Note: Today’s blog post is much longer than my typical post. If you have feedback about the length, feel free to send me an email.

Much news has been made of Uber’s sky-high valuation. This valuation rests on the assumptions that (a) Uber is much bigger than just taxis, and (b) it is very hard if not impossible for another company to beat them in their market.

What would it take for (b) to be false? In other words, what would an Uber-killer look like? First it’s worth understanding the structural advantages Uber has.

Uber’s primary advantage is the users they have. This aggregation of users who collectively start their trips on Uber means that drivers have little choice if they want to get trips. But similar to messaging apps, this advantage is surprisingly geographical in nature. For messaging apps, this is because people generally talk to people they already know, and people they already know are local. For driving, it’s unlikely that an uber driver will drive 100 miles to find customers, and impossible that they will drive 1000 miles, at least in the short term. This means that even if Uber has many users in a given city, the network effects are limited mainly to that city.

Second, Uber is able to capture users earlier than taxicabs or public transit. For the former, you can get a ride from the comfort of your home, for the latter you need to venture outside before you know if you will be successful or how long you will need to wait. This certainty is valuable. However, it is at risk of users finding rides even earlier in their decision process, for example when they find a restaurant on Google Maps or Facebook.

Third, Uber has benefited from the extreme focus on driving in the 20th century, at least in the US. Their drivers use cars bought for another purpose, on roads built by taxpayers, to destinations that are easily accessed by car, with a density and design that makes transit and walking unpalatable. Yet this isn’t the only model of development, and it’s unclear that cars will enjoy as much focus in the 21st century as they did in the 20th.

Anyways, back to the question. What would an Uber-killer look like? For today I’ll just describe one form it could take. 

An Uber-killer would be branded as “local”, supported directly or indirectly by the city, and would generally only operate in a one or few markets in contrast to Uber’s global presence. The underlying technology would be powered by a white-label supplier. And the costs would be heavily subsidized by advertising, perhaps with a core partnership between a single advertiser and the transportation service. An experiment is already underway in Minneapolis. Notice that 75% of the cost of the ride is subsidized by the city, similar to the subsidy given to many mass transit systems. This is impossible for Uber to match.

Can this scale? Yes, at least for bike shares. One company in Canada provides the technology for the bike shares in DC, NYC, SF, among others. The bike shares are sponsored by large brands like Citibank and Ford. They enjoy special support from the city in the form of dedicated bike lanes and street space for their stations.

There are several opportunities here. For cities to retain their leverage by encouraging the creation of local ridesharing companies. For the white-label provider who can focus on the technology and enjoy dozens or hundreds of large customers. For the rideshare operator who can  compete with Uber in a way that was unthinkable before.

Cities still have leverage over the roads they build and who can use those roads. For example, a city can paint a road red and make it busses only, or paint it green and make it bikes only. It can build rail tunnels and subway stations, spending on infrastructure that Uber currently only dreams of.