When I was a Masters student at MIT, I studied Operations Research...you know, queueing theory and the shortest paths problem and such. Around this time, I started to get annoyed (mainly on principle) at the people running the checkouts at the CVS in Porter Square. (Perhaps this indicates a pathological condition on my part, but hear me out.) One day I went into this CVS and saw a sign above the checkouts asking people to form separate lines in front of each register. In addition, this sign specifically pointed out that the goal of the proposed system was to provide faster service. As one who has though a lot about these types of problems, I knew that the stated goal couldn't possibly be achieved. Apparently Whole Foods got my telepathic memo.
Whole Foods has decided to implement a single-line policy in its Manhattan stores, which is in contrast to the typical one line per register setup. This has, not surprisingly to me, resulted in shorter wait times for customers and greater equity. (Wow, a time where efficiency and equity are not at odds, at least not for customers.) Think about it- with one line, you never have to worry about whether you picked a "good" line, you wouldn't be held up by the guy in front of you with 100 coupons, and you would never get upset that the guy that entered the line next to yours 2 minutes after you got in line is now happily walking away with his groceries while you are still standing in line.
Theoretically, this is great- why didn't companies think to do this sooner? My guess is that they probably did but were afraid that they would upset customers. Why? There is a tendency for people to believe that a wait is longer than it actually is when they see a very long line. This psychological phenomenon would lead people in the single long line to either perceive the store experience to be more unpleasant or to balk altogether and leave without a purchase. This congitive bias leads to the persistence of an inefficient setup, since consumers are largely incorrect in this case about what will make them happiest...in other words, a Pareto improvement could be achieved if the consumers' biases could be reduced or eliminated.
Update: Apprently I'm not the only one that finds this interesting...Tyler Cowan has a post on his web site about the economic relevance of supermarket lines as well (on which I made an ass of myself trying to post a comment, but that is another story). Economists are already looking at supermarkets- Alex Mas and Enrico Moretti, for example, have a paper that gives evidence in the supermarket realm that workers are positively impacted in terms of speed and productivity by having more productive co-workers around. I am curious as to whether there is also a productivity impact of seeing more people in line. That said, I am not going to try to guess whether a cashier would work faster if she saw a long line or just get frustrated and register a productivity decrease.