The Zipcar Prize proposal: enabling one-way car sharing trips | Hacker News Reader

The Zipcar Prize proposal: enabling one-way car sharing trips

troy.yort.com

53 pointstroydavis20 comments

Rather misleading title, since it's not actually about a prize, it's about a prize that some guy thinks should exist.

The problem of allowing one-way rentals is not one that I think Zipcar is particularly interested in solving. No matter how you slice it, it winds up leading to a less efficient allocation of parking spots than the current one spot per car model, it increases the probability of a pissed-off customer who finds that the car from his local spot has semi-permanently vanished to four hundred miles away and nobody is currently planning on one-waying another one in, and it wouldn't necessarily lead to enough increased revenue to make these two downsides worthwhile. In fact it might well lead to less revenue, since someone making (say) an overnight round trip could just book two one-ways instead of a return and wind up saving money.

I don't think the title is misleading, and it definitely isn't intentionally. Half of the post is an email to Zipcar proposing it, and the title needs to be short.

I think the other issues you identified are real, are solvable, and are addressed in the post.

The post explicitly says not to allow a one-way trip to empty out a lot/neighborhood (permanently vanished car). An "I miss this car" button on the Web site would make it easier for Zipcar to know when to reward someone for parking a specific car type in a given lot.

In many cities, Zipcar has dozens of cars in the urban core. The use here is going from Seattle to Bellevue or Mountain View to San Jose, not BFE 1 to BFE 2.

I agree that Zipcar hasn't been not particularly interested in solving this problem. I think that may change now that they have competition in most markets.

As far as less revenue, most of the trips this enables aren't happening today because so few people spend $60-$80 (day rate) to drive 30 minutes then park for 8 hours.

If someone could book two one-ways cheaper than a round-trip, it would be because it's helping Zipcar. If my driving balances out two one-way trips booked earlier, I'm helping reduce Zipcar's costs by slightly altering my behavior (driving different cars).

No, I think it's pretty definitively misleading without context. And there's a simple, short solution: "The Zipcar Prize proposal: ..."

Thanks. Title changed.

What if such a system would only allow one-ways if A) nobody had reserved for any extra gap created (or the gap-cost were charged and split), and B) another person was intending to do a one-way in the other direction, with a similar-grade vehicle? ie, only fairly-precise trades / N-way rotations. This, of course, makes one-ways to/from small cities very hard to achieve, but it gives you one-ways in cases where it does work for almost free.

I think the bigger problem with zipcar is having a car sit around for hours while you're visiting a friend or sight-seeing.

As a zipcar user, I often want to drive somewhere that's 30 minutes away and stay there for 5-8 hours. That means I'd have to pay for the 5-8 hours that the car just sits around. (Which means I just skip it and hunt for a cab.)

On the reverse side, on sunny weekends, I often can't find a car to rent. Since I live near a part of the city that's a big draw on the weekends, I bet there are Zipcar customers who visit my neighborhood and leave their Zipcars in parking lots all day. I would pay extra to use their cars for a few hours.

Good point - my first reaction was just do two one-way trips, but I imagine a lot of the sight-seeing that people want to do isn't within easy walking distance of a zip-car lot.

Zipcar's hourly pricing is so low because the rental for these extra 'waiting' hours is built in. If you are driving at 60 miles per hour and the car gives 20 miles per gallon, the cost of driving for one hour is over $10 at current gas prices. Zipcar's hourly rentals don't cost much more than $10.

Your idea of using cars others have parked locally is a good one, although I wonder how many people rent zipcars for 7-10 hours and park it for 5-8. It would be almost certainly cheaper to rent by the day from Enterprise et al if you did that.

It often is cheaper to rent by the day from companies like Budget, but zipcar is incredibly convenient. In DC, I have 5 locations within 4 blocks. The major car rental companies are all 1 cab ride away.

Won't enterprise come pick you up?

I really like your idea but if Zipcar let you rent someone else's parked car, and you were late, they would almost certainly lose the stranded customer.

I know how much people hate waiting for a ride because my startup RideCell worked on real-time carpooling for some time before we pivoted to fleet automation. Even now, one of the main problems our product solves is reducing wait time for transport/service vehicles by auto-dispatching tasks to the nearest one, routing them intelligently, and making the wait more bearable by letting people track their transport/service vehicles (Think Uber for everything).

Of course, the fact that this is a hard problem makes it even more interesting :)

The point of the Netflix prize was to improve something Netflix already understood well and was getting diminishing returns for its research dollar. The Netflix Prize was able to attract serious interest from legitimate academics and engineers because the problem was refined to the point where it was legitimately very hard.

ZipCar doesn't even offer one-way car trips. I'm sure it could do them with adjustments to inventory, management, and pricing. Traditional car rentals are evidence there is a model that works. But ZipCar needs start one-way car trips, much less get to the point where it has reached diminishing returns on its research investment, before it can expect worthwhile contributions from crowd sourcing.

Isn't this already done by (for example) U-Haul? They would quickly go out of business if you couldn't rent a truck one-way, or if they couldn't manage their inventory.

U-Haul's business model is not precisely the same as Zipcar's, but it's pretty close.

Car2Go (a Daimler company) is tackling this problem with the concept of an operating area. You can pick up and drop of your car from a car2go designated spot or any legal parking spot within this area. Here is an example of their operating area in Austin

http://www.car2go.com/portal/austin/page/mybookings/mapEnlar...

They are also innovative on other fronts such as charging by the minute and a feature-rich touchscreen in each car.

The ultimate car service would allow me to find a car a short walk from where I am and then do a one way trip, without a booking. I've always felt that the only barrier to providing this service was a critical mass of drivers and zipcars. Most cars are currently idle for the majority of their lifetime, taking up a lot of parking space. Imagine if the streets were lined with shared cars instead of privately owned cars. Once you get to a certain density you should be able guarantee availability in all but the most remote locations. There are large scale migrations like commuting that take place over the day/week but the availability of cars should match this. Given a rising population density it seems inevitable that such a service will have to exist as we will run out of storage space for idle cars.

Do any zipcar-like rental companies allow you to return the car to somewhere other than where you picked it up? Doing so obviously dramatically changes the complexity of reservations. It is a feature that more traditional rental operations can offer only because they have very large centralized inventories. Even so, they still have to orchestrate moving cars from popular destinations back to popular origins.

The proposed solution is definitely interesting, and I think it might be possible to make it workable, but I think it is a different order of complexity from the Netflix prize.

Just signed up for Car2Go.com in Vancouver, which launched here this month and offers one-way rentals.

Their model is per-minute pricing with an hourly cap.

You "return" it by parking it anywhere within a huge area (several square miles) that effectively makes up anywhere you'd be car-sharing anyways (living outside this area means you'd be pretty likely to own a car).

You find a car to rent with their android or iphone app, website, or by calling their 24/7 phone number. All cars are tracked using GPS.

Pretty sweet model that is a good substitute for taxis, assuming the car supply keeps up with consumer demand.

Is it just me or does this smell like a great social media job by Car2Go.com?

I think the solution would be variable pricing, much like buying or selling a financial asset. If the rates for things could be adjusted to encourage activity around some number (say, to keep the number of available cars in a region at 1 for every 10 zipcar members), using some real data, they could just let people drive cars one-way. If it turns out that everyone wants to borrow a car from A and return to B, that just means it's too cheap to do so, or it's too expensive to borrow from B and return to A.

It would certainly be an interesting model to try.

I worked on developing a model like this while at the Media Lab. We were trying to model dynamic pricing for one way rentals (with and without restrictions as described on this thread) but ultimately found it very difficult to simulate due to the lack of data.

We tried using proxy data by scraping popular bike sharing websites but it didn't seem analogous enough. Any ideas where to get zipcars reservation data with zipcar offering such a prize?

Reply on news.ycombinator.com