How to tame those long lines

nce again, it’s election season, and once again, the New York City Board of Elections is under fire. This disaster is as predictable as the phases of the moon.

To be fair, the long lines for early voting at sites across the city, which stretch down and around blocks and in some cases take hours to get through, are the result of many factors, some of which can’t be predicted precisely. The biggest variable is the number of people who will show up to vote. Small changes in demand can result in large changes in waiting time.

But time-consuming slogs aren’t inevitable if we make some smart changes.

I teach operations analysis and queuing theory at Columbia University. Those two esoteric-sounding topics mean I teach how to make things work. In particular, operations analysis looks to uncover bottlenecks in a production or service process, and queuing theory helps figure out how to reduce and improve the lines on which we spend lots of our lives waiting: at supermarkets, on tech support lines and so on.

To speed a slow line or queue, we first need to understand what people are actually waiting to do. Looking at voting, people arrive to vote, they wait in line, check in, get and fill out their ballot, then insert the ballot in the voting machine.

In this process, the bottleneck depends on how many check-in people are assigned to a voting location, how many privacy booths are there, and how many voting machines are in service. From my observations, the check-in takes usually a minute or so, filling out the ballot takes about five or six minutes on average, and the actual voting takes no more than a minute.

Lines are primarily the result of more people arriving than the chokepoint in the processing facility can handle. Understanding the processing bottlenecks is the key to solving the problem.

Case in point: On Monday, I went to vote at the Wagner Middle School on the East Side of Manhattan. The waiting lines went around the building twice. People started arriving at 6 a.m. At 3 p.m., there were so many people in line, officials stopped allowing new arrivals. (Alas, I didn’t get to vote.) Clearly demand was much greater than processing capability. Why?

The voting process inside started at 7 a.m. and continued until 7 p.m. The facility processed 2,935 voters, or about 250 per hour.

The facility had 23 privacy booths. Each could handle, at an estimated five or six minutes per voter, 10 to 12 per hour. Therefore, altogether, the privacy booths could, in total, process about 250 to 275 per hour, closely matching the overall calculation above.

But people arrived at the rate of about 325 per hour, substantially more than could be processed each hour, hence the lines.

The Board of Elections is, of course, in a bind. Clearly, they had too little capacity to handle the onslaught. If they had had too much processing capability (as they often have in off-year elections), critics would claim they wasted taxpayers’ money.

So how could they have handled this situation, and how should they handle similar situations in the future?

Short of shifting to all mailin voting, the first imperative is to have a better forecast of the likely volume of people showing up early. Surveys taken a month before early voting starts could’ve helped them better calibrate the resources needed. Although even if they knew, they may not have had the budget to provide what would be needed.

Second, scheduling — giving people the chance to reserve a time to vote — is a way to match the demand to the capacity. We do it at the golf courses, at Apple Genius Bars and in countless other contexts. Why not for voting? You could ask for and get a time, say a half-hour interval, and when you arrive, you will have very little wait, like FastPass at Disney World.

These and other innovations should be implemented not just in NYC but nationally. The right to vote is the bedrock of a democracy. We need a national voting-line analysis project to find what it will take to make it equal and fair for all Americans to vote.

My friend Dr. Richard Larson of MIT — nicknamed Dr. Queue — has written that America needs waiting standards for voting the way good businesses understand what their customers expect in order to remain loyal customers.

Make the voting experience easy — maybe even pleasurable — and many more people may line up to do it.