Measuring Your Market? Consider Timing.
10 years ago today, NYU finance professor Aswath Damodaran, well-known for his work on corporate valuations, posted an article on Uber. The rideshare company was already a unicorn (a new term at the time), with a $17B valuation.
Yet, Damodaran ran the numbers and came up with a more modest $5.9B valuation. The title of his post, “Uber Isn’t Worth $17 Billion” summed up his thoughts. Lest we forget, many others back then also thought that companies like Uber (as well as other unicorn startups) were overvalued.
A month later, Uber investor and board member Bill Gurley posted a rebuttal. In “How to Miss By a Mile,” Gurley laid out an argument for Uber’s valuation.
But you have to appreciate Damodaran’s analysis. He showed his work and provided his Excel model. Gurley, on the other hand, revealed fewer numbers or didn’t want to share specifics for a then-private company.
To be fair, Damodaran had to rely on public information where Gurley had insider knowledge. But the main difference between the two approaches was on whether the world would remain the same in the future related to local transportation (the Damodaran model) or whether the world of transportation would change in a meaningful way (the Gurley model).
But how could you ever know if the world is ready for change or if things would largely remain the same?
Questions like that involve the role of timing.
Timing Changes The Market Size
As part of his process, Damodaran estimated the market size of the global taxi industry. He summed up measures of the largest taxi markets in the world (Japan, UK, USA) as $50B and doubled that for the rest of the world. This estimate came to $100B/year, with the large markets like the US stable and real growth coming from emerging economies. He averaged annual growth at 6% / year, meaning that the taxi market that he put Uber into was projected to grow to be $183B by 2024.
Among the logical mistakes I believe Damodaran made was in believing that (as he wrote): “Unlike technology companies in other businesses, like Google, Facebook and eBay, the network effect and winner-take-all benefits are limited. Having a global network of tens of thousands of cabs doesn’t make a difference to a customer looking for a cab in New York City. That, along with the regulatory restrictions protecting the status quo and the competition Uber faces from Lyft, Hailo and others, lead me to estimate a market share of 10 percent.”
Let’s look at why these assumptions proved to be incorrect.
Uber didn’t immediately roll out drivers around the world. They intentionally rolled out new geographies when they could manage (and create) supply and demand. There are winner-take-all benefits even for rideshare, whether from the view of passengers (don’t need to find the local rideshare app for each trip) or from the view of the rideshare companies (accrued benefits to the tech, mapping, funding of new market expansion).
Damodaran also overestimated how much power municipalities had to prevent Uber from entering their cities. Uber’s tactic here was to enter without permission, provide service better than the baseline taxi companies, gain passenger support, and then negotiate with municipalities that wanted to force them out.
In his model, Damodaran took the 10 projected years of free cash flows, the estimated terminal value after year 10, assumed that Uber would cap out at 10% of the overall market size, and that it would take 10 years (that would be 2024) for the company to reach that stable state.
Way too small to read as an image. Check out the actual spreadsheet.
When it came to VC Gurley’s rebuttal, he used this short description of why projections can mislead.
Seeing that Tweet, I looked at the markets for horses and cars when doing research for my Why Now book. It tells the story. The market for cars does not cap out at the size of the previous market for horses. Rather, the market for cars becomes much bigger than the market for horses. The market for cars opened up new transportation possibilities.
Gurley focused on some key differences between the legacy taxis and new entrant Uber. With Uber, pick-up times were faster, there was improved “coverage density,” or the ability to grow the service area, there was no need for cash payment, having driver and passenger ratings produced better behavior, and resulted in higher trust and safety.
Gurley claimed that Uber was highly price elastic, meaning that as it lowered its prices, more people would choose it, even choosing it over other options, such as not traveling at all. When Uber would start to strategically alter its pricing, it would benefit from that price elasticity.
Gurley provided other use cases in Uber’s favor, but stressed the potential for the company to be an alternative to owning a car. In the US, few cities have such good public transportation as to eliminate the need for a car if you could afford one. But Uber’s entrance potentially does just that.
Interestingly, Damodaran’s 2014 estimates of what the taxi market size would be in the year 2024 ($183B) wasn’t that far off from more recent estimates ($230B in 2023).
However, the sum of the taxi and ride-hailing markets became about twice what Damodaran estimated for 2024.
Uber’s market share of the combined ridesharing and taxi market ended up being 25% globally, as opposed to Damodaran’s assumed 10%.
Since those previous projections are now history, I used Damodaran’s model and plugged in the actual numbers for 2014 – 2024 for the taxi and ride-hailing markets and increased Uber’s market share to 25%. Writing in 2014, Damodaran obviously didn’t factor in a market decline for Covid. So I went back and smoothed the growth rate for the years 2020 to 2024.
Damodaran’s valuation after those changes? $28.4B, or almost five times his original estimate. Uber’s last valuation before IPO was $80B. Current market cap is $135B.
But what changed the market size? I think about this in terms of timing.
The general process I follow is to look for Timing Drivers (there are 12 I consider), to understand how those Timing Drivers impact the company’s business model, and then to consider what company is best positioned to take advantage of those changes.
Rideshare businesses only because possible around the year 2009 or 2010 when Uber launched. Timing Drivers that enabled their existence included critical mass of smartphones with GPS chips that were fast enough to provide turn-by-turn directions, user comfort with online payments and ratings systems, and gig economy demand. While pieces of those Timing Drivers existed earlier, their combination was essential.
Those Timing Drivers combined to enable a new business model. Namely, a business model where Uber could provide higher value to customers (clarity on when a car will arrive, driver trust, no need to carry cash), revenue for the business (charging passengers’ credit cards), and lower per ride costs (car and fuel costs are pushed to the drivers, while Uber provides the overall system).
Avoid Damodaran’s mismeasurement. Go through what I call a Why Now Session to evaluate the future potential for a new business or evaluating whether it’s the right time for a specific product or service to grow. I describe the process, with frameworks and examples in the book Why Now: How Good Timing Makes Great Products.
