Last two posts on Zoom Video and Etsy, I mentioned network effects. The term has popped up everywhere for tech start-ups and IPOs. In my layman’s opinion, if the network doesn’t provide much pricing power and / or participants have low switching costs, there is not much “network effect”. Case in point is Uber. To be fair, Uber has brought great benefits to consumers with its ease of use. It’s just the economics of its business model that is challenging. A more precise definition of network effect can be found here on Wikipedia.
Uber Business
Uber started as a ride hailing business and has since expanded into food delivery, freight and other businesses. First look at ride hailing. In the initial stage, you do need network effect so you have enough drivers and riders to gain traction and create value. However, once certain scale is reached, the platform runs into constraints. It’s a two-sided platform, or you can call it simply an intermediary, in a way similar to Zoom and Etsy. But worse with real-time and location constraints.
Platform Constraints
As Uber attracts more drivers, soon enough they may exceed the demand of riders. Business will be lower for all drivers, leading to attritions. Drivers may switch to other platforms, taxi cabs, other ride service companies, food delivery or just do other things. On the other hand, if the number of riders out-grows the number of drivers, it will increase wait time and pricing (sometimes sharply). Riders would instead choose other means of transportation.
While this demand-supply balancing act is quite eloquent on its own, it creates constant friction cost to network value creation. What the platform requires is for the two populations of drivers and riders to grow and be sized proportionally to create more value. That’s not an easy task. Demand for rides changes on hourly basis and is based on lots of uncontrollable factors like weather. And assumes supply of drivers is fully flexible, which may not always be the case.
Furthermore, ride hailing only works in localized markets. Yes, you can use the same app. And yes, you have brand recognition. But drivers in one market simply cannot serve riders in another. Thus each local market faces its own network inefficiency as described above.
The same dynamics apply to food delivery, except it’s more complicated, as it deals with 3 population groups: the restaurants, people who deliver, and people who order. Any one group not coordinating with the other two would bring inefficiencies to the system. And it’s even more localized and time constrained with food sitting in transit.
As for the freight business, Uber is competing in the digital freight brokerage market. It has its own set of market dynamics, customers and competitors.
Financial Performance
Uber takes a spread on its ride hailing and food delivery gross booking values as its revenue. Growth in ride hailing has slowed down considerably in 2018 and 2019 to around 20%. 2020 was an abysmal year for ride hailing due to stay-at-home orders. In 2019, ride hailing gross bookings reached $50bn and revenue ~$11bn. But Segment EBITDA was only $2.1bn, which was not enough to cover corporate G&A and platform R&D of $2.5bn. Food delivery averaged 100+% growth over last couple of years, reaching $30bn of gross bookings and $4bn of revenue in 2020. Yet Segment EBITDA continued to be negative at $900mm, before any G&A and R&D.
Looking at the company as a whole, if you account for both cost of revenue and operations and support as part of gross profit, gross margin is around mid-30%. Sales and marketing comes in consistently at ~30% of revenue despite growth. So it basically eats up most, if not all, of gross profits. Then G&A and R&D combines for another 30% of revenue.
Why Is The Business So Unprofitable?
I assume there are much fat to be cut in each bucket. But given the brand recognition that Uber already has, I wonder if the sales and marketing dollars (e.g. driver incentives and customer discounts) are needed to subsidize and sustain the level of ridership and delivery. 2020 was a favorable year for food delivery, yet the segment had deep deficit. The company thinks it can reach break even on food delivery by end of this year, 2021. But it’s still on segment EBITDA level, which is before any G&A and R&D. I attribute this unprofitability to the company’s general lack of pricing power and low switching cost.
First, competition is intense among the similar apps as every one of them is offering essentially the same services and more or less selling at below cost. In ride hailing, this had led to massive consolidation couple years ago. For example in the US, it’s more or less either Uber or Lyft at this point. That said, new structures may still be formed to compete, e.g. Drivers Cooperative in New York City. Internationally, Uber sold itself in various markets in exchange for stakes in the local dominant players, for example Didi in China, Yandex in Russia and Grab in Southeast Asia. In food delivery, the same consolidation phase is happening right now. Uber bought Postmates last year. Just Eat Takeaway acquired GrubHub, which owns Seamless.
However, in my opinion, the industry faces even bigger competition simply from all existing means of transportation, public and private. Emergence of Uber and apps alike forces existing options to improve quality of their services, e.g. development of its own hailing and payment apps for taxi cabs. Or if delivery price has become too high versus the meal, people may just pick up themselves. Given the abundance of alternatives, most riders and eaters are price sensitive. It will be key to see when Uber increases prices or decreases driver incentives and customer discounts, how the demand-supply will react.
Valuation and Outlook
To be fair, I do believe there are select markets that are profitable, groups of customers that are less price sensitive, and business verticals like Uber for Business that are more profitable than others. However, based on current construct and level of revenue, I have a hard time to see EBITDA margin to go much above 10%. There are further cost pressures from drivers demanding higher pay and potential re-classification in various states of drivers from independent contractor to employee status. Recently I came across this blog, The Ride Share Guy, who has much deeper perspective from the driver’s point of view, if interested.
Say with recovery in ride hailing and somewhat continued growth, perhaps one day they can do $20-30bn of revenue. At 10% margin, that’s $2-3bn of EBITDA. As of writing, share price is around $50/share. Enterprise value of $100bn. That’s still 30-50x my future EBITDA. They do have ~$10bn of investments on the balance sheet, primarily stakes in Didi, Yandex and Grab as alluded to earlier. But they all face similar challenges.
In the longer term, there is further existential risk from autonomous driving, which would render all Uber business verticals, from ride hailing, food delivery to freight services, displaceable. Uber had to develop autonomous technology so it may not be left behind. In Dec 2020, the company decided to sell its self-driving unit to Aurora Innovation, while retaining 26% stake in the combined entity. Maybe that’s for the best so they can focus on getting to profitability for the first time since founding in 2009. Challenging outlook, nonetheless.
