber has had a lot incentive programs for drivers). This, kinda, goes to the cost of a ride. I have no idea how much money this is overall. A lot maybe.

> Uber only counts their cut as revenue not the full cost of the ride.

except that revenue != income. (http://smallbusiness.chron.com/net-sales-revenue-vs-net-inco... )

> Net sales, or net revenue, is the money a company gets from doing business with its customers. Net income is profit -- what's left over after the company has accounted for all its revenue, expenses, gains, losses, taxes and other obligations.

Uber almost certainly counts the revenue. This is a standard silicon valley way of getting a higher VC valuation.

This is important because if the money is flowing through Uber's pockets then Uber has the ability to adjust the diversion of revenue into income. ( and the VCs know this )

Look at the random silicon valley startup marketplace that reports how many millions of dollars they sold. All revenue - but the income may be pitiful.

Uber almost certainly counts the revenue. This is a standard silicon valley way of getting a higher VC valuation.

It does not, as you can plainly see if you look at the nakedcapitalism link. Revenue is a separate, and much smaller, number than total passenger payments.

What is the correct percentage?

AFAIK there is not enough publicly available data to answer this question. But total_corporate_revenue(1) / total_corporate_spend(2) is an almost nonsensical way to calculate that number.

(1) which is not the same thing, at all, as the total amount customers are paying for rides

(2) which is not the same thing, at all, as the cost of providing all uber rides

    > (1) which is not the same thing,
    > at all, as the total amount customers
    > are paying for rides

Why not?

Because they are two different numbers in the report in question. Maybe go look at it?

Any percentage given is very misleading. In nearly all cases, Uber will make a profit when you take a ride. However, they built such an expensive company (huge R&D department, zillions of free coupons, ads everywhere) that they cannot pay for it without making a loss.

If they suddenly had 10x as many rides or would shed coupons and R&D, they'd probably be profitable. But they'd also lose their edge they hope to have over Lyft & Co.

NB: There are of course areas where they offer rides below the driver's cut, but that's probably rare after they exited China.

I know companies subsidize product costs to get larger market share, but subsidizing 59% seems a bit crazy.

As I said, they aren't doing that.

So what's the real number?

AFAIK there is not enough publicly available data to answer this question.

Oh no! It's probably even worse then! /s

Payments cover 79% of the ride cost (if we assume that the ride cost should covers the driver's cut and raises Uber's EBT to zero).

Reply on news.ycombinator.com