Geeks On Ice

Slashdot brings us a story about An Algorithm To End The Lines For Ice At Burning Man:

I skipped burning man this year but went for the first time in 2013. One of the only goods for sale at Burning Man is bags of ice — to keep your own food cool, or simply to refresh yourself, you can line up to buy bags of ice that are sold by Arctica camp out of the back of a refrigerated truck under a tent. Bags cost $3 apiece.

photo: Nellie Bowles

During peak times last year, the lines were up to an hour long. This year, so I heard, the lines on the first day were even worse, because two of the three distribution points were unable to open due to closed roads, so everybody lined up at the only sales tent that was operating.

Regardless of the conditions, the procedure when you get to the front of the line is the same. You specify how many bags of ice you want, and deposit cash in a container on the counter. Then a volunteer walks back to the ice truck to fetch one or more bags from the truck and brings them back to the counter. You collect your bags and continue on your way.

OK, before reading any further — based on what I just wrote, can you think of a way to speed up the line? No cheating — read the preceding paragraph and think of what you might do differently. Spoilers follow!

The thought that occurred to me almost immediately after I got my bag of ice, was: Why not just have the volunteers carry the bags of ice from the truck to the counter, before people place their order? As long as the line is moving, no bag of ice would sit on the counter long enough to melt. And then each transaction at the front of the line would be reduced to: Customer pays for bag(s), customer picks up bag(s) and leaves. By eliminating the time to walk back to the truck and fetch the bag(s), the system would significantly reduce the per-customer transaction time.

I’d asked a handful of Burning Man veterans about this, and they said that Arctica had tried this at one point, but was required to stop by Nevada health code regulations, which treated ice as a “food product” and therefore said that it could not be moved out onto the counter until an order has been placed. This sounded puzzling to me — don’t cafés place other “food products” out on a counter all the time, where they can be bought and picked up by customers? And for the ice bags, why would it matter in practice anyway — even if the state of Nevada is worried about germs starting to multiply as soon as the bag is removed from the refrigerated truck, the time the bag spends sitting on the counter is still negligible compared to the time the customer spends transporting it back to their own camp.

So I emailed the Nevada State Health Division to ask them what the regulations actually said, and if they would allow the ice vendors to load bags of ice onto their sales counter before they had been paid for by a customer. One of their Public Health Engineers replied and said, “I can assure you that we do not require the ice to remain in the truck until it is ordered” (and dryly added, “It is common for vendors to blame the health authority for imagined regulations”). Regarding the resulting long lines, he also advised me, in the spirit of Burning Man radical self-reliance (if not practicality), “You may consider bringing your own ice to the Playa rather than purchasing it from them.”

So that’s it. There’s no regulatory reason why the ice can’t be brought to the sales counter before it’s paid for — where it wouldn’t even have time to start melting, if there are customers eagerly waiting to carry it away — and no reason why the line couldn’t probably move 5 to 10 times faster as a result. (I emailed Arctica to ask if they would start having volunteers bring ice bags up to the counter before customers place their orders, and showed them the email from the Nevada Health Division saying it would be legal. I received a very friendly reply, mostly asking me who I was and why I was concerned about the issue; I said I had no stake in the matter except hoping to reduce the wait times and hence the aggravation and health risks for people waiting in line in the sun. I have not received a reply to any subsequent inquiries after that.)

In a previous article I’d theorized about an algorithm for speeding up the vehicle exodus at Burning Man. (Basically, have a “priority lane” where cars can exit at different times of day, depending on the last character on their license plate. So one hour where the priority lane is set aside for cars whose license plates end in “A”, another hour where the lane is used by cars with plates ending in “B”, and so on. This means that drivers who want to use the priority lane, can just wait for the designated hour, instead of spending five hours queueing up to leave.) That was intended more of an intellectual exercise, as a jumping-off point for a discussion about which algorithms would work best under different theoretical assumptions, and with only the small possibility that it might ever actually be implemented at the real event.

The call to speed up the ice lines is not an intellectual exercise. Unless there’s a non-obvious major problem with making this change, this is something that could be done the very next year, and would save people thousands of person-hours waiting in line in the sun.

My other suggestion would be to have a “turbo” line even faster than the main one, designed for people to complete each sales transaction in seconds. Every customer in the “turbo” line would be required to have exact change (or be willing to overpay and let the vendor keep the change), and every customer would be required to have their cash fanned out in their hand like playing cards when they got to the front of the line. (A volunteer could walk up and down near the front of the line to verify that people already had their cash displayed properly.) A transaction at the front of the line would simply consist of, “Three dollars — bag”, or, “Six dollars — two bags”, where the customer shows their fanned-out money, dumps it into the cash receptacle, and picks up one or more bags from the counter.

With or without the “turbo” line, at first it might seem like it would take extra labor to keep a supply of ice bags moving constantly from the truck to the counter, but that’s not the case. For a given number of bags to be sold, every bag has to be moved from the truck, to the counter, exactly one time. So the total amount of labor is always going to be the same, for a fixed number of ice bags. To have a steady supply of ice moving quickly from the truck to the counter, you might need to have more volunteers working at the same time, but that just means that rather than having 5 volunteers with one-hour shifts spaced throughout the day, you’d have those same volunteers working simultaneously to keep the bags moving.

With the lines moving that much more quickly, what if the ice bags run out halfway through the day? Hopefully the vendor can just send the trucks back out to fetch more bags of ice to be brought back in and sold in the afternoon. But even if they can’t — even if, for some reason, the number of ice bags sold per day has to be fixed at X — you’ve still done an enormous amount of good by reducing the wait time from 30-45 minutes to 5 minutes. Because you still sell the same number of ice bags, but you’ve eliminated the pointless deadweight loss of all the time the customers were previously wasting in line.

And if the vendors can bring in more ice whenever their existing stock sells out much faster, that’s a win too — regardless of whether they’re selling the ice for profit or just for altruistic motives. If they’re selling ice to help people, then selling more ice is better. If they’re selling ice for profit, then selling more ice is better, too.

I’m being fairly pedantic here because I want to make it clear that I think that I think there’s no counterargument to be made to this, under any combination of reasonable assumptions — whether the vendors can bring in more ice or whether they’re stuck selling a fixed number of bags per day; whether the goal of selling the ice is for altruism or to make a profit. Bring the ice out before it’s paid for, shave the transaction time down to the bare minimum of the customer paying money and then grabbing their ice bags, and everyone will be grateful they don’t have to wait an hour in the sun.

And if you’re an adventurer thinking about going to Burning Man, my tips for making it (slightly) easier include bringing your own cooler (separate from any food storage cooler) so that you can buy a bag of ice each day, dump it in the cooler, and have your own supply of ice water. That’s well worth it, whether the wait time in the ice line is five minutes or an hour.

Makes perfect sense to me. So, will it happen?

Like many suggestions for improvements to Burning Man, the first response is “NO”. We’re told “we tried that and it didn’t work” and “the authorities won’t let us do it”. Kudos to Bennett for doing the work to fact check these statements: sure enough, they’re false. There is no regulatory reason for making Burners suffer in queues in the desert.

Having applied his brainpower to the Exodus and Arctica lines, perhaps Bennett can now turn his grey matter to the Will Crawl problems. Hint: mailing tickets to the 20% of Burners coming from outside the US will halve the number of people who have to go to Will Call. Fewer people should mean shorter lines.

Will BMOrg listen to the Burner community, and try something new to make things better for their customers?