April 06, 2020 - 3:30pm EST by
2020 2021
Price: 25.62 EPS 0 0
Shares Out. (in M): 1,727 P/E 0 0
Market Cap (in $M): 39,416 P/FCF 0 0
Net Debt (in $M): 0 EBIT 0 0
TEV ($): 0 TEV/EBIT 0 0

Sign up for free guest access to view investment idea with a 45 days delay.




What and Who


Let me introduce you to our What, Who, and How Much framework.


Imaging you are in 1998; this new internet thing is complete chaos; it’s the Wild West. Wild West because there are a lot of possibly very profitable and certainly disruptive business models emerging. Chaos because it’s incredibly difficult to see which business models and which companies are going to succeed. So in 1998 you don’t know what and don’t know who.


Fast-forward to ... 2003. At this point we can clearly see that advertising and classifieds dollars will shift online. We know what will happen. But we still don’t really know who – the market is very fragmented and no obvious leader has emerged. Yahoo!, Ask, Bing, AltaVista, MySpace, and Google are still duking it out for market share.


Fast-forward again to 2010. At this point no one has any doubt that digital advertising is the future and analog (or paper, to be more accurate) is a relic of the past. Any ambiguity is removed from what, and the who is Google (and Facebook, but I’ll ignore Facebook here for the sake of simplicity). Google has become a verb; the company dominates search – its biggest competitor at this point is Bing, which has a tiny, shrinking 2.5% market share. People are Googling, not Binging.


Google’s scale gives it an enormous competitive advantage: The more data it has from past searches, the better results it can provide for future searches. Google made some brilliant decisions along the way. Instead of fighting Apple in its own domain, it gave away the Android operating system to hardware manufacturers. Today Android powers half of the mobile phones sold in developed countries and 80% of phones globally.


The problem is that, with the exception of a very few moments in 2010, Google has looked statistically expensive on the basis of current or next-year’ earnings. However, if you looked five or ten years out, you’d have realized that digital advertising would be taking market share from analog and thus would double and then quadruple. Since Google was the who (its dominance was likely to grow), it would capture the bulk of the profit from growth of the search engine advertising market. If you looked at Google’s earnings power through a spyglass instead of a microscope, it was insanely cheap and significantly undervalued, though it would not show up on a single value screen.


Here is what we learned from this: When we see a tsunami of disruptive change (the what), we need to identify the who.


A lot of times it is easier to identify the what than the who. For instance, we spent a lot of time looking at the legalization of online sports betting in the US – it is clear to us that this industry will explode. But we could not identify the who – a company that today we can clearly see will be able to capture a meaningful chunk of the profits in five or ten years. A few years down the road it will be easier for us to see which companies will dominate that space, but we will have paid a price for this insight, because the who will have appreciated a lot by then. At the same time, there will be a lot of companies that looked like reasonable contenders but vanished from the map. So identifying the who is not always easy.


Finally, how do you figure out how much? This is where it gets really difficult, because you can’t build models analyzing these companies with any kind of precision; you need to put your sharp pencil away and take out a crayon. Here is where the first three of our Six Commandments of Value Investing come to the rescue. First, margin of safety: Make sure you buy cheap enough that if the future is not as bright as you anticipate, you don’t lose money. Which brings us to the second commandment: The true risk is not volatility – temporary gyrations – but permanent loss of capital – a stock declining and not coming back. And finally, have a long-term time horizon. When you look at the who in the industry, ask yourself, five years out, what is my most conservative, most uncreative estimate of value? These types of companies have a good chunk of their value not in today’s but in future earnings. Using conservative estimates of revenue and earnings growth and giving them a conservative valuation multiple five years out, you can arrive at a worst-case valuation. If your current price is not much different from this number, you have a margin of safety.


This brings us to ridesharing – the what.


The ridesharing market was created by Uber’s competitor Lyft in 2012. At first, ride sharing was competing with taxis, then with car rentals. But ease of use, instant availability (it often takes just minutes from the time you push the Request Ride button to the car arriving), and affordability expanded the TAM (total available market). Today ridesharing competes with car ownership in a gazillion trillion-dollar global market.


Over the last five years ridesharing gradually slipped into our lifestyle choices. A friend gave up his second car – he and his wife share a car and use Uber when they need to be in different places at once. My son Jonah got his driver’s license a year later than his friends and took Uber when he had to get to movies or friend’s houses when we couldn’t drive him. It was still cheaper than buying him a car and paying for a teenager’s car insurance.


We are no longer asking relatives to take us or pick us from the airport, or for a ride from the mechanic when our car is in the shop. My friends leave their cars at home and take Uber instead when they go out on weekends, so they can drink and not worry about driving. These little anecdotal stories are not unique to me, my family, and my friends; they represent a significant change in our behavior.


There are benefits to not owning a car. You don’t have to worry about car payments, insurance, repairs, maintenance, or looking for and paying for parking. There are deep societal shifts toward urbanization that create a significant tailwind for ride sharing. Owning a car in the city is a headache – there is a very good reason why NYC is the largest ridesharing market in the US. (Our guestimate is that it represents about 20% of all rides.)


Americans used to have a love affair with their cars, but the younger generation prefers experiences over ownership. They’d rather tweet than drive. They are willing to give up owning a car or a house (they’ll rent) in order to enjoy Chipotle burritos or Starbucks cappuccinos.


I don’t expect all consumers to give up their cars and start using Uber instead. But the traditional family with picket white fences, two and a half kids, a dog (in the worst case a cat) is not the majority but a minority of the US population. Two thirds of households in the US don’t include children.


Ridesharing is better than hailing a taxi.


Along with the aforementioned advantages, ridesharing offers a much better quality of service. How can I make such a statement, you may ask? Easy. Unlike with a taxi, the ridesharing passenger and driver get to rate each other. Two years ago, I took one of the world-famous London taxicabs, whose drivers have to go to school for two years. The driver was texting throughout our whole ride and was not particularly friendly.


If an Uber or Lyft driver texts while he drives or is rude to his passengers, this will be quickly reflected in his rating. Uber fires drivers if their rating falls below 4.6 stars out of 5. This feedback loop impacts driver behavior and weeds out bad apples very quickly. Taxis don’t have that. This feedback system is also the reason why we are willing to trust perfect strangers when we buy items on eBay or let strangers stay in our Airbnb houses.


Ridesharing is safer than a taxi. With ridesharing you know who picked you up, and you can share the location and status of your ride with your spouse or friend. I know this is where I’ll get the most pushback, as there have been reports in the news about Uber drivers using their friends’ IDs to drive Ubers.


Each incident with Uber receives greater attention because of its novelty, just as, if a Tesla car bursts into flames in Wichita, it will not be just national, but international news.


But did you read about the Lexus ES of 2011 vintage busting into flames a month ago in Denver? One did, but I bet you a sesame bagel you didn’t hear about it! A friend of a friend’s Lexus was on consignment at the dealer, sitting on the parking lot with other cars. It burst in flames out of nowhere, taking another car with it. The fire department investigation ruled out foul play, determining that the fire came from inside the car. You didn’t read about it because Lexus isn’t Tesla. Lexus lost its novelty a few decades ago and is no longer newsworthy. (Sorry Lexus, I don’t make these rules.)


Tesla and Uber are new kids on the block and thus every little news blip is amplified. However, this amplification creates an even stronger incentive for Uber to make its platform as safe as possible.


There is nothing especially good or bad about people who drive Ubers or taxis. They are people, and thus they come in different shades of good and bad. They all go through similar background checks.


They may not go to school for two years like British taxi drivers – thank god! British cab drivers have to memorize each and every street in that town. I’m sure that was a valuable skill 1970 or even in 2010. But not in 2020. In my unhappy experience taking a London cab, the driver dropped me off in late evening and said, “That house you’re looking for is somewhere between this and that next block.” I had to use my iPhone’s GPS to find the house.


This reminded me an episode of Seinfeld where Kramer decides to give up his watch and use the sun instead to tell time. When asked how he knows the time at night, he mumbles, “Yes, it is more difficult.”


Uber is a platform. Just like any other platform that involves humans, it will have bad actors. Uber, just like any other platform owner, will do anything and everything to keep its platform safe and its reputation intact. Facebook and Twitter have a political ad problem – they are spending billions to fix it. Every company will use technology to fight bad actors. We have no reason to believe they won’t succeed, since their existence depends on it and the solutions are mostly amenable to technology.


Here’s the punchline: Ridesharing is toothpaste that got squeezed out of the tube, and you can’t put it back. It has forever changed our behavior, and this change will continue to snowball far into the future. And the most important part: People love it!


So we’ve figured out the what. People spend trillions of dollars on cars and transportation globally, and ridesharing is roughly a $100 billion industry today (it did not exist nine years ago), so we still have a long way to grow.


Uber – the Who


I admired Uber from afar for a long time as a passenger and an Uber Eats customer. At the same time, I was puzzled why Uber’s losses continued to increase as its revenues skyrocketed. But before we get to Uber’s cost structure let’s talk about its moat – which is enormous. At the core of Uber’s business lies a two-sided network – drivers and passengers (customers). The richer one side of the network is, the more it is valuable to the other side. These networks are incredibly difficult to create and thus extremely difficult to disrupt. The biggest threat to Uber is… we’ll come to it.


Unlike traditional transportation companies, Uber doesn’t own assets. At its core Uber is a switchboard (a software system) that connects passengers and drivers. This is the digital part of Uber, putting it in the category of companies like Facebook, Google, eBay, or Booking.


However, if you look at Uber’s digital counterparts, as their revenues increased their margins (and profitability) exploded, because revenues grew at a faster pace than costs, which turned into fixed costs.


This is not what happened to Uber: The faster it grew the more money it lost – the headline loss number for 2019 was $8.5 billion. Journalists predictably drew parallels between Uber and the dotcom disasters circa 1999 or WeWork circa 2019. And they are partially right… but.


Here is where the analog part of the company comes in. When Facebook enters, let’s say, France, its incremental cost of entering this market is almost zero – it just needs to translate its user agreement into French. Today Uber is in 62 countries and 500 cities, But in every country and city Uber needs to have a physical presence. It has to develop infrastructure to run background checks on drivers and to inspect cars. It also has to create a two-sided market where one did not exist. In some markets Uber has literally had to pay drivers to do nothing, just sit around and wait while it convinced passengers that it was open for business in that country. If this was not hard enough, in almost every city it entered it was going up against the status quo of a taxi monopoly. Thus these new markets came with large legal bills. (Uber is still banned in Germany, the largest country in Europe, and in South Korea. Also, when Uber went public it had large costs in creating a financial reporting system and stock options that were granted in pre-IPO days got vested when Uber went public, inflating headline cost numbers by a few billion. Though the digital part of its business was at scale long time ago, the analog side took years to catch up. This point is paramount.


Finally, the biggest cost to Uber was the irrationality of the market. Low and negative interest rates have flooded the venture capital market with billions in easy money. This has led Uber and its competitors to resort to uneconomic behavior to capture market share. Even if Uber wanted to behave rationally, as long as its competitors and their backers were willing to burn money in attempt to capture market share, Uber had to play ball.


And then came the WeWork moment.


The WeWork blow-up is the best thing that ever happened to Uber and its competitors (really, its competitor, in the singular): The era of unlimited capital is over (at least for now). Venture capital markets and Softbank sent a clear message to their companies: Your business needs to be self-sustaining. Uber and Lyft (its US competitor in ride sharing) started to raise prices, but most importantly they stopped uneconomic behavior and are now focusing on taking out unnecessary costs.


This was when we got really interested in Uber. I listened to Lyft’s earnings call, where management talked about how they started to raise prices and Uber followed. Then they stopped issuing discount coupons and Uber did not, and so they resumed. Lyft’s CEO said that they would be “the follower” not the leader in price setting from now on. I get a feeling that Uber management heard this call, too. There is no reason why Uber and Lyft cannot have a cozy, rational duopoly in the US (Lyft is a US- and Canada-only company).


Similar rational behavior is happening in other markets. Via and June, competitors in NYC, shut down their operations there. Uber sold its Uber Eats business in India, where it was subscale and losing market share to a local rival. (Uber still dominates the rideshare business in India, where it has 57% market share.)


When journalists write today that Uber is losing money, they are doing what most journalists do –looking in the rear-view mirror. To understand this we need to zoom in on each of Uber’s businesses.


Traditional companies in our portfolio go through gradual changes. Uber is different. When you analyze a company like Uber it is almost like you are observing an accelerated maturation process from child to adult. The process is anything but linear and isn’t happening uniformly across the company, because Uber owns businesses at different stages of maturity.


Let’s start with Uber’s core ridesharing business.


Over the last twelve months customers paid $50 billion for Uber rides (this is called gross bookings). Out of that, about $40 billion went to drivers. Uber charged a 20-25% fee for its platform, connecting passengers and drivers, and Uber recognized that $10 billion as revenue. This business generates about $2 billion of cash flows.


This business is finally at scale. Let me explain what this means financially.


In our models we assume that ridesharing will grow 15% a year (it grew 20% last quarter) and incremental margins will be 50% (they were 80% last quarter). In other words, in five years Uber’s ridesharing business will have $20 billion in revenue and earn $5 billion of incremental cash flows (50% of the $10 billion in additional revenues), putting its total cash flows at $7 billion.


In this business Uber’s competitive advantage will be solidified by data and scale. Driver earnings (revenues) per hour really depend two factors: per-hour/per-mile rate and utilization (how much time the driver actually spends driving passengers). Due to its size, Uber has more data than anyone else on these variables. This data should allow Uber to better match drivers and passengers and thus improve utilization.


Here is an example. If a driver charges $40 per hour but only drives passengers 30 minutes each hour, his gross (before expenses) earnings are $20. If the same driver works every minute of the hour and charges $30 per hour, his gross earnings are $30. To simplify a very complex topic, in the long run AI will allow Uber to use its data to improve the utilization rate, which in turn will result in higher wages for drivers and lower the cost of rides. Scale will allow Uber to spread its R&D and other costs over 62 countries, as well.


Today, instead of fighting for market share, Uber is focusing on customer and driver retention through loyalty programs. This will further freeze market share, as it will discourage both customers and drivers from flipping between Uber and competitors.


Uber Eats


When I moved to the US in 1991, I was pleasantly surprised to learn that food could be delivered to my house. Unfortunately, my choices were limited to pizza and Chinese food. I love both, but one can only have so much sweet & sour chicken. Uber Eats brings the menu of any restaurant within a few miles’ radius directly to your stomach.


Food delivery is transforming the restaurant industry. A new phenomenon has emerged: cloud kitchens, restaurants that are not open for walk-in customers but only exist to prepare food for delivery.


Red Robin’s and Chili’s’ incremental sales will have to compete with cloud kitchens that occupy one twentieth of their space, don’t require customer parking lots, and don’t have to pay servers. I would not want to be in their business. Meanwhile, Uber Eats has signed a deal with Rachel Ray to open dark kitchens.


Uber Eats’ business is less than two years old and, stunningly, it generated $15 billion in bookings in 2019 (customer food orders and fees) and had revenue of $1.3 billion and losses of $1.3 billion. It is still a toddler fighting with other kids for survival.


Uber Eats’ success here will vary market by market. In the US – its largest market – Uber Eats is competing with GrubHub, DoorDash, and Postmates. Uber is the second largest player. Size is important in this fight, but so is staying power. Uber has by far the best financial position (with $12 billion of cash and soon its very profitable ridesharing business will be able to support Uber Eats), which is paramount, considering that almost every player in this space is losing money.


The size of the food delivery market is very easy to estimate. Just take the size of the US food market – about $6 trillion – times our laziness, and you get a number in the ballpark of $250 billion in the US.


It is reasonable to assume that after the war of attrition is over this business will go from losing billions of dollars to breaking even to eventually making billions of dollars. The weaker players will either go out of business as they burn through their cash or they’ll be acquired. According to Bill Gurley, a gentleman among venture capitalists and one of the very best early stage investors in the business, DoorDash and Postmates each have less than a year and half of cash left based on their current rate of losses. They’ll have to start making money or raise more capital, which will be difficult because they’ll have to do it at lower valuations than when they raised money in the past.


Either way, this industry will consolidate. However, we are not assigning any value to Uber Eats. In our analysis we are assuming it will either turn breakeven or Uber will shut it down. If it turns profitable it will just be icing on the cake.


Uber Freight


Uber Freight is an infant. Here Uber is going after the enormous freight shipping market (think of containers carried by semi-trailer trucks on the highway, not UPS packages). Uber Freight is trying to bring digitization to a business that is still conducted over the phone and fax machine. Imagine if you were playing the modern version of “rock, paper, scissors”: “fax, phone call, app.” I’d bet on app every time.


Truck drivers can look on the Uber Freight app and choose which load they want to carry and where. This market is incredibly fragmented: The top 50 players account for 38% of the market. The market is dominated by mom & pop (mostly pop) operations with one to three trucks. Uber will have competitors in this market, but it has an opportunity to dominate – after all, few companies know real-time logistics better than Uber; and just as in ridesharing and food delivery, companies need to have the capacity to suffer – to lose money for years before turning a profit.


And then Uber has embryo businesses – self-driving flying cars (yes, you read that right). We assume that in five years this business will either break even or be shut down . We assign no value to it.  Most likely it will be sold or merged with another entity.


And then there are businesses that have not been conceived yet. I have a close friend who lives in Chicago and whose daughter is a freshman at CU Boulder. They went skiing in Vail and then left her skies at my house on their way to the airport. A few weeks later she needed them. I called Uber, loaded her skis. Forty-five minutes, $53, and 40 miles later she had her skis. I basically used Uber ridesharing as a courier service. That Uber service doesn’t exist yet, but I don’t see any reason it won’t in the future. There is a lot of upside option value in what Uber can do that is not priced in.


How Much?


When we model Uber’s earnings five years out using a crayon, our worst-case estimates are around $2.5-3 per share. If these crayon earnings come through then we’ll make money, but not enough to justify buying the stock. Most importantly, though, we won’t lose money. Uber management has guided that the whole company will be cash-flow positive a quarter sooner than expected, at the end of 2020.


If we let our crayon go wild, there is a lot of positive optionality in Uber’s future. Uber’s growth rate may actually accelerate; the incremental margin may end up being higher than we forecasted; Uber Eats and Uber Freight could turn profitable and start making billions of dollars. Uber also owns 15% of Didi, the leading ridesharing company in China, a share that today is worth $10 billion. Five years from now it may end up being worth a lot more. At today’s price we are not paying for that optionality.


Elephants in the Room


Before we bought the stock, we had to get comfortable with an important issue: Do Uber drivers make enough money? Do they make a fair wage?


It is important to understand that the drivers don’t work for Uber (though this has been disputed by the State of California). The bulk (75-80%) of what passengers pay is paid out directly to the drivers; the rest is Uber’s revenue, which also goes to pay for the insuring of both passengers and drivers. Though fares are officially set by Uber, in reality they are driven by the supply (drivers on the road) and demand (number of passengers requesting rides) dynamics of the market. This is why fares fluctuate from market to market.


How much do drivers make?


One study found that drivers’ share of fares varies from $5 per hour in Akron, OH to $25 per hour in Hawaii, with median in the US being around $15 per hour. However, these are gross numbers before all expenses – car depreciation, gasoline, taxes, insurance, repairs.


Here is the crux of the issue: If drivers don’t make enough money after expenses, then the ridesharing model is not sustainable. There is an argument that drivers don’t properly account for all of their expenses, since most of their expenses (other than gasoline) will occur in the future, and thus they are earning less than they think.


Anyone who lived through the financial crisis or who has been involved in the stock market knows that there is plenty of financial illiteracy and irrationality out there. With that in mind, we’ll note that about a million people in the US and four million globally make a decision on a daily basis to spend their time driving for Uber. It is easy to start driving for Uber, and it is just as easy to quit – in fact it is easier than quitting a job: You just stop driving. Barriers to entry and exit are very low. Driving for Uber is not a big emotional decision like buying a house; it’s very easy to unwind.


There is a feedback loop: How much money do you, the driver, have at the end of the month or year, and does that result keep you driving? About a quarter of Uber drivers have been driving for longer than a year, and one fifth for longer than two years.


I suspect that nationwide Uber drivers’ earnings are close to minimum wage, and in some areas like Akron, OH, they may be below. This begs the question, is it a fair wage? The answer might get political very quickly, so let’s add this question: How much should a person who works at McDonalds make? The great thing about capitalism is that the invisible hand answers this question for us. Hate going economist on you here, but it’s all about supply and demand. The pool of people that can follow basic instructions is incredibly deep, and thus these people will (and should) make less than an engineer who has invested time and money in his education and is thus in shorter supply.


Considering that 87% of the people in the United States over the age of 16 have a driver’s license, an Uber driver’s skill set is not much different from that of a person who works at McDonalds. Now, given a choice, would you rather work for McDonalds or drive Uber? Driving for Uber you get to choose your own hours, work as much or as little as you want, and you don’t have a boss. I bet most people would choose Uber. Thus there is a nonfinancial value in driving Uber that makes this vocation more desirable than flipping burgers, thus creating an even greater supply of labor that is willing to work at minimum (or below) wage. For the majority of people driving Uber it’s not their main but rather a supplemental income: According to Uber, 80% of drivers work fewer than 35 hours a week.


Again, just as a reminder, the price of fares is set by the free market (supply and demand) and drivers get to keep most of the earnings. Uber, for its part, provides an ever-improving technology that matches drivers and passengers, insures both drivers and passengers during rides, provides customer services, and earns a profit.


Are Uber drivers workers or contractors? This is a complex issue. California’s new law AB5 argues that they are employees. I am not going to write what I really want to say about AB5, because it will quickly spill into a political discussion, which I avoid like the plague (or coronavirus). But AB5 is going to cripple a lot of other businesses and create significantly frictional employment costs. I suspect the law will not exist in its current form in the future. But here is how we pragmatically look at it: People love ridesharing. It is not going away. Drivers want to set their own hours; they don’t want to be employees as that law forces them to be. Despite the law, neither California nor any other state wants Uber or ridesharing to die.


Law-making usually lags behind technology. Here is an example. After the Great Depression, states passed laws prohibiting auto manufacturers from selling cars directly to consumers. Tesla fought this law. Until January 2020 you could not buy a car directly from Tesla in Michigan – the home of the Big 3. Facing the reality that people wanted to buy cars directly from Tesla and that a nearly one-hundred-year-old law had lost its relevancy and lagged today’s economic reality, Michigan changed that law.


No one, other than the taxi union, wants Lyft or Uber to leave California; thus we suspect California’s law will morph into one that will satisfy both Uber and its drivers.


Uber’s Management


Despite the negative reputation that Travis Kalanick, Uber’s first CEO, acquired in the media, we think he was instrumental in building Uber to what it is today. Remember, Lyft was the company that came up with the ridesharing idea. The only difference between Lyft and Uber was Travis. Yes, Travis was aggressive – he was a fighter, a skill set that was paramount in Uber’s entering new cities and battling against local taxi monopolies. While we don’t endorse his behavior, this aggressiveness was incredibly important in Uber’s early days. But as the company matured it required a more “gentle” CEO who would soften the corporate culture. In 2017 Travis was replaced by Dara Khosrowshahi, a very capable executive who previously ran Expedia.


The ultimate threat to Uber is the autonomous car. I spent a lot of time analyzing this when I researched Tesla (research is cumulative). We are far, far away from self-driving cars presenting a serious threat to Uber. This technology, both software and hardware, is not there. I drove my Tesla Model 3 on wet roads and my autopilot stopped working because sensors (cameras) got covered with dirt. We won’t see autonomous vehicles in wide use for a long time, except in small, geofenced areas.




Value investing is a lot more than just buying statistically cheap companies that trade at low P/Es.  It is a philosophy, a framework.  It is about buying companies that are undervalued, even if this undervaluation may not be staring in your face today.  It is also availing yourself of positive optionality (significant positive outcomes that are hard to evaluate today) and most importantly not paying for them. 


Uber and Coronavirus.   Uber will be impacted by coronavirus: when people are locked down at home, they don’t need ridesharing.  However, the coronavirus will pass, and we believe, based on our modeling, that even in an adverse scenario that Uber will not only survive, but continue to dominate the ridesharing market long term.  In the meantime, Uber has $12 billion of cash to get it not just through the virus but through nuclear winter.  It has $6 billion of debt with maturities starting in 2023 and extending to 2026. 




I do not hold a position with the issuer such as employment, directorship, or consultancy.
I and/or others I advise hold a material investment in the issuer's securities.


Cash flow breakeven (adjusted EBITDA level) in Q4 2020/FY 2021.

Normalization of losses in Eats; Post-COVID recovery

    show   sort by    
      Back to top