Wondering About Uber’s Financial Model

(Cross-posted and revised from Facebook)

Kevin Werbach
Mar 13, 2017 · 6 min read

Has anyone seen a good recent financial analysis of Uber? I’m sure private equity investors and Wall Street analysts have built models to evaluate whether to invest (now in private offerings, or later after the IPO). Is there anything publicly available?

Everyone points to Uber’s huge fund-raising numbers and rapid growth. But as the old saw goes, if you’re losing money, you can’t make it up on volume.

For the moment I’m not asking whether Uber is good or bad for society. Nor am I evaluating the economics of ride-hailing or the sharing economy in the abstract. I’m just considering Uber’s business prospects, based on what we know of its finances.

We know how Uber generates revenue: Riders pay for rides. What are its costs? Obviously the company is private, so we don’t have all the data, but it’s possible to come up with a rough model from what’s out there. I try to below. I didn’t put too much thought or research into this, so I’m probably missing mis-estimating. Please let me know in the comments. Based on my back-of-the-envelope effort, something doesn’t add up.

It was widely reported from a leaked fund-raising deck that Uber lost $800 million on $1.7 revenue in Q3 2016, and was projecting $5.5 billion in revenue with a loss of $3 billion for the whole year.

So, that’s $8.5 billion of expenses. In one year. Presumably $10-$12 billion in 2017. From what?

The revenue number already excludes the 80% of ride fees that Uber pays out to drivers. It’s just the portion Uber takes for itself. Perhaps there’s a big number on the balance sheet for special promotions, where Uber grants freebies or discounts, or temporarily cuts into its 20% to build market share. But would that be listed as a cost? It’s foregone revenue, not money Uber takes from its coffers and pays back out, I’d assume.

Uber reportedly had 11,000 employees (again, not the drivers) as of February. If one assumes an average salary of $75,000 + benefits, that’s about $1 billion per year. Double that at most. (The tech and business staff make more, but a large number are probably customer service folks.) Let’s guess $1,000/worker/month for office space — Bay Area rent ain’t cheap, but a lot of the Uber workforce must be in the cities where they operate. That’s another $150 million or so. It’s spending $250 million on new office space, which we’ll throw in even though it’s a bit of double counting. Marketing? Uber is doing more TV, radio, and online ads to recruit drivers, but it’s not exactly Proctor & Gamble. Call that $500 million/year, generously. Lobbying? It’s tiny number in this context, a few million a year.

Uber drones promoting its carpool service to drivers stuck in traffic in Mexico.

I’m still only up to $3 billion in costs, nowhere near the $8.5 billion last year or >$10 billion that must be projected for 2017. Uber provides an insurance policy for riders and does screening for drivers — how much are those costs? Uber charges a $1 “safe ride fee” (actually $2.50 now in some places) and did 2 billion rides last year . So let’s there’s no margin on the fee, and give Uber another $2 billion in costs. And then there’s cloud infrastructure. Uber runs its own data centers, but is reportedly switching to a public cloud provider like Amazon or Microsoft. Those costs could be a healthy number, but Uber doesn’t need to handle billions of videos like Facebook or Google. Hard to see it spending more than $1 billion/year on data centers.

R&D on autonomous cars? If we’re talking about above and beyond the people and office space for its research facilities (which are already counted) could that be more than a few hundred million per year? Throw in internal IT costs of $10,000-$20,000 per employee, and we’ll round up those two numbers to half a billion.

OK, that gets me to $6.5 billion of Uber’s $8.5 billion in 2016 costs. Is the rest stock options? That would be about $200,000 per employee on the balance sheet, or roughly 3% of Uber’s equity given away in one year. Seems high. And it’s not clear if stock options are even included in the $3 billion loss for 2016. We don’t know if that’s a GAAP or non-GAAP number.

So my first question from this exercise is, what’s missing or mistaken?

My second question is, what percentage of these costs will scale more slowly than Uber’s future revenues?

Companies like Facebook and Google and Apple have tremendous leverage — their incremental users and advertising revenues over time are much higher than the incremental employees and infrastructure to support them. I have a hard time seeing that in the Uber numbers above. Its costs seem to scale relatively linearly with its growth in rides.

Now, it may be that Uber’s revenue will scale more rapidly with future ride growth. Either because it raises rates, cuts driver payments, or finds major incremental revenue sources from its non-passenger delivery services. And the shift to autonomous cars would be a huge change as well. That’s going to take many years to scale up to hundreds of thousands of vehicles, even optimistically if autonomous vehicles are widely legalized in 2–4 years. It’s also going to come with its own huge costs of owning and operating a fleet, direct liability costs, etc.

I’m not writing this because I want Uber to fail, nor have I mentioned any of its various controversies. And its true that many leading tech companies were unprofitable during their initial period of hypergrowth, and later generated huge regular profits. To do so, however, there need to be fundamental economies of scale or scope. Google gives better answers, and better targeting to advertisers, as it grows. Facebook can do much more with a billion users than it could with 10 million. Amazon has lowered its per-unit costs and improved its value proposition dramatically over the years by plowing potential profits back into the business. I’m having a hard time seeing that for ride-hailing. The example of Austin, Texas, where seven small competitors quickly moved in when Uber and Lyft left reinforces this view.

Ride-hailing may be a terrific business, just not one that inherently lends itself to massive scale. Kind of like the old taxi industry.

Kevin Werbach

Written by

Wharton prof, tech policy maven, digital connector, feminist. Co-author, For the Win; author, The Blockchain and the New Architecture of Trust (coming Nov. 9).