The "On Demand" economy (elsewhere, I've called it the crowd-scaled economy and it sure isn't accurate to call it the "sharing" economy) is getting lots of attention, and financing, these days. Just yesterday, Uber announced a round valuing the company at $51 Billion (but see here for how inaccurate most of the reporting is around these Unicorn financings).
Like everyone these days, I've used a number of the services. Lately, I've been wondering how each "on demand" marketplace is affected by "leakage", that is, the proclivity of the service provider and the customer to circumvent the marketplace where they met and first transacted in all subsequent interactions. The incentive, of course, is to avoid paying the marketplace fee, so that those savings can be shared by the service provider and the customer in the manner on which they agree. This age-old issue is inherent in any business model in which one party charges for "brokering" a transaction. Once the buyer and seller of anything have transacted once, they have an incentive to conduct future transactions (with each other) off the marketplace or platform.
Does Uber have greater or lesser leakage than the recently failed Homejoy? What are the factors that might affect the degree of leakage?
I have no data on leakage across different verticals, and would love to hear from anyone who has such info.
But, here's a simple armchair theory on how and why it might differ across verticals.
- Predictability of need: I've had Uber drivers offer me their business card and suggest that I call them directly the next time I need a car service. Without confirming or denying that I've taken them up on this offer, the risk of leakage seems to be inversely related to how far in advance the customer knows they'll need the car service. For example, most people have reasonable advance notice of a trip to the airport, so it might make sense to save a little money to arrange a pickup in advance outside of the Uber platform (there's actually a company, Wingz, that is trying to exploit this business model (Note: I have no financial or other interest in Wingz, although I have used its service once)). But, ending a business meeting somewhere "downtown", then needing to get cross-town to another meeting is unpredictable enough that it's not worth the trouble of directly calling a driver who gave you his card. Plus, he's not likely to be in the area, etc.
So, given that some car service trips are planned and some are spur of the moment, I would guess that Uber and other car services have a medium level of "leakage"
- Repeat Usage: For home cleaning services, such as those (until recently) offered by Homejoy, the need for the service is ongoing (i.e., you need your house or apartment cleaned some some periodic basis). Because house cleaning is something of an intimate, trusted service, customers who have good service providers want to maintain the relationship for as long as possible. If the customer likes the service provider they found on the platform, then there's a strong incentive to continue to transact off the platform and save some money, and split the savings somehow. I would think the same factors would apply to other home services to a greater or lesser degree: plumbers, gardeners, dog walkers, etc.
So, given that these kinds of services are repeated and the desire to keep working with a trusted service provider in this more intimate, trusted setting, I would guess that these kinds of platforms have a high rate of "leakage".
- Importance of Novelty: I would guess that most people who use AirBnB or HomeAway for vacation rentals do not usually want to go back to the same place year after year, so are seeking novelty as one of the factors in their choices. In this context, the cost of discovering and renting new places is much higher off the platform than on it.
So, given that vacation home rentals are hard to find off one of the major platforms, and that novelty is likely an important attribute of the choice, I would guess that these kinds of platforms have a low rate of "leakage".
- How Commoditized is the Service: Where the service is, more or less, commoditized (e.g., Uber -- within limits, a car service experience is somewhat commoditized), there is less incentive to expend effort in maintaining the relationship off the market or platform. On the other hand, a home cleaning, or dog walking, or home chef service is not commoditized.
So, I would guess the the more commoditized the service is, the less "leakage" one would expect, and, conversely, the less commoditized (or, put the other way, how important personal preferences are to the continued use of the service) the service, the more "leakage" one would expect.
But, again, I have no data on whether this theory holds any water -- so I'd love to hear from anyone who does. There are, no doubt, other factors that would also play into the dynamics here. Product design/user experience, for example, likely plays a big part. An app like Uber that is so easy to use probably reduces the level of leakage they would suffer if their app weren't as good.