Lyft has completed its Initial Public Offering, and at this writing the price has since fallen 35%. Uber’s IPO is expected soon. Both will now be publicly traded companies, reliant on many people’s judgments about whether they can be good investments. Uber loses billions of US dollars every year, while Lyft, which is smaller but growing faster, is getting close to losing $1 billon/year for the first time.
Why invest in these companies?
Anyone who says “Amazon lost money too at first” is just not thinking about transportation. Amazon can grow more profitable as they grow larger, because they can do things more efficiently at the larger scale.
Uber and Lyft are not like this, because their dominant cost, the driver’s time, is entirely unrelated to the company’s size. For every customer hour there must be a driver hour. Prior to automation, this means that no matter how big these companies get, there is no reason to expect improvement on their bottom line. Any Uber or Lyft driver will tell you that these companies have cut compensation to the bone, and that they already require drivers to pay costs that most other companies would pay themselves, like fuel and maintenance.
If Uber and Lyft could rapidly grow their shared ride products, where your driver picks up other customers while driving you where you’re going, that could change the math. But shared ride services don’t seem to be taking off. My Lyft app rarely offers me the option, even when I’m at a huge destination like an airport, and when they do it isn’t much of a savings, which suggests that it’s not really scaling for them.
Of course Uber and Lyft could also go into another business, such as bike and scooter rental, but in doing that they’re entering an already crowded market with no particular advantage apart from capital. The single-customer ride-hailing is the essence of why these companies exist, and there’s no point in investing in them unless you think that product can succeed.
Please correct me if I’m wrong, but it seems to me the possible universe of reasons someone would invest in these companies is the following:
- Confusion about the basic math of ridehailing, outlined above. Hand-waving comparisons to Amazon are a good sign that this mistake is being made.
- Extreme optimism about Level 5 automation, which would indeed transform the math by eliminating drivers. I no longer hear many people saying that commercial rollout of Level 5, in all situations and weathers, is imminent, as many people believed around the time Uber and Lyft were founded. (And no, it makes no sense to have a huge crew of drivers ready to take the wheel only when the weather looks bad. Nobody can live on that kind of erratic compensation.)
- A naive belief that if you love a product, or find it essential to your own life, it must therefore be a good investment (a rookie investing mistake).
- A belief that while you don’t believe any of those three things, enough other people do that those people will drive the price up, and you can get out before they discover the truth. If this goal were intended clearly and honestly, it would be Ponzi scheme. So surely it can’t be that.
So I must be missing something. What am I missing?
There’s still money to be made turning the customer into the product. TNCs could expand revenues via in-vehicle hyper-targeted advertising to a captive audience (people love this on taxis & transit) or by selling data on people’s movements. Not that either is desirable to the consumer. But perhaps the additional financing available to public companies staves off any VC-funded new entrants that would provide a substitute.
The ads wouldn’t be mandatory. Consumers could pay to remove them. But would allow greater levels of price discrimination for the single-ride product. And just as those who want to select seats in advance pay an extra fee for airline tickets, those who want to turn off the ads could pay fifteen cents per minute, easily bundled into the ride charge.
Selling a passenger’s freedom of movement? That’s already done with information/data?
I don’t think advertising revenue is going make much of dent in the cost of operating a motor vehicle. An ad shown to one person for one minute is not going to generate anywhere near 15 cents of revenue – probably, not even one cent.
Yes, a reasonable idea, except advertising revenue only makes sense when it costs very little to get those advertisers. You just don’t make that much money from online advertising (which is why newspaper publishers struggle). As Jarrett pointed out, each Uber trip costs a lot, even when there are millions of trips happening each day. Google is different, because they scale (each use of Google costs a fraction of a penny). Google may make only a penny on each search, but it costs them way less to provide it.
Of course Amazon is still around, but no one remembers all the losers:
https://en.wikipedia.org/wiki/Dot-com_bubble#Notable_companies
Someone also thought Blue Apron was a good idea.
https://finance.yahoo.com/quote/APRN?p=APRN&.tsrc=fin-srch
They seem to be thinking that once the eliminate driver cost with self driving cars they will be profitable. I think this is severely wrong. Today they have to give drivers a cut. Yes. But they have no capital costs, no maintenance costs, no storage costs because they drives are absorbing all the costs related to car ownership, maintenance, and storage. Get rid of the drivers and suddenly Uber needs to buy, maintain, and store millions of self driving vehicles just to maintain the same footprint. I don’t see it.
Agree. plus (as FT Alphaville wrote yesterday) to be even remotely economically viable Uber et al need to have a stable “peak hour” capable fleet. That also means that they’re not able to handle cases where a lot of people need to be moved more or less at the same time (concerts etc). And of course there’s just no way for human drivers to just wait for Uber to throw a bone their way –it’s just no way to make even pocket money, let alone a living.
The only way that makes any kind of sense is if they can get people who own driverless cars for their own use to release them to Uber when they are not using the car themselves. This could give them access to a very large reserve fleet in the evenings (for concerts, sporting events, etc), but would leave them very short at peak commuter times.
Uber drivers often underestimate how much they are spending on owning, insuring and maintaining their car – lots of people look too much at the fuel cost per mile and not enough at the increase in the cost of ownership caused by the mileage. I suspect that owners who aren’t actually driving the car will be worse at this, not better.
The only way it makes sense is if the whole point of the driverless R&D was to make the whole enterprise look more like a tech company and less like a taxi company, because investors think tech companies are cool and shower them with money, while taxis are boring old stuff and not particularly profitable.
Is Blue Apron having difficulties? After the initial burst, they seem to have settled in to a particular fraction of the market, and there’s no obvious reason why they would fail to be profitable.
“Against stupidity the very gods. Themselves contend in vain”
Technophilia and the lingering glow faith in the rockstar, damn-the-rules CEO?
Seriously, though, if you’ve missed anything, I can’t see it either. I do think there is a financially sustainable market for pre-automation ride-hailing, but it’s much smaller one with typical prices at or near those charged by cab companies.
Taxis have historically been priced in a way that is disconnected from the price of actually operating the vehicle, with the initial pickup under-priced, and mileage, severely overpriced. The estimated cost to drive a typical car one mile is usually estimated at about $0.55, including gas, wear&tear, and everything. The marginal cost to drive each additional mile in a taxi is nearly $3. On the flip side, somebody too who just wants to ride a couple of blocks because they’re too lazy to walk, the entire fare is close to $3, even the driver had to travel a couple miles and/or sit in traffic for 15 minutes to get to the pick-up point.
Uber and Lyft have, for the most part, shifted this with the cost per mile being way cheaper than a taxi, and the initial charge for the pickup being the same as a taxi, if not a little bit higher.
I personally think a fair pricing scheme might result in customers paying more the aggregate, but would require even more disconnect between the distance where the customer is in the car and the fair paid. For example, keeping the current rates for time and mileage, but charging the customer for the time and mileage to drive to the pick-up point, in addition to the time and mileage while the customer is actually in the car. It would also mean that if you end your trip somewhere very remote, you would be charged for the time and mileage for the driver to get back to somewhere where he could reasonably expect to pick up another fare.
The other issue with current Uber/Lyft pricing is that the pooled rides seem far to cheap, based on the amount of actual ridesharing that happens, and probably represent the bulk of the investor ride subsidies. At times, I’ve ridden “pool” rides, whose pre-tip cost barely exceeded the cost of just gas and bridge tolls – without even being matched. I’ve also had “pool” rides where I ride 9 out of 10 miles alone, then pick up another passenger right before my destination. Obviously, discounts on the order of 50-60% for “pooled” rides like these are not fiscally sustainable. I have a strong suspicion that the solo rides may actually be coming close to breaking even, and it’s the pooled rides that are getting almost all of the investor subsidy. But, as that information is locked up behind the corporate firewall, there is no way to know for sure.
Uber could pivot/muscle into MaaS platform? If subscription based transit can become a business model Uber is in a good place to shoehorn PT subs into their app?
Naked Capitalism has a multipart series entitled Can Uber Ever Deliver?, by Hubert Horan on the finances of Uber, beginning three years ago. Short answer: “No.”
So, in answer to your question: you’re not missing a thing. Uber will never be a profitable company, the IPO is the only hope of the early investors to get out without a total haircut, and it is a Ponzi scheme.
I suspect that most investors are banking on Lvl 5 autonomy. I share the collective skepticism on whether that will actually happen soon enough to justify Lyft’s valuation….but I think that’s what this is. The notion that transportation is going to shift from private automobile ownership to Transportation as a Service within a decade, and that Lyft and Uber have a leg up on other companies.
Apart from that, I think there’s a belief that there’s still some juice in the lemon of taxicab oligopolies (ie. that Lyft and Uber could eventually raise prices and still be cheaper than taxis), as they absorb the value out of medallions. They’re also eventually going to come for public transit subsidies for coverage routes in second-tier cities, where transit agencies are already subsidizing at the rate of several dollars per trip. Note that since autonomy will destroy the economics of coverage route transit due to safety concerns, leaving a big opportunity for car service to pick up some of those subsidies.
But at the end of the day, it’s a bet on TaaS/Maas in an autonomous world.
I think it’s absolutely that rookie mistake thing — people use Uber and Lyft, they like using them, the app truly is lifechanging (particularly for people who live in areas where you can call a taxi and it never shows, like Hyde Park in Chicago) and so they feel that this thing has to be profitable. Somehow…
“areas where you can call a taxi and it never shows, like Hyde Park in Chicago”—as someone currently living there, I’ve never had trouble calling a cab.
More generally, is this a real problem? My impression is that some places are harder to *hail* a cab than others, simply because local taxi use is low, but that, no matter where you are, if you call the dispatcher, you’ll get a cab.
All of the above, plus perhaps a believe that they’ll be able to radically increase prices once they’ve built a monopoly.
Which can’t happen. You can lock markets with very high entry costs, but anyone can be a taxi for a very low entry cost.
“For every customer hour there must be a driver hour. Prior to automation, this means that no matter how big these companies get, there is no reason to expect improvement on their bottom line.”
Surely not, right? Just because the majority of cost is currently labor, that does not mean a transition to majority capital/maintenance cost will translate into scale efficiency for ride-sharing. The individualized transport model guarantees that for each new ride, the cost will rise proportionally. All that automation does is increase safety and change the total cost of operating service, as opposed to changing the relative efficiency of modes with different spatial efficiency.
One other thing: I may be understanding this wrong but what is your reasoning that peak only services will be more plentiful under automation? If the marginal difference between owning a vehicle and running it and owning a vehicle and not running it is small then doesn’t this mean vehicles run more often during the day? I am not familiar with the cost structures but I feel like there is potentially some need of taking all the automation theory you have mentioned here and there into something more elaborate.
One of the big fallacies in the transport game is that peak service is the gold mine and amid-peaks time is a money loser. In fact, it’s the exact other way around. A service that meets peak demand only has coverage of payment for capital and operations for four to six hours per day. A service with consistent vehicle usage for most of the day has 19 hours to cover the cost of capital and to pay for operations, a much more sustainable model.
In addition, many of the rosily optimistic projections involving AVs almost solely depend on the cost of travel. However, the cost of travel is a complicated question, and actually has very little to do with mode choice decisions in many circumstances.
Another major fallacy is that AV-equipped TNCs are a first mile/last mile solution. In fact, first mile/last mile is almost never a primary choice, as it infers a transfer between modes that is almost always a time waster. Transfers are a necessary evil in the transit world, but are avoided as much as possible, because there’s a big penalty in the overall utility of the trip if even one transfer is involved, no matter what the mode. With TNCs, I really think that first mile/last mile, is just a hip, non-threatening way of saying “we want the whole trip.”
why would I put trust in a person who is an instant taxi driver? A TriMet driver goes through many hours of testing and background checks.
Hype
I find it really interesting that a lot of the “problems” of transit (like a transfer penalty, or the cost of paying a human bus/train driver, for relevance in the discussion above) don’t really exist in places where transit is done right, like Japan… e.g. the service runs on-time, is very frequent, the cost is reasonable, and it goes exactly where you need it to go — thus, tons of people use it so it becomes highly profitable (most mass transit in Japan is privately owned!)… I don’t understand how these things are such a mystery in the rest of the world? These types of issues really only come up when transit is done poorly, right? I mean seriously, who wants to ride a bus that comes only once an hour? Or a train that can’t be relied upon to arrive at its scheduled time?
Some other things to consider:
– For younger people from the suburbs who need a car to live as an independent adult (and attract a mate), but don’t necessarily have a steady career situation, it may help to generate a little side income and write off the enormous costs of auto-dependence as a business expense.
– Across the transportation industry, and especially in the case of Uber & Lyft, it seems driver costs can only go up. In transit (and trucking) it even seems there is a driver shortage. Yet compared to 5am pull-outs and 12-hour split shifts… don’t underestimate the value of a job with flexible hours that lets you fit in errands on travel you are being paid for. There is lots of opportunity to “take advantage” of drivers by paying them a lower wage.
– Regarding “A naive belief that if you love a product, or find it essential to your own life, it must, therefore, be a good investment”… (For urban-dwellers also seeking to avoid the enormous costs of auto dependence,) Those who use the service would naturally assume they are paying the full cost. That’s how markets are *supposed* to work, right?
– I think these companies are banking on extensive subsidies by governments & transit agencies… who are equally duped by the currently low (privately subsidized… for now) prices.
– Strong Towns Charles Mahron talks about the Growth Ponzi scheme… Of course people will buy into Uber/Lyft if it allows them to believe an auto-dependent development pattern could work out ok. If/when the system collapses, buying a few shares of Lyft will be a small thing to regret. Which leads me to…
– There’s always a market for taking advantage of the poor. As cities become more expensive and poverty moves increasingly to the suburbs, there are probably more and more people turning to these services because they have few other options for transportation. Market prices probably limit demand more than subsidized public (para)transit, but some base level of demand will always be there, for which private sector can extract most of the value they create. Perhaps Uber & Lyft will become more like the payday loans of transportation.
I believe both Uber and Lyft have already ventured into the bike & scooter rentals. Uber owns Jump (https://jump.com) and Lyft already has scooters (https://www.lyft.com/scooters). It feels like they’re desperately trying to diversify in attempt to reduce some of those losses–it remains to be seen how successful those will be, but there’s certainly potential upside form an investor perspective.
Some of what the investors are responding to is the data implications of being able to uniquely identify uber/lyft/ride-hail passengers and correlate that with other marketing data. Obviously Facebook is doing a crazy amount of that kind of work and it appears that Wall Street has appetite for more of this kind of work. I’m critical of the implications of creating marketing avatars of everyone who uses an internet-based service; however, it seems like Wall Street believes in the profitability of that venture.
Re-visiting this post today, in light of Uber’s rather dismal IPO.
I have no idea why investors would be jumping at the opportunity to invest in Uber/Lyft, (in reality, they don’t appear to be, given the initial public stock performance of both firms), particularly in light of their ongoing financial losses. It’s even more interesting when Lyft essentially admitted in their prospectus to investors that they may never make money off their traditional business….ever.
I can’t explain the desire to invest in these firms, but have some theories on what the two firms are hoping for the future:
1. MaaS/Taas play, with large established subscriber network – Assuming that we eventually get to an automated fleet at some point in the future, the business dynamic for how you pay for local transport may change to a cost per mile/extreme short term lease dynamic, rather than the purchase of a fixed asset that sits in the garage 22-hours/day. If that switch were to happen quickly (highly unlikely), than the company with the largest subscriber network has a distinct advantage and is “valuable”? (“valuable”, somehow assuming that a MaaS/TaaS model is profitable, I guess?). How many years of $1B losses investors are willing to endure in the meantime?…not sure, but VC’s are clearly not willing to cover the tab any longer.
2. If MaaS/TaaS does play out, take over target by vehicle manufacturers – Assuming significant automation occurs and the market for local transport switches to short term vehicle leasing (ie: pay-by-the-minute), than traditional vehicle manufacturers theoretically lose a large portion of their traditional customers. Both Uber/Lyft (with their subscriber networks) become an attractive takeover target. This thought helps explain the stakes that some of the vehicle manufacturers have taken in Uber/Lyft….it’s their insurance policy.
3. Cost offloading, from Private to Public sector – Assuming these firms are effective at “partnering” with transit operators for last-mile/coverage type service contracts, they presumably could eventually make the argument that they should be eligible for federal/state/local funding to help support their service offerings, or risk having to “cut service” to these areas. VC’s have been happy to subsidize firm losses for a while, the market will do so for a while too, but eventually watch out for the slow transition of cost/subsidy to the public sector for offering these “transit lite” services.
So I’m the boat with Albaby – probably MaaS/TaaS optimism, with some betting on them being a potential takeover target, if their automation tech is really at the forefront.
If they are waiting on AV what’s to stop a competitor from gobbling up that market. Especially a competitor who does not depend on the taxi business as their primary source of income.
Tesla will likely eat their lunch with AV. May even beat them to the AV punch. Tesla could easily undercut them.
Then in countries with real governments the transit agencies themselves could get into AVs for coverage and last mile trips.
Tesla’s plan is to use its own customers as taxi providers. Which shifts the capital costs to their customers. How can Uber counter that? Waymo(Google), the other main likely winner in the Level 5 race, would just own their fleet and flodthe market when ready.
I don’t think anyone sane would deny the ability of Google or Elon Musk(founder of Paypal) to release a software working well enough to do the web interface for customers. If they’ve got the vehicles as well, why would they cooperate with Uber or Lyft? And in this domain, having the best software is not enough – you need those level 5 cars as well. Uber & Lyft are far beyond in the Level 5 race.
You are missing one thing. Even if Level 5 automation would be available, there still would not be economies of scale in the taxi industry. Replacing the driver (labor costs) would just lower down the costs to a certain percentage related to the miles driven. The span between revenue and profit would rise for a single car but there would be no upscaling. You still have the problem of variable costs. So a single ride must cost “something” and can not sink in price related to the upscaling.