Lyncredible Navigating the tech stack of engineering and management

Uber had no upside

I enjoyed a great read of Ride-Hailing: Is It Sustainable? by Philo from MD&A1. It sent me back to late 2017 when I decided to leave Uber. At the time, many friends and colleagues asked me why. The answer was because I saw no upside in Uber. In this post, I would like to take a detour from the usual engineering management topics and talk about business. I hope it is useful since it is probably more important to decide whom to work for in a career. Please bear in mind this is not investment advice though.

(Edit: this post sparked some inspiring discussions on the Internet, which encouraged me to write a follow-up post. I do suggest reading both to get a full picture.)

Uber stabilized in late 2017

Uber had a turbulent year in 2017. From the Delete Uber campaign in January, to Susan Fowler’s heartbreaking blog post in February. From the unearthing of other harassment and business conduct claims throughout the spring, to the tragical death of Bonnie Kalanick, mother of Founder & CEO Travis Kalanick, over the Memorial Day weekend. The chaos finally came to an end when Dara Khosrowshahi took over as CEO before the Labor Day. Dara brought hope to the company when he projected an 18-month timeline to take Uber public2.

I started interviewing in October, right after Dara started, and gave my notice in December. So it was understandable for my friends and colleagues to ask the “why now” question. A lot of Uber employees left in the chaotic first half of 2017, but things had stabilized over the fall. After all, who did not want the IPO excitement now that there was a time table to go public?

I was also doing well at Uber. I was an established tech lead trusted by my team, my manager, and my organization. I just had a stellar performance review in the spring of 2017. The accompanying stock refresh pushed my yearly compensation package to a level unmatchable from other tech companies3. Why leave?

Profitability was baked in

The answer was because I saw no upside in Uber. I did and still do believe Uber created massive value for the society. It expanded the ride-hailing market by at least an order of magnitude4. It was and still is literally impossible to get a taxi from the Seattle-suburban neighborhood I live in, but now I can get a Uber within 10 minutes just at my fingertips.

Uber was still losing money at the time, but I believed that the Uber business could be sustainable. However, I suspected that Uber’s sky-high valuation had already baked in long-term profitability of a low-margin business in a highly competitive market. Uber would need to deliver new high-margin business at scale to justify any future upside.

One data point was the valuation. I joined Uber in October 2015, when the valuation was $50B at a per-share price of $48.77. When I left Uber in December 2017, the valuation had risen to $72B, but the per-share price stayed flat at $48.77. That meant Uber the business did not create any value in the two-year span from the investors’ perspective. New capital infusion was responsible for all the valuation increase. All existing shareholders, including employees, were diluted as a result5.

Another consideration was Dara’s performance as the CEO of Expedia prior to joining Uber. It appeared that Dara excelled at optimizing operations and doing mergers & acquisitions, but he was not as good at leading product innovations and creating new business opportunities. If that were to hold true for Dara’s tenure at Uber, the best case scenario would be cutting losses and achieving profitability, which again implied limited upside.

These were my primary considerations when deciding to leave in late 2017. While I stand by that decision today, my thinking has certainly evolved over the past four years. The sections below have more.

The missing monopoly profits

There was an argument that Uber should have acquired or could still acquire Lyft to reap monopoly profits. That claim seemed dubious to me. We can leave the regulatory approval question aside, and just think about market dynamics for the moment. Uber didn’t seem to possess any deep moat other than the network effect built up in a two-sided marketplace connecting drivers and riders. That moat was fragile though, as Lyft had shown when they took advantage of Uber’s forgettable 2017 and grew from ~10% to ~30% market share in the US. It was not a stretch to imagine a new player entering the market and undercutting Uber, had Lyft been acquired or driven out of the business by Uber.

The counter argument is why a new investor will still fund a Uber-like business when they know it is a race to the bottom. That is a legitimate question, but I think someone might still do:

  • If Uber were a monopoly and were to raise prices, a sufficiently capitalized competitor would have the incentive to enter the market, compete on prices while offering similar quality of service, and take share from Uber.
  • If Uber, fully aware of the competitive threat, didn’t raise prices despite being a monopoly, it would continue to live on thin margins hence having limited upside.
  • Even in a competitive market where companies race to the bottom, there are still plenty of newcomers every now and then. The airline industry is capital intensive and known for being loss makers to promote their loyalty programs. There are also quite a few commercial airliners on the market.

Another counter argument is Uber’s network effect. The theory goes like more riders lead to higher utilization for drivers, which leads to higher earnings and more drivers on the platform, which then brings down prices and attracts more riders. AirBnB seems to have a strong network effect. It has been successfully fending off competitors so far. Why would Uber not benefit from the virtuous cycle and fight off any potential competitors? Here is a hypothesis:

  • Uber’s network effect is local to metropolitan areas. I believe most Uber trips are taken by local riders and operated by local drivers. A new upstart could focus on just one city, run promotions and pay incentives, and quickly undercut Uber’s business in that city. The capital requirement is manageable, and it could raise more capital and expand to other cities once it has shown some success.
  • AirBnB’s network effect is global. A typical traveler would not go to the same destination for repeat vacations. As a result, it would be hard for new upstarts to compete. They could focus on finding hosts in one hot destination, e.g. Hawaii, but they would still need to run national or global promotion campaigns to win mind share from travelers. Alternatively, they could focus on getting travelers from one source area, e.g. Seattle in winter, but they would need to sign up hosts in all the popular destinations. In other words, AirBnB’s network effect is more robust.

The hypothesis is partially validated when DiDi, Uber’s almighty competitor in China and globally, went public along with its financial data. Didi merged with Kuaidi and acquired Uber’s China business many years ago, but there are no signs of monopoly profits at all. Instead, Didi faced fierce, localized competition from Meituan and iZuche.com after the ride-hailing war in China was declared over in 2016 when Uber exited.

Could Uber lobby for favorable regulations to prevent other companies from even entering the market? It is possible, but it is probably not good for Uber either:

  • Regulations imply cost. They will cut into Uber’s already razor-thin margin. Just look at how hard Uber has been fighting against regulations to treat drivers as employees.
  • Artificial monopolies created by regulations tend to be fragile6. One only needs to look at the market value of New York City taxi medallion, a government-created monopoly that Uber destroyed not long ago.

The underachiever in Uber Eats

Uber Eats was believed to be the next growth engine after rides, and it is indeed a larger business7 than ride-hailing for Uber in 2021 due to the coronavirus pandemic. Nevertheless, it still under-achieved relative to its potential. Uber Eats was on track to be No.1 in the US food delivery category in 2017, but it is now a distant second to DoorDash. What if Uber Eats were to maintain its leading position? Would the food delivery business unlock a lot more upside for Uber?

I would still be quite skeptical in that world. Meituan is probably the most successful business built on the back of food delivery, but its profit driver is the high-margin ads business to rank restaurants in its app8. Amazon employs a similar tactic in its retail listing page to extract margin from third-party merchants. Meituan does the same to restaurants who want to be ranked higher in the app. That ads business requires high-density metro areas with hundreds of selections, which does not work in most of US suburbs9. There are exactly four Asian restaurants that Uber Eats can deliver to me under 30 minutes. It is irrational for any of them to pay Uber in exchange for higher ranking.

The pipedream of self-driving cars

The final argument is that self-driving cars will turn Uber into a cash cow. Uber’s largest expense is driver payouts10, and they can drastically lower prices for rides and still literally print money with self-driving cars. After all, isn’t that why Kalanick acquired Anthony Levandowski’s startup and got into an expensive legal battle with Waymo, Alphabet’s self-driving unit? The payoff has to be extremely high to justify the investment, right?

I am not buying that argument either. First, it was not a given that Uber would be the first company to commercialize its own self-driving technology. Of course this option is now officially dead after Uber sold its Advanced Technologies Group to Aurora. However, even before the deal, the fact that Uber had the largest ride-hailing network did not appear to give it any advantage in developing self-driving technology. Tesla has millions of cars and volunteer users/testers on the road. Waymo has the longest actively-operated fleet of test vehicles. Startups like Cruise and Aurora are focused bets that will either make self-driving work or go out of business. All things considered, it was unlikely for Uber to deliver either the first or the best commercial deployments of self-driving technology.

Now that Uber is no longer developing its own self-driving technology, could it license such technology once it is available and mature? Maybe. Would that unlock a lot more upside? Probably not. Let’s say some other company develops such technology and is ready to scale it. I think they have a lot of options favorable to them and unfavorable to Uber:

  1. License the technology to Uber but ask for steep royalty premiums. Uber would still accept the deal because it would be cheaper than human drivers, but the leftover margin to Uber would be thin.
  2. License the technology to Uber and Uber’s competitors. Capture all the value while the ride-hailing service providers race to bottom. This has happened in the PC industry where Intel and Microsoft captured all the value while the OEMs competed to death11.
  3. Create their own ride-hailing service and compete directly with Uber. They could offer superior experience and much cheaper rides thanks to their self-driving technology.
  4. Partner with an established consumer brand to create a new ride-hailing service and compete with Uber. This is the same as #3, but they could offload the rider relationship to an entity with much better brand recognition, for example Apple, Amazon, Facebook or Google.

Uber definitely has the competitive advantage in scaling the ride-hailing marketplace. In particular, Uber has the operational excellence in managing its fleet of driver partners, but this is irrelevant in the self-driving future.

Will I be proven wrong?

I saw no upside for Uber four years ago. I still see very limited upside for Uber today. That does not mean I am right though. I am writing this to share my reasoning rather than my conclusion. In fact, I will be happily proven wrong if $UBER takes off some day, as I still hold a meaningful portion of my employee equity grant12.

  1. I found the article from the weekend longreads via The Diff by Bryne Hobart

  2. Uber did go public in May 2019, roughly 20-months after Dara became CEO. 

  3. I interviewed at Stripe and Facebook, among other companies. None of them were able to match my yearly total compensation package at Uber based on the then-current valuation for each company. 

  4. To learn how Uber expanded the ride-hailing market, read Ride-Hailing: Is It Sustainable?, the article I linked at the beginning of this post. 

  5. Uber went public at $45 per share and a $82.4B valuation. In other words, it had more dilution at lower price; late-stage investors lost money; most employees were underwater. 

  6. On the other hand, natural monopolies are more robust against regulations, even when the regulations were aimed at the monopolies themselves. GDPR was supposed to contain Facebook and Google. Instead, they became even more unstoppable in the EU because smaller competitors cannot afford the cost of GDPR compliance. 

  7. According to Uber’s Q3 2021 report, Delivery is ~30% more than Mobility in gross bookings, and marginally higher in revenue. 

  8. See online marketing services from Meituan’s 2020 annual report

  9. One primary reason for DoorDash’s success was their focus on suburbs

  10. Technically most driver payouts are counted in gross bookings but excluded from revenue, which means it is not an expense line item. Uber only books its own cut as revenue. 

  11. Remember IBM was the original PC OEM but they cut their loss long ago. Uber’s dominant position in the ride-hailing market does not provide any safeguards. 

  12. This post is not investment advice. I may transact in the securities mentioned in this post without providing any updates here. 

Share via