Uber’s Balance Sheet Says… We’re Not A Tech Company

Uber contends it is a software company. A software firm can elude onerous transport regulations. But more importantly, it can get more investor capital early on— software monopolies have big profit margins, transportation firms do not.

In their infancy, software firms tend to favor user growth over profitability. The marginal cost of adding just one customer is nearly zero; the main cost is in writing and maintaining the software, which does not increase linearly with adding customers. As Mark Suster puts it, an unprofitable software startup may simply be growing their staff ahead of revenue.

Software companies have incredible economies of scale because their marginal cost of serving additional customers is low. While internet software firms have margins of at 24%, transportation firms have net margins of 2–5%.

Growing early can even entail ‘locking in’ customers to your platform and warding off interloping firms.

The problem? Neither of these apply to Uber.

Uber’s cost per ride was some 90–95% of revenue — it had gross profit of $865m on $9,786m in revenue, or an 8.8% gross profit margin. Unlike software, an additional $1 in sales comes at a cost of about ninety cents.

Gross profit margin also doesn’t account for operating expenses like “Operations and Support,” and “General and Administrative.” Those expenses rose 42% quarter over quarter, while gross profit only rose 15%, an alarming diseconomy of scale. This is probably due to increases in say, service quality, rather than strictly a result of more customers.

Uber’s Financials. Source: WSJ

Even under panglossian conditions, Uber’s profit margin would be less than 10%.

For Q2, year 2016 vs 2017, Uber’s revenue doubled, but it’s operating expenses declined only 3 percentage points, or 15%. For Q3, operating expenses declined 5 percentage points, or 22%. Further reductions, beyond say 15%, may prove increasingly difficult, as revenue growth slows.

Non-GAAP. Source: The Information

One ubiquitous technique for estimating costs in manufacturing is learning curves: as the volume of production of something doubles, its unit cost declines by some consistent fraction, for that category of goods. Uber’s reduction of unit operations expense of 15% after a doubling in volume is perfectly normal for… precision manufacturing, like aerospace.

Learning Rates in manufacturing. Source: Strategos Inc

The other problem: Uber has no moat. If Uber charged enough to have profit margins over 10%, a startup or foreign firm would enter the market and slash their profits. Drivers and riders can easily switch to another firm offering higher pay and lower prices, respectively. The Chinese ride-hailing giant Didi Chuxing is already circling Uber’s turf.

Despite having a fraction of the market share, Lyft has had response times similar to Uber for years now, in major markets (based on personal experience in Los Angeles). The promise of a profitable monopoly through “scale” was a cultic corporate mantra, not a sound strategy.

Uber’s Growth Is Slowing

In 2016, Uber’s gross revenue increased 26% from Q2 to Q3. In 2017: 11% over the same period. Annualized, those figures are 150% and 52%. From 2016 to 2017, revenue doubled in the second quarter, but ‘only’ increased 78% in the third quarter.

Non-GAAP. Source: The Information

In 2018, revenue growth rate will slow further, even if ride subsidies are maintained. They won’t be. Uber will cut them to become profitable, further hampering growth.

Driverless Cars Will Hurt, Not Help

Uber has new mantras to separate investors from their money. Instead of “network effects” and “scale, baby, scale,” it’s now autonomous vehicles and flying cars.

Today’s batteries simply aren’t capable enough for a viable flying car. As for AVs, even if they are viable soon, it will be cold comfort to Uber. As of January 2018, there were 49 companies cleared to test AVs in the state of California alone.

If even a fraction of these field a ride-hailing service, the competition will be blistering for Uber. Uber was already dead last (sixth) in AV reliability in 2016. Competition against dozens of rivals, instead of just one, will not help Uber’s bottom line.

Parenthetically, with AVs, it’s even plausible that carmakers could charge riders below cost and still turn a profit. Tesla has promised to sell AVs to customers and let them operate on Tesla’s own ride-hailing service. It could pay car owners a rate above marginal cost of a ride but below the total amortized cost of ownership. With Tesla’s devoted fan base, mass ownership may persist despite being financially irrational.

Uber can’t do this, because it does not sell cars. It will have to compete with vertically integrated rivals like GM.

Perhaps the biggest barrier to entry of new ride-hailing firms currently is driver recruitment — and that will disappear where AVs are viable. Apps that aggregate pricing will quickly channel riders to the cheapest ride available. Carmakers like BMW and GM are already deploying carshare services to use later with AVs.

How could Uber cut costs and prosper?

In its first few years, Uber was fighting for a toehold in the market, and getting the basic app functionality down. By contrast, future progress is in incremental gains through product refinement.

Make Uber safer

Uber spends some 10% of revenue on insurance. Fewer crashes will mean fewer payouts, and less spending on support and administration.

  • Consult with insurers and designers on how to reduce crashes further. Uber already has a division that parses telematics data to enhance safety, and to give feedback to drivers about how safely they’re driving.
  • Pay drivers for time, not speed. Higher speeds increase the frequency and severity of crashes. Uber’s pay structure currently rewards speeding, by having high per-mile rates and low per-minute rates.
  • Reduce drowsy driving. Studies have found drowsy driving to be comparable in danger to drunk driving, yet Uber resists limiting driver hours unless forced to by local authorities. Lyft, for one, forces drivers to take a break after a 14 hour period. Uber could have a “field alacrity” test built into the app that regularly test’s drivers senses after extended driving. 
    Admittedly, regulators will probably need to step in, as Uber has shown no willingness to change voluntarily.
Source: The Information

Equal Service, Equal Price: Eliminate cross-subsidies

Until recently, Uber didn’t charge riders for the ‘drive-to-rider’ phase of a ride. That meant passengers would pay the same price regardless of how much time and gas drivers would have to spend reaching the passenger — a ‘sprawl subsidy’ for riders in areas with lower driver density, implicitly paid by other riders.

  • Uber has begun charging for the ‘drive-to-rider’ phase of a trip if it exceeds 8 minutes. This period is still too long — with driver wages at say $15/hour, that’s $2 on a trip that may pay a driver as little as $2-3. (See here for an excellent look at fees.) Passengers should pay for the ‘drive-to-rider’ phase — this will make low demand areas more rewarding for drivers, and assure availability for passengers.
  • Thankfully, Uber started charging tardy riders after a 2 minute wait, whereas before passengers could be up to 5 minutes late at no extra charge.

Scrap rider subsidies

  • Cheaper rides -> more rider demand -> more drivers needed -> more costly recruiting and marketing required (whose costs were 6% of revenue, from WSJ). Ride subsidies are a vicious price spiral. Surge pricing can mute this somewhat, but price volatility will alienate riders from using it for their daily routine, beyond bar-hopping and airport run — which is precisely Uber’s ambition.
  • Marketing, at 5.8% of gross revenue and 28% of net revenue in Q3 2017, seems too high when Uber is already a household name, though it’s difficult to say without insider knowledge — most Americans still have not used the service once. 
    Ride-hailing firms have focused on costly recruiting of new drivers instead of keeping the existing ones, while even Walmart takes the opposite tack, through higher pay.
Source: Hall & Krueger, 2016. p 16; for third party 2017 data see here
  • In a maturing market without economies of scale, and razor thin margins, demand subsidies do not make sense. The continued existence of rider subsidies is best explained by excess investor capital and a hubristic loathing of losing unprofitable market share to rivals.


Ride-hailing has no doubt improved the experience of taking a cab. The use of smartphones ignited innovation in an otherwise conservative sector. The core product is valuable, one for which people will gladly pay. But some grim economic realities remain — principally the low economies of scale inherent to taxis. Uber has not found a way to repeal these realities. At best, it has used a combination of innovation and deregulation to streamline logistics, eliminate economic rents, and reduce driver wages to achieve moderately lower costs and better service.

Consumer prices will have to rise some 10-20% for firms to attain profitability. (In September 2017, Uber raised per-mile prices 5–7% on UberX in all major California markets, and… hardly anyone noticed.) As for profit margins, the trucking industry is a reasonable benchmark, and it garners a 3–6% profit margin.

Interlopers on the Horizon: Bikeshare, Didi, AVs

While ride-hailing is concentrated among few players, competition from alternatives puts a lid on pricing. Bikeshare in particular is a cheap and rapidly proliferating rival, and already does ~3x the ride volume of Didi Chuxing, at ~60m vs ~20m rides per day. Car ownership is still abundant of course, so driving is cost-competitive where parking is cheap, for sober trips.

A New Rival: Electric Dockless Bikeshare. Source: Limebike

Rival Didi Chuxing, fresh off raising $4 billion in capital and investing in bikeshare, could restart bitter pricing wars in Uber’s strongholds, just as Uber’s subsidies were supposed to taper off.

With the sector awash in competition, capital and change, profits will be meager to nil for the near future, as Uber’s valuation continues to decline to reflect its finances.

Addendum: Is Uber like Amazon? Is Lyft like Uber?

Amazon famously has postponed earning profits until recently, in pursuit of growth… and has made Jeff Bezos the richest man in the world. Perhaps Uber is simply taking a page from Amazon.

After paying for the cost of goods, i.e. drivers and insurance, Uber has less than a 10% gross profit margin. Amazon’s margin, by contrast, keeps growing, to 36% for the latest period available. Uber has very little room to squeeze out a profit even before operating expenses, without increasing the price.

Amazon data from Morningstar

When comparing earnings before income, tax, depreciation and amortization (EBITDA), Uber is writing its own book: -10% versus Amazon’s sustained 

  • Capital Investments & New Markets: Amazon and even Lyft can point to investments for future sales growth (e.g., new supply centers) and entering new markets to explain losses; by contrast, the bulk of Uber’s business occurs in places where they are already established, with a product that is largely the same since it introduced unprofitable shared rides (UberPool in 2014) . While it’s still adding users at a torrid pace, these have done little to change the fundamental lack of an economy of scale.
  • Moat: Amazon has the finest consumer logistics operation in the world, with delivery times markedly faster than its rivals. It has shown an unrivaled ability to enter and dominate new industries in short order. 
    Uber has a service that has already been more or less matched by rivals (Lyft, Didi, Grab, Go-jek).
  • As for Lyft, comparable data was not available to provide a similar analysis. In its favor, Lyft has managed to avoid the legal and governance crises surrounding Uber, and until recently, has resisted making large investments in AVs. With Uber hungry to finally make a profit, Lyft may go along with Uber raising prices. (Parenthetically, it always seemed that Uber was the one initiating price cuts, and Lyft survived thanks to its smaller market share.)
  • Just as Uber and Lyft end their profitless price wars with price increases that enable profits, a well-rested Didi may enter the market to erase profits anew.