Dash.net and circularities of externalities

Dash.net is a start-up that has recently announced VC backing and emerged from stealth. Quoth CrunchGear, "Dash saves drivers significant time and enables them to find and explore new destinations. Dash will not only deliver real-time and relevant information to the car, but also create a network of drivers who help each other avoid traffic and share information about their destinations". There is much uncertainty around the details of the actual technology, but generally it seems to be a "social network of traffic data" (GigaOm). It will apparently transform the driving experience in a way that is similar (the website modestly alludes) to the way in which TV transformed entertainment and the mobile transformed communications.

Paul Kedrosky asks the pertinent question: "if you had a best route from A to B, would you ever tell anyone? Unlike other social networks, traffic is an example of an application where there is little incentive to help anyone". Indeed. Traffic is a classic externality - using the road yourself has the undesirable consequence for other (actual and potential) road-users of marginally diminishing the utility of their road-usage. A system that worked out where the traffic between two points was lightest would render that data obsolete by disseminating it as everyone diverted to the suggested route. Private data in this scenario is rendered valueless by sharing, so a social network becomes impossible.

What if only a few drivers have access to Dash? Well, assuming a small number of users can create useful data about optimal traffic conditions (and that's a big assumption) then a small number of users could genuinely benefit from Dash. The few in the know could take the quickest, least congested route and in fact even the drivers without access to Dash benefit because everyone is moving more quickly. It's an easy win-win.

But that quickly breaks down as more drivers are introduced into the system - because making use of the information Dash sends out about traffic conditions changes traffic conditions more quickly than it generates more data about traffic conditions. Dash tracks the speeds of drivers in the system, so it won't know immediately that so many drivers have taken its advice to divert that the new route is about to become congested too. Add to the model the friction and high switching costs that bedevil driving in the real world, and Dash quickly hits a point at which the impact of additional users has a marginally negative utility. This is, of course, the exact opposite of how a social network operates.

The value of a social network functions by maximising the benefits that additional users bring to existing users. But here additional users introduce not just connections but also externalities, or a dysfunction. Perhaps a social network that aims to minimise externalities by sharing an optimisation strategy with the creators of the externality necessarily fails, at least as a networking model. And perhaps there is a number of users at which optimal traffic data is generated, shared and actioned...but that just raises more questions for Dash (and their backers) about pricing and growth. Once Dash has the optimal number of users in each, say, city, perhaps it can stop selling any additional (units? licenses?) and let an auction system prevail. But that just opens up the market to a second or a third provider not caught in Dash's feedback loop in a way that is, again, diametrically opposed to how real social networks generate their defensibility.