The Elevator Dilemma


I received an interesting email from a fellow tinkerer that I met recently. It’s one thing to solve a design problem, it’s another to hypothesize on how to approach it. Here is the email I got:

I recently discovered a programming game called Elevator Saga. It challenges you to write code to move elevators for a building – the goal being to transport passengers as quickly as possible.


I noticed that sometimes it makes sense to reverse the direction of your elevator. For example, say you’re on the 19th floor, going down. Someone on the 20th floor pushes the down button, too. It makes a lot more sense to go up and get that person.


But this creates an interesting dilemma….You’re basically harming individual user experience (pissing off some people by reversing elevator direction) for the sake of making the overall system much faster.


I’m curious if anyone has tried to optimize an elevator in this way. It would be a fascinating design / UX challenge to figure out how to communicate the benefit to the user.

Elevator Saga Screenshot

Before tackling the design issue, there is an assumption here: You are sacrificing the individual user experience for general efficiency.

I would define an individual user experience or riding an elevator as having to conform to what Jon Wiley once described as three parts to a great user experience.

  • Usability – Can I walk into an elevator and communicate where I want to go? Do I feel like it will respond appropriately and predictably? Is it intuitive enough to use having never stepped into this particular elevator?
  • Utility – How useful is the elevator? Will it get me to where I’m going in a reasonable time?
  • Desirability – Am I comfortable while riding in the elevator? Are the benefits over alternatives clear? (I could take the stairs, but will be sweaty? will I be able to keep sipping my coffee without interruption?)

The issue related to the assumption is really about usability and utility.

If we think about the elevator experience as a riding inside experience and a waiting outside experience then assume that the general efficiency benefits the entire experience. Primarily by shortening the waiting time outside. Sometimes the riding experience is about the same as usual, sometimes a little slower, sometimes a little faster. Overall the time from calling the elevator to destination averages out to be shorter than most elevators we have today.

The average elevator rider of today with a western worldview (individual success over the success of the group) selfishly doesn’t care if strangers sharing an elevator get to where they are going in time. Even if it was made abundantly clear that a detour was for general efficiency. The detour would be jarring and hurt the experience as moving in the wrong direction won’t feel like it responded to your request. You could call this a learning curve or growing pains, but in the end you are indeed introducing pain where it didn’t exist previously.

The actual questions is how would you communicate the general efficiency at the cost of individual pain.

Unless this type of elevator design was adopted as a universal standard (with a widespread pain period), it would be very difficult to communicate the benefits. Most elevators get the rider to where they want to go seemingly faster and behave in a similar way. I would speculate that this cannot be solved by a elevator design change alone, but rather by a strategy to increase adoption.

You would have to identify and communicate some piece of data that articulates the benefit at an individual level. You could use energy savings in terms of some emotionally charged resource: “your ride saved N amount of oil”. (No-water urinals do something similar,but I’m not convinced the two could be compared since the cost to the user of not flushing is actually a convenience.)

Today the issue of elevators stopping often (and lack of space) is solved by having a special-case freight elevator in addition to general purpose elevators. There is a choice between riding the economy elevator and an express elevator. There could be some sort of reward system for individual contribution to energy savings. Over time turning a reward system for economy into a cost system for an express service.

This year we will give discounts to shoppers who bring their own bag to the grocery store. Starting next year we no longer offer a discount and will be charging customers who require a supplied bag.

Another potential solution would be to allow elevator users identify their destination from outside of the elevator as they call it. Rather than just saying “up” or “down” they select the floor they want to reach. The rider is then told which elevator to enter and an estimated overall time to destination (+/- a minute or two). This would eliminate the unpredictability of the whole thing and (in the best case) highlight the overall efficiency. It could also provide the rider with another data point to consider. Should I just take the stairs based on this time?

I’ve clearly not had much time to write in some time since moving to Seattle. So thanks to Cody Romano for inspiring me.