Uber In a Box : Open Source & The New Economy

Open source software has played an important role in the development of the Internet and the web. Tools like Linux, MySQL, PHP and countless others dramatically reduced the cost of building new services and companies.

Let’s take a database as an example. It wasn’t that long ago that if you wanted to use a database in your application (any useful application probably interacts with one), you had to license software from a vendor such as Microsoft or Oracle. Database technology is not easy to build, so it was usually cheaper to pay an expensive software license than try to build your own database server in house.

Open source software turned this model on its head. It turned out that a group of volunteer programmers could build and maintain complex projects without the usual economic incentives. Several open source database platforms (PostGRES and MySQL being among the better known) emerged, and became part of the standard set of tools new companies used to build their solutions. Instead of paying a large tithe to Microsoft or Oracle, you could use free software and hire a handful of subject matter experts to support these tools (and pay much less in total).

The current leaders in the “sharing economy” all seem to be betting that by buying market dominance today, they’re insuring they’ll be in a dominant position tomorrow. At the moment, the cost of building a platform like Uber is very high, both because its expensive to create a marketplace, and because these companies have to build the technology behind the marketplace (dispatch, payment systems, etc). That may be true today, but its not going to be true in the future. I’ll focus on car services as an example, though the same points apply to other labor marketplaces.

A car service is a good example of a service that is pretty much the same wherever it is offered. Whether its Uber, a licensed taxi operator, or another car share service, the service has a common set of required features, such as:

Dispatch : ability to request, price and accept/reject service, track status
Feedback : rate drivers and riders (this data feeds into dispatch)
Billing : process and settle payments, resolve disputes
Customer service : track service inquires
A stable and well defined API that allows for different components (such as mobile apps) to be developed fairly independently of each other

This is a situation where an open source solution will do very well. Open source excels in situations where you need a 90–95% solution, meaning it does almost everything you need out of the box, and can be adapted to your specific needs without the original vendor’s involvement. At the moment, the major players build and maintain their own software, which gives them an advantage over new entrants that have to build this as well.

But what happens when someone decides to build a decent open source platform? This could come from any number of places:

A group of independent developers, perhaps motivated by altruistic reasons, wants to create a more level playing field for smaller operators.
One of the second tier operators goes “paws up” and decides as a final act to open source its technology.
A group of smaller operators decide to pool resources and technical talent and end up open sourcing much of their work (similar to what Flywheel is doing, although Flywheel has not open sourced its platform).

Once someone releases a fairly complete and reliable platform, pretty much anyone will be able to compete on a technical level. Building a carshare service will no longer involve recruiting the world’s best technical talent, but rather hiring a much smaller number of people to adapt existing tools (much like you might hire someone to customize an online shopping platform, rather than build the entire platform from scratch). Note that new entrants will still need to recruit and retain high quality labor to their marketplaces, but the availability of a more or less turnkey logistics/marketplace platform will lower technical barriers to entry.

Open source also has an important role to play because of its utility in busting monopolies. In the long term, its not in consumers’ interest for a marketplace, and the services brokered in the marketplace, to be controlled by a single company. Imagine you are a local car service a couple of years from now. Would you rather trust your revenue to a system that you do not own and control? or would you rather have control over your dispatch system, pricing, etc?

This is where my skepticism about sharing economy comes from. At the moment, the leaders enjoy high barriers to entry, which I suspect are more due to the fact that the services are difficult to build, especially for companies that are historically not technical (e.g. taxi operators). That advantage will erode pretty quickly once “Uber in a Box” is available under an open license, as a cottage industry of vendors will emerge to support this sort of solution (think Wordpress for example), and that will enable another wave of entrants. Meanwhile, legal pressure to treat contractors as employees will erode one of their main cost advantages.

Its unwise to extrapolate future growth based on the recent past. Every now and then, a new technology like the telegraph, radio dispatch, or the computer comes along that gives certain operators an advantage, but then access to that technology becomes more or less universal and that advantage dissipates. There is a reason that services like taxis have historically been dominated by smaller local operators (not just in the US but everywhere).

This isn’t to say that the sharing companies are doomed, just that current and future competition may be stronger than expected, and may come from unforeseen places.

And lastly, for the software developers and entrepreneurs who are reading this. Many of you may be thinking that the window of opportunity has closed for this space. If you’re trying to build a vertically integrated business, it probably has.

But what if someone builds a general purpose platform for managing a labor marketplace, something that’s relatively easy to use, applicable to multiple uses, and extensible? Something like Wordpress?