Tuesday, October 28, 2008

Part of the New Way

Talking with the silicon valley whiz kids confirmed something that I've been wanting to believe for a while; that we're entering an era where openness about good ideas is recognized and rewarded. I've had a few really good ideas (imho) that I've held back from writing about for fear of seeing them taken and used without attribution. However, I like them enough that I'd be happy to see them implemented at all, even if I don't directly benefit from them, and therefore I think I'll share.

For instance, my ideas on the implementation of autonomous vehicle technology. The "cars that drive themselves" idea is clearly neither mine nor very original, but I have some neat ideas about how they might be brought to be. I wrote a "whitepaper" on the subject and did some rudimentary modeling, and I think there's a lot of promise to the approach. Thus:

<<< Transportation Marketplace Whitepaper

This year the DARPA Grand Challenge involves racing a fully autonomous (self-driving) vehicle through a city with traffic, and both GM and Volkswagen are speaking of marketing a fully autonomous vehicle as soon as next year. It’s reasonable to assume that the technology to enable fully-autonomous vehicles will be available to the mass-market within the next five years. The question is: how will it be rolled out?

The obvious answer is to just put the technology on new cars as they’re produced. But this approach could take as much as twenty years to replace the entire vehicle base and there are compelling reasons to encourage the change to happen faster (i.e. 40k vehicle deaths yearly, lost productivity due to traffic jams, the need to spend $25B over 20 years nationwide on highway capacity improvements, environmental damage from inefficient auto use…)
Another approach is to retrofit existing vehicles with the technology. If old vehicles can be converted, as the DARPA contestants have demonstrated they can, there’s no need to sell a totally new platform. What, then, could encourage a large number of vehicle owners to make this expensive conversion? Consider that after a vehicle has taken a commuter to work, it will normally sit idle for nine or more hours until the owner uses it to return home where it will sit for another even longer period of time. If a vehicle is capable of driving itself, it could very well return home to be used by other members of the same family, or it could be hired out for other commuters to use. If the owner is willing to take this latter option and there are a large number of these networked and self-driving vehicles available, an on demand transportation marketplace could emerge.

Assuming a large number of both sellers (the owners of the vehicles who would hire out their cars) and buyers (the people needing transportation, whether to work, household shopping, or school), competitive bidding could emerge. Peak demand rates and odd routes would presumably be more expensive, as determined by market forces, and both buyers and sellers could adjust their transportation patterns to take best advantage of the variable rates.

Why, then, could both buyers and sellers be expected to actually use the service in the first place? Consider the fee for service: a fixed rate will be required to cover fuel and depreciation on the equipment, plus a variable rate that is up for bidding. As long as the rate provides enough income to cover the cost of the seller’s conversion of the vehicle, but is still less than what it would cost to the buyer to acquire a vehicle, both parties can benefit financially. Add this to the prospect of being able to spend commuting time productively (or in relaxation) as well as the long-term potential of improved traffic flow, and we can hope for good market-acceptance.

A key assumption, as noted above, is a marketplace that is large enough to enable competitive bidding and reliable fulfillment on both sides (i.e. sellers can consistently hire their cars out and buyers have transportation readily available; supply and demand are well-balanced). How can this be accomplished? Consider that the initial investment could be borne by a third party, be it a profit-seeking investor or a safety/capacity-concerned government. In such a case, the third party pays to install the technology and earns the commissions on the vehicle’s trips (at least until the cost of the installation is recovered, plus some appropriate profit), the owner of the vehicle obtains an autonomous vehicle, and people who don’t own vehicles obtain flexible and affordable transportation. The combination of private financing, for-profit investors, and (as a last resort) government could enable rapid large-scale deployment on a level that could fulfill this proposal’s need for an active market. The best approach may even be to literally give the technology away for free and earn money from commission on the transaction. If that prospect seems unbusinesslike presently, consider that the price of technology in general deflates at about fifty percent a year; the expensive systems on the DARPA vehicles will soon be available at hobbiest cost.

The really innovative part of this proposal is the marketplace in which the transactions will take place. Imagine a user-interface combining EBay with Google Maps and Paypal, with back-end services that handle the routing, load balancing, and vehicle maintenance scheduling. Add flexible ways to access the marketplace: via smartphone, voice-recognition on a regular phone, SMS, or web browser. Finally, integrate a social reputation rating (like EBay’s) where users can rate how reliable their market counterparts are. These elements together with an installed base of autonomous vehicles create a system that users can take control of and morph to suit their needs.

The success of the marketplace is highly dependent upon network effects; i.e. it’s only useful if there are lots of buyers and sellers. This point has an interesting implication for competitive strategy: openness to market entrants will be more beneficial than building closed systems. If the venture earns money from commission on individual rides, the request for a ride could originate from a competitor, and both could benefit by sharing the commission or charging commission twice.

So what would commission rates be like, and what kind of money could be earned from them? Large sums can be earned from a high volume of small transactions, as Google amply demonstrates. Consider a heavy-traffic market like LA. Assuming only 5% of car drivers could be persuaded to use the service, assuming they make only two trips a day, and assuming they are only charged one cent commission per ride (on top of fuel and maintenance), the venture earns more than a million dollars a year. Each of those factors is likely to be much larger, and there are many other cities whose residents would also welcome a relief from traffic problems. Setting rates around the cost of owning and operating a vehicle and using more ambitious market acceptance rates when considering the entire nation yields a multi-billion dollar business.

Autonomous vehicles are an inevitability. Introducing already-successful information-technology business-models to the concept can yield vast benefits, both to the fast-movers who back the project and society at large. This market-based approach has the potential to reduce emissions, congestion, traffic deaths, and the average cost of travel, while simultaneously allowing vehicle owners and the purveyors of this technology to earn a handsome return on their investment.


My copywriter friend, Rachael Kauffung, even came up with a brilliant and marketable name for the system: the Avee. Get it? A.V. = Autonomous Vehicle, and the name has a charm to it that might help relax some technophobes' fears about letting a robot car drive. It also makes me smile because it reminds me of Snatch (the movie, you cheeky monkey). It Also "borrows" a bit of brand equity from the Evo, though I'm not so sure that's a good thing.

I recognize that there are lots of efforts underway to make the technology feasible, but I havn't seen too many holistic solutions for getting it rolled out. There's benefit-a-plenty to doing it fast (50K+ lives a year! Also money!), and I don't think its out of line to suggest that it could be implemented within the decade, at our lively rate of technological progress and internet-saturation.

So here's to hoping that a person with the resources to make it happen reads this, likes it, and makes it happen. Of course, I'd love to be involved, or to be the conduit for someone's resources to make it happen. I'll do what I can from inside Toyota, but I get the feeling that some entrepreneurial energy is required. I should also point out that all the information used to generate the idea is publicly available and in no way subject to my non-compete agreement (says me).

And for good measure, here's a first-shot at a custom ambigramatical logo:
The V is in the middle, and it reads the same way in a rear view mirror. Now that I'm looking at it, it reminds me a bit of the Mercedes Maybach logo... how's that for being presumptuous! Plus the fact that you can read all the letters of my last name in it (if you really try).

No comments: