The problem with platforms in the sharing economy

Platforms are considered essential to the rapidly-growing sharing economy.[1]

Mass adoption of the modern smartphone is resulting in vertically integrated firms being replaced by private markets hosted on platforms and populated by a plethora of much smaller firms. In the case of the sharing economy these smaller firm are typically individuals trying to stitch a career together from a collection of gigs.

Conventional wisdom has it that smart money is investing in building scalable platforms. If firms are to be replaced by platforms surrounded by clouds of platform inhabitants, then it is the owners of these platforms that will harvest that lion’s share of value from the sharing economy.

There are a number questions being asked around the appropriateness of this platform-driven approach. Should the supplier on these platforms be considered employees or contractors? Where does healthcare fit in? (Less of a problem in left-leaning countries like Australia with national healthcare schemes, more of a problem in right leaning countries such as the US where healthcare is tied to full-time employment). Does our society want the on-demand jobs created by these platforms?

There is, however, one important question that isn’t being asked.

Are these platforms a necessary part of the sharing economy?

These platforms might be sufficient to kick-start the sharing economy, but they’re not necessary for its long term survival. There are alternative approaches to creating sharing economy solutions that do not rely on a centralised platform.

Platforms solve what we might call the discovery problem. When we’re creating a market  it needs a mechanism for buyers and sellers to discover each other.

One common solution is to address discovery via rendezvous. Buyers and sellers agree to meet at a common location, a place, to enable proximity to simplify the process of matching supply and demand.

Lloyd’s of London, an insurance market, began in the 17th century in Lloyd’s Coffee House; a popular place for sailors, merchants and ship owners. The NY Stock Exchange grew out of a group of speculators that took to meeting under the shade of a buttonwood tree on Wall St. in the 18th century, near the foot of the slave market on the corner of Wall and Pearl Streets. Taxi ranks are also an example of using rendezvous to solve the discovery problem.

One side effect of using a place as a rendezvous is that the owner of the place can extract an economic rent from the market participants. They can charge for access or, even more profitably, take a clip of every transaction that passes through the place.

We can see the same process in action with platforms (virtual places) in the sharing economy. Platform owners – the Ubers, Homejoys, and so on – use their ownership of the rendezvous to take a clip of every transaction that passes through the platform.

There are, however, other ways to solve the discovery problem.

One other solution is for a buyer to broadcast their needs to all suppliers within range. This is what you’re doing when you hail a cab from the street corner. You signal your need (shout “taxi” or wave your arm) and then pick from the suppliers who respond.

The relative advantage of using rendezvous vs. broadcast for discovery depends on effectiveness and efficiency of broadcast vs. point-to-point communication. If there are few suppliers in broadcast range than it will be more effective and efficient to travel to a rendezvous location (either a physical or virtual place) where you can use point-to-point communication. This is why you walk to a taxi rank when you cannot see any cabs nearby. If, on the other hand, there are many suitable suppliers in broadcast range, then it is probably easier to broadcast your need to these suppliers, and not bother to head to a rendezvous point.

For a very long time broadcast technology has been expensive and cumbersome, limiting its use. On the other hand, point-to-point communication has improved in leaps and bounds, from telegraph through telephone to the modern internet connected smartphone. This is what has driven many physical rendezvous locations to move to the virtual world.

Once we could only hail cabs on the street or via taxi ranks. The invention of phones resulted in modern dispatch systems where we use point-to-point communication to talk to a dispatcher who then radios a taxi in range to pick us up. Ride sharing services – such as Uber and Lyft – bring this into the digital age by swapping the phone call and radio for smartphone-hosted apps, streamlining the experience in the process.

However, broadcast technology has started to improve.

One example is IPv6, the most recent version of the Internet Protocol (IP) that underpins the internet. IPv6 includes a few new ways to distributing messages in addition to the traditional unicast (point-to-point) and multicast (one-to-many, where the many must be nearby in network terms). One of the new message distribution approaches is explicit provision for what IPv6 calls geocast. Geocast enables you to send a message to a geographic destination (a geographic point, circle or polygon), with the message delivered to any interested internet server that identifies itself as being in the geographic destination. It is also possible to build geographic-based discovery infrastructure as common services much like email etc. We can see similar developments on a smaller geographic scale with solutions such as iBeacon being used to interact with individuals based on their geography. Mechanisms akin to geocast are also being developed as part of Internet of Things technology suites.

With our newly developed geographic broadcast technology in hand, it is now possible to imagine an alternative to rendezvous platforms.

If the sharing economy service we’re interested in is geographically based – such as taxis – then we can use geographic addressing to discover service providers that are physically close to us.

First we create a message outlining what we want: a taxi now, from where we are (home) to the airport, along with any preferences you might care to express. We would do this via a user agent, a commoditised application that allows us to create and respond to transport messages (much as our email app enables us to create and respond to email messages).

This message is then sent to a geographic address. Any transport dispatch service within the geographic address will receive the message and can choose to respond. This might include a government run dispatch service catering to independent drivers, created to ensure that there is sufficient supply to make the market work. It could include traditional taxi companies who have integrated their existing dispatch systems. It might include the new ride sharing services from the sharing economy. It might even include autonomous cars via car sharing services such as Go Get, Flexicar et al.

Finally, your user agent collects the responses and enables you to pick your preferred option (or it might even automatically pick for you, based on stored preferences). The car turns up and off you go.

Clearly platforms are not essential to the sharing economy.

So why haven’t we already created sharing economy solutions that use these distributed approaches to discovery?

The first explanation is opportunity, as technologies such as IPv6 though iBeacons have only recently started to see reasonable amounts of adoption.

The second explanation is that firms, driven by a profit motive, find it easier to build their businesses on the back of some form of property right that enables them to extract an economic rent. Firms prefer rendezvous solutions as it is easier for them to make more money.

We can see signs of this in the RFC process, which is the process used to create internet standards. All manner of distribute solutions have been created in the past by the RFC process, solutions such as USENET, and even email. However, these solutions dried up once the web emerged and the internet became commercialised. The RFC process was replaced by profit driven innovation that prefers close centralised solutions over open distributed solutions. The few exceptions to this rule – such a BitTorrent – appear to be special cases.

However, this may be changing.

The recent emergence of blockchain – a distributed ledger solution – from the shadow of Bitcoin might be a sign that something has changed in the environment, something that is tipping the advantage away from centralised solutions and toward distributed ones.


[1] The definition of “platform” is somewhat fuzzy in general use. We should also note that the sharing economy is poorly named as it neither involves sharing nor is it a standalone economy, though we’ll set that aside for now.

Want to stay up-to-date?

Stay on trend and in the know when you sign up for our latest content