How to avoid the "local elevator" problem?

I live in a high-rise apartment building in New York. Our building has four resident elevators which each serve every residential floor. (An additional fifth elevator (physically separated from the four described herein) serves as a freight elevator for movers, contractors, and building staff.)

However, "elevator #1" is special - it is the only one of the four that can reach the rooftop, as well as the only one of the four that can reach the parking garage in the basement. It is also slightly larger in size than the other 3 elevators, to accommodate wheelchair-using residents and strollers.

As such, the elevator banks on each floor have been outfitted with two elevator call panels - the one in the center of the bank (panel "B") can call any of the four elevators, while the buttons on the panel next to elevator 1 (panel "A") specifically call only elevator 1:

1               2                 3               4
┌─────┬─────┐   ┌─────┬─────┐     ┌─────┬─────┐   ┌─────┬─────┐
A  │     │     │   │     │     │  B  │     │     │   │     │     │
┌─┐ │     │     │   │     │     │ ┌─┐ │     │     │   │     │     │
│▲│ │     │     │   │     │     │ │▲│ │     │     │   │     │     │
│▼│ │     │     │   │     │     │ │▼│ │     │     │   │     │     │
└─┘ │     │     │   │     │     │ └─┘ │     │     │   │     │     │
│     │     │   │     │     │     │     │     │   │     │     │
│     │     │   │     │     │     │     │     │   │     │     │
└─────┴─────┘   └─────┴─────┘     └─────┴─────┘   └─────┴─────┘

A problem commonly arises (especially during peak times) that residents will walk by both panels and select their desired direction of travel on both, even if they are intending to go to a floor which is serviced by all 4 elevators. This becomes a challenge when those riding in elevator #1 are now stuck stopping at every floor in which another resident has pressed a button on the leftmost panel, whether or not the resident has already boarded one of the other three elevators headed in their direction. In these peak periods, this can easily add 2-3 additional minutes to a resident's travel time, effectively becoming the "local" elevator (as opposed to the "express" - for anyone who's ridden the subway, you'll certainly know the difference).

How can residents be guided to use the special panel ("A") as intended (only when intending to ride elevator #1)?

• Is this a problem you are tasked to solve, or is this an interview question? Jul 20, 2021 at 17:16
• Neither, actually; I literally live in a building where this is a consistent issue, and wanted to understand if there were any UX approaches which could assist in solving it. Jul 20, 2021 at 17:18
• A ux vigilante! Sounds great. Jul 20, 2021 at 17:54
• @pajonk I currently don't have very much in the way of material limitations - I'm personally interested in understanding the "art of the possible" in this scenario. It'll go a long way if/when I and other residents propose some alternatives to our building's management. Jul 21, 2021 at 13:45
• Are users pushing both buttons because they're confused or because calling two elevators reduces their expected wait time? Jul 21, 2021 at 16:34

Replace the panel A with buttons “to rooftop” and “to garage” instead of “up” and “down”. Now elevator 1 is reserved for special floors and most people use elevators 2-4.

• I've been to a building that does this. You don't just press an elevator "call" button but you press your floor range instead (though, there's only one panel instead of multiple - but with multiple call buttons). I don't remember how many call buttons there were on each floor but they were grouped like "1-5", "6-10" etc. Jul 21, 2021 at 4:38
• If there are regular users they are likely to figure out that 'to rooftop' doesn't preclude using the elevator to get to the other floors though. People will experiment. Jul 21, 2021 at 8:35
• But isn't the first elevator also used for people using wheelchairs going to non-roof floors? "To rooftop" and "To garage" don't capture this usage. Jul 21, 2021 at 8:52
• One issue with completely discouraging elevator 1 is that the goal shouldn't necessarily be to direct almost everybody to elevators 2-4, because that means even if nobody needs rooftop/garage/special access, one elevator is effectively out-of-order, and the elevator bank wouldn't be running at full capacity during peak times. Jul 21, 2021 at 9:39
• @jamesdlin the panel for elevators 2-4 also calls elevator 1, as per the question's third paragraph Jul 21, 2021 at 9:46

On further thought, another solution could be to implement a singular, digital wall panel where users can indicate to which specific floor they are intending to go, much like the ones found in many newer high-rise office buildings and hotels:

Original image by Access Matters on Flickr, used under CC BY-NC-SA 2.0.

Once the floor is selected, the main computer decides which elevator is most appropriate to handle the request based on capacity, proximity, and ultimately whether a specific elevator can service the requested floor or not.

While I don't have access to any specific figures, I would imagine this is more cost-prohibitive than other solutions, and wouldn't make sense in many scenarios similar to the one posed above.

• What about wheel-chair users or people with strollers? Jul 21, 2021 at 7:56
• @JannPoppinga The panel just would need an extra button for special accommodations. Jul 21, 2021 at 9:44
• @CaiusJard sure but you need one of these per floor, and it's got 4 elevators so it must have a lot of floors! Jul 21, 2021 at 15:49
• @JannPoppinga, jamesdlin: The panel in the photo does have a wheelchair button. Jul 21, 2021 at 16:01
• Valid answer but this is rather \$\$\$\$ to install these and their current elevators circuits may not even support this. Its like saying drop this new engine in your car to make it run faster and the engine wouldn't even fit. Jul 21, 2021 at 19:58

In an office building where I worked, they solved a similar problem by making panel A a bit less obvious than panel B. Panel B was right out in the open, on the wall between two of the elevators, facing the lobby.

Panel A was located in the kind of door jamb space of elevator 1, perpendicular and right next to the elevator doors.

• This was my office, too. Also, pushing the special elevator button lit up the main elevator panel, so that you couldn't call both. (It's possible that you could get on another one if it happened to stop there, but "get the special elevator guaranteed or another one if another one happens to stop first, but potentially have several normal ones go by while waiting for the special one" is much less tempting than "get whichever elevator is first available".) Jul 21, 2021 at 17:33
• Although this is still abusable, slightly, just less so--see comment on Martin Kealey's answer. Jul 21, 2021 at 17:40

My thoughts would be to consolidate both sets of buttons into a single panel and draw focus to the primary action buttons.

• DOWN

• Car Park

• ... but also "up in a wheelchair"? Jul 21, 2021 at 9:39
• @teedyay: Why would you have a button specifically for "up in a wheelchair"? Jul 21, 2021 at 10:10
• @musefan because elevator 1 serves all floors and is the only elevator wide enough for a wheelchair (and strollers and whatever else) that does so. Jul 21, 2021 at 10:27
• This answer would be substantially improved by an explanation of how this functions to reduce/remove the problem. Jul 21, 2021 at 10:29
• If the system still functions the same, bringing elevator 1 only if you press the extra buttons, why would people not press both for faster service? If not, explain how. Jul 22, 2021 at 7:17

This is a people problem.

This can be fixed by using some scotch tape and a piece of paper which reads:

Please use this panel only if you need roof or basement access, or need extra space for a wheelchair or other large belongings.

• I do not believe in this. People just won’t realise „Oh, there is information I should process“. They tend to not see such things. Jul 21, 2021 at 15:03
• @peter_the_oak Then dispense spiders from panel A whenever a transgressor uses the panel but candy from panel B. Jul 21, 2021 at 15:21
• People tend not to read signs like this. They just add clutter and further annoyance because the delayed elevator users are now annoyed both that people have unnecessarily pressed the button and annoyed that they have ignored the sign. Jul 22, 2021 at 16:00
• Unfortunately, having elevator 1 only summoned by its own buttons will increase the incentive to use both call buttons. Jul 22, 2021 at 19:45
• @Falco I agree, thanks! Jul 23, 2021 at 13:28

The variant:

If this is not clear - this is bad solution.

1. Of course this is just a mockup.
2. Add accessibility icon at Panel A as most people respect this mark and will not use marked buttons.
3. Add notice to Panel B for rooftop and garage to make clear what to do - explicit allowance to use Panel A appropriate buttons.
4. Panel A button pairs Up and Rooftop as well as Down and Garage are electronically the same, so only panel should require upgrade but not whole elevator schematics.
• The pictures you have made my head spin. Why two wheelchairs? Also the wheelchair icon is OK def not great. The garage and rooftop icon... really bad. No way in the world I would think those meant garage or rooftop. Maybe if you gave me 50 guesses I might say one of the words eventually. Why wouldn't the button just say Rooftop? What happens when you hit that button? I would expect that it would already be pressed for me in the elevator. Also I don't see how this solves the existing tenant issue... they will just hit the up and down with the weird icon next to it. Jul 21, 2021 at 19:36
• @blankip Sorry about your head ) Try to answer in order of your questions. Once more - this is mockup just to illustrate an idea. Real designer possibly will depict one wheelchair and group in some way two buttons. Icons - available icons just for illustration. Buttons with text labels - the best solution I think too. You are right about expectation - this is possible an issue. But this could be solved by updating elevators schematic and firmware if it is possible.
– Serg
Jul 21, 2021 at 19:52
• @blankip Addition to previous comment: Another solution is to remove extra buttons and label up button as 'To rooftop' and bottom button as 'To garage'.
– Serg
Jul 21, 2021 at 19:59
• Why not use a single panel with 6 buttons? E.g. buttons arranged similarly to the arrangement of the icons on your current Panel A, with an icon on each button (just add the arrows to the wheelchair icons). The choice of the elevator would be left to the controlling circuit, which would decide the best option depending on user's needs. Jul 22, 2021 at 9:31
• @blankip Of course this isn't a final solution. Don't be so critical. Jul 22, 2021 at 14:13

The heart of the problem is that people going to every floor except the roof & basement will on average get better service by pressing both buttons. Any modification needs to remove that incentive, or remove the ability of inappropriate users to "double call".

For the floors other than lobby, I would make the selectors mutually exclusive: pressing a call button for elevator 1 cancels the other call (if any) and disables the other call button until elevator 1 arrives. You want to make your trip slower by calling E1, go right ahead...

If there are busy times when lots of people leave a floor at the same time, there might be a need to relax the mutual exclusion rule. One way would be to cancel both calls when either elevator leaves a floor, unless the other call button has been pressed since it arrived. I would have the other button start blinking to warn people that it's about to cancel.

For the lobby there are several approaches.

One possible approach would be for elevator 1 to always go directly to the rooftop or basement directly, and only stop on other floors on the return. Other users would quickly learn that it's not in their interest to use that elevator.

Another would be installing a destination floor selector in the lobby, plus buttons for "wheelchair" and "freight". Elevator companies are notorious for charging excessively for trivial things, but the software logic is simple enough, and could be implemented as an after-market kit that simulates button presses within the elevator. (I suggest leaving the buttons inside the elevator in case the software screws up, but put them inside an alarmed cupboard to discourage people from using them.)

If you are in a jurisdiction that makes it difficult for "amateurs" to modify an elevator system, but where you have a centralized locking system then that could be used to enforce the elevator allocation without modifying the elevator system itself: instead of call buttons, the person taps their card in the lobby, and the system tells them which elevator will next take them to their home floor. If they get into the wrong elevator, their card simply won't work. There's a small chance of a sub-optimal outcome where someone doesn't want to go to their home floor. And in the short term, people may try to game the system by pressing "freight" until they realize their waiting time for an elevator is worse on average. It might be helpful to have separate card readers for calling "up" and "down" elevators.

(Other solutions that partially alter the UI have merit, but neglect wheelchair users' need for the extra space in elevator 1.)

Directing people to specific elevators also makes it easier to enforce elevator passenger limits, particularly when a town is in a Covid-19 lock-down. Tapping a card rather than pushing a button is also a hygiene measure.

• How does a "card operated" system deal with lost or forgotten cards, visitors without cards, etc? (For example a group of residents all leave together using only one card to call the elevator, but when returning separately some of them discover they forgot to bring their own card) Jul 21, 2021 at 12:07
• This can still be abused slightly (source: office I worked in). Since the special elevator will prioritize calls that can only be serviced by the special one, if most people are only using the regular call panel, then using the special one will still get you an elevator more quickly. In a perfect knowledge situation where people always chose which one was fastest, this would average out anyway to make the special one the same as the others. But without perfect knowledge, people do weird things--probably some will always use the regular panel, and some always the special one, regardless. Jul 21, 2021 at 17:40
• @alephzero I'm not suggesting that such a system be installed where it doesn't already exist; when I said "where you have a centralized locking system", I mean you already have such a system. Jul 22, 2021 at 19:57
• @user3067860 If people can game the system by calling the "dedicated" elevator, then the prioritisation is too strong. It should be tuned so that it gives slightly worse than average times for everyone using it, roughly in proportion to the number of extra floors it needs to traverse. Remember that the point is to improve trip times by removing vacant stops. Jul 24, 2021 at 19:23
• @MartinKealey Machine learning elevator, perhaps? I don't think elevators really support that much logic--at least not the garden variety elevator for mid-rise buildings. Who is even going to gather statistics to decide what the average wait time is? Jul 26, 2021 at 13:41

I think this is especially tough because the combination of these issues:

1. E1 is the only way to access the roof and garage
2. E1 is the main intended elevator for for wheelchairs/strollers/accessibility
3. E1 is associated with panel A AND panel B

I'm going to assume that most users already have some way of knowing (existing labeling?) that E1 is the appropriate elevator for roof access or accessibility.

So looking at the main question...

How can residents be guided to use the special panel ("A") as intended (only when intending to ride elevator #1)?

A couple options with the goal of discouraging users from pressing panel A unless they need elevator #1, while not causing confusion if elevator #1 opens when panel B is pressed:

The nudge

Make Panel "A" more clearly associated with only elevator #1:

1               2                 3               4
+-----+-----+   +-----+-----+     +-----+-----+   +-----+-----+
A|     |     |   |     |     |  B  |     |     |   |     |     |
+-|     |     |   |     |     | +-+ |     |     |   |     |     |
| |     |     |   |     |     | | | |     |     |   |     |     |
| |     |     |   |     |     | | | |     |     |   |     |     |
+-|     |     |   |     |     | +-+ |     |     |   |     |     |
|     |     |   |     |     |     |     |     |   |     |     |
|     |     |   |     |     |     |     |     |   |     |     |
+-----+-----+   +-----+-----+     +-----+-----+   +-----+-----+

The colors

(I'd have made diagonal stripes but who has the time)

• I like your answer, it's good UX and design. The OP's users already know the left panel is elevator 1 only though, so how do you expect your suggestion to change their behaviour? Jul 22, 2021 at 11:08
• My thought is that making panel A "special" would add a block to the mental flow of using panel A, resulting in (1) reduced unconscious use of both panels just because a user is in a hurry, and (2) reduced "selfish"/conscious use of panel A due to having that slight extra block of "this looks different from the rest" Jul 22, 2021 at 13:43
• Very nice. Perhaps especially so if it had a disabled icon on it somewhere too. Jul 22, 2021 at 15:48

Solution1

No hardware modification required. Just add two software rules:

• full priority to the destination floor selected inside the elevator 1 when at the garage level or at the roof level
• full priority to the elevator call buttons at the garage level and the roof level

This way the elevator 1 will not be affected by people pressing both buttons on all the other floors

Solution2

Have this:

• I like your first suggestion very much. Also like the panel design but what I am missing there is something that discourages people from pressing Panel A if they do not need it. (however your solution 1 may be enough to discourage the frequent users, but not the new users). Aug 7, 2021 at 12:46
• I really love Solution 1. It's a very elegant solution. However, it does become a problem for wheelchair users on other floors: they'd end up waiting forever to get anywhere. Aug 8, 2021 at 21:10

Many of the proposed solutions involve combining the panels or relabeling Elevator 1's controls. The problem with this is accessibility: anyone who needs the larger car (especially if they are not tenants -- visitors or delivery people mostly) may not understand they have to use that panel to get Elevator 1 if their destination isn't one of the specially-marked destinations.

Instead, I propose putting a giant blue-and-white wheelchair sign above the left panel (ideally with braille at the bottom, or below the panel):

Possibly include signage stating "ELEVATOR 1 ONLY", as well as signage above the center panel stating "ELEVATORS 1-4", so more astute users will realize the center panel can summon the first elevator as well.

This should be a clear indicator to disabled users (and stroller users and delivery people) that this car is a larger-capacity car better suited to their needs, while (at least hopefully) discouraging casual usage of the panel by people who are going to use the main panel anyways.

• Hmm... Although a parking area specifically designed for disabled people (sometimes) discourage casual usage, I think a totally normal but roomy elevator with that sign might not work so well, though it would definitely allow concerned people to quickly understand that this is the way. Mostly speculation though. Jul 23, 2021 at 16:59
• Ultimately, user behavior modification is what we have to rely on. Short of having the elevator panel staffed by someone who can say "oh this person has a walker, they need panel A; this person wants to go to the garage, they need panel A; this person doesn't need any accommodations, we can use panel B"... you're not going to be able to make people stop using panel A when panel B would serve them just fine; the best you can do is discourage it. Jul 23, 2021 at 17:45
• I think you'd also need a small sign by panel B saying "For roof and garage, use elevator 1" Jul 28, 2021 at 17:48

I think the functionality of the elevators and two panels are fine. The issue is with user behavior.

You have two sets of users:

1. Users who are in a wheelchair or want to go to roof or basement.
2. Users not in set 1. ie. non-wheelchair users who want to go to general floors.

For each set of users, we want to encourage and discourage a certain behavior.

1. For these users, we want to encourage them to use (special) panel A and discourage them from using panel B.
2. For these users, we want to encourage them to use (general) panel B and discourage them from using panel A.

Note that we will allow both users 1. and 2. to use the elevators simultaneously. Both panels will be used, but at the same time, both special and general elevators that get called down will be used by users 1. and users 2., respectively. We want to avoid the case where a user calls down an elevator and abandons it.

Design:

First, we can put a simple sign on top of each panel. panel B will have a sign "TO 1-N", where N is the topmost non-roof floor. panel A will have "TO B-ROOF <wheelchair symbol>". These signs are just to give a general idea and may be vague to the user. But we will try to reinforce/punish to get a certain behavior out of them.

To encourage users 2. to use panel B, we will put it closer to the entrance. There will also be a chance that panel B calls down the special elevator and that would also encourage users 2. to only use panel B as they would see that there is no need to use both panels.

To discourage users 2. from using panel A, we will put it a little farther off from the entrance. Not too far so that it would inconvenience users 1., but far enough that users 2. will be too lazy to walk the few extra steps.

The idea is that a user sees panel B before panel A. They have to actively decide to press panel A in addition to panel B by walking extra steps. Having panel A come up first is bad because it leads to users accidentally using panel A. They would see panel A first and use it. Even if they go and correct their action by using panel B, they can't cancel their mistaken use of panel A.

We already encourage users 1. to use panel A since it is the only elevator that can go to the special floors and accommodate wheelchairs.

We discourage users 1. from using panel B because they might get 3/4 wrong elevators. If they are on a wheelchair, they will find out that there is not enough space. If they want to go to the roof or basement, they will find out the buttons for those floors are missing.

In addition to all of this, we can put up a sign with a few sentences on elevator etiquette for those that are not too lazy to read. They can help spread the word. For example, if a person sees a wheelchair user waiting for a panel B elevator, he can kindly tell the wheelchair user to use panel A.

• This is one of the few answers I've seen here that actually considers users who need the first elevator for something other than roof/garage access. I think perhaps combined with my answer (put a large "disabled" placard above panel A) this might be the best solution. Jul 21, 2021 at 16:49
• Thanks for the recognition. Yes, I was specifically trying to utilize the first elevator to improve efficiency. We don't want an elevator sitting around when there are no group 1. users. I like your idea. I actually typed <wheelchair symbol> in my original answer, but it didn't show up. I guess it was an escape sequence. I edited my answer now. Jul 21, 2021 at 17:30
• Maybe have panel A require a key or card to use. Then no one wants to dig out a card instead of press the panel B button. Jul 22, 2021 at 15:40
• That's a good idea. They can just reuse their apartment key/card, so we don't have to pass out any special key/card. As long as we don't mind preventing people without cards from using panel A. Jul 23, 2021 at 5:42
• Hello! Can you clarify what you mean by "There will also be a chance that panel B calls down the special elevator"? Thank you! Jul 24, 2021 at 1:47

Current system:

• Panel "A" - Calls elevator 1
• Panel "B" - Calls elevators 1-4
• Panel "A+B" - May end up calling two elevators

Below are 2 proposed fixes, intended to be used at the same time.

Fix 1:

1. If a user pushes any button on any panel, the corresponding light will turn on.
2. If a user pushes a direction on panel "A", it will silently cancel that direction on panel "B" (the light on panel "B" will remain on).
3. If a user pushes a direction on panel "B" while panel "A" is active, it will silently be ignored (the light on panel "B" will still turn on).

This has a few effects:

1. If multiple people want to use an elevator, only one elevator will be sent.
2. Users will learn (or be told) using both panels is slower (i.e., because if elevator 1 is in use, elevators 2-4 won't arrive). If they do so anyways, they're only hurting themselves.
3. If a user is fine with any elevator but someone else wants elevator 1, they'll be penalized. This is mitigated by fix 2, since having multiple people request an elevator at once is more common in the lobby.

Users can still game the system by pushing the wrong direction and hoping an empty elevator shows up.

Fix 2:

Set elevator 1 to rest in the lobby when not in use.

This has a few effects:

• Optimizes delivery of freight/packages (freight typically only goes in one direction; the delivery person can go down on any elevator).
• The lobby will tend to call an elevator faster.

Notes:

The people who know/care/understand the current system should already realize that pushing both panels is not necessarily faster, since panel B already calls elevators 1-4. The goal of this solution is not to penalize people who abuse the system. Rather, the intent here is to ensure that people who use both panels don't slow down the elevators for everyone else (i.e., by using 2 elevators when they only need 1).

• I like fix #1. Another addition may be to add presence detectors to the elevator lobby so that calls will get canceled if everyone who was there, leaves. Jul 22, 2021 at 21:15
Elevator1   Elevator2   Elevator3   Elevator4
~4 min      ~1 min      ~3 min      ~3 min
>CALLED<

The problem seems to be people believing they get a benefit out of pressing both buttons, when in fact they will not. I think this can be addressed by communicating the state of things more clearly. By displaying an estimated waiting time for each elevator (calculated by distance and requested stops on the way) the users will clearly see the fastest option is reserved for them after Pressing Panel B.

The estimated times are always displayed - so this will encourage people to use Panel B if one of the three Elevators displays the smallest estimated waiting time. And if Elevator 1 has the smallest waiting time it does not matter if they press B or A.

This solution could be quite costly to retrofit on your building, but I think it could solve the problem quite nicely by informing the users.

• Make panel B light up the buttons on both panels, summon any elevator, and have both go out when any elevator answers.

• Make panel A light up only panel A, and summon only elevator 1.

• Make elevator 1 responding to a panel A summons or use NOT respond to panel B summons. This mode would probably end when either the elevator is closed and not directed, or when it reaches the end of travel and turns around.

The fact that panel A doesn't get all the elevators but panel B does should filter through to most people over time.

The one problem with this is that some people will work out that they can get an express elevator... so you might need a key or password system to discourage this.

• The problem is, if someone who needs elevator 1 approaches and the button is lit, they may assume that elevator has already been summoned, and not press anything. Perhaps make it blink, or light up in a color, unless it's been explicitly pressed? Jul 21, 2021 at 16:46
• @DoktorJ The psychology of elevators is that you press the button, even if it is already lit. Someone who needs elevator 1 will push the panel A button even if it is already lit, thus requesting elevator 1 in addition to the current panel B request. Jul 21, 2021 at 17:48
• How about having one set of call buttons with a sign "Push and hold for handicapped or roof access", and a light to indicate when that was selected. Selecting handicapped/roof access would indicate that only elevator 1 should stop at the floor. Otherwise whichever elevator was closest (which might happen to be #1) would be dispatched. Jul 22, 2021 at 20:56
• @DavidG. if a UPS delivery person or a visitor with a walker comes in, how are they going to just "know" they're supposed to push the button on panel A when it's already lit? A lit button is well-known elevator UX for "this button has already been selected" Jul 23, 2021 at 17:40
• I think Supercat has it: the only way to stop people pushing 2 buttons is to have one (modal) button. The default mode calls any elevator. And to switch to disabled mode is just a long press. Jul 23, 2021 at 19:28

I'd put a timer / a usage sensor in the circuits for Elevator 1 during peak periods, so once its been called it will always go directly to the set destination ( the roof or the basement )

Anything to do with trying to get users to modify their pattern of behaviour is going to be too complex.

EDIT: this doesn't address the disabled usage issue which I couldn't resolve. Of all the answers and comments to the Question, a comment by Supercat is the simplest solution which cracks both problems.

• I don't think modifying user usage pattern is "complex", but that it just doesn't work. People do random stuff. Jul 21, 2021 at 1:59
• I think this will reduce the amount of time you spend waiting while on the elevator but increases the amount of time you wait for the elevator by almost the same amount. The last person to call Elevator 1 still has to wait for it to visit every other call before reaching their destination, plus all the round trips that wouldn't otherwise be needed. They just wait outside the elevator instead of inside. It'll prevent someone from waiting due to a call that came in after they board, but at the cost of increasing wait time for whoever made that call. Jul 21, 2021 at 13:17
• The problem is, how is a disabled user (or delivery person, or stroller user, etc) going to get to the third floor? Jul 21, 2021 at 16:45
• This will make elevator 1 the express elevator. Everyone will want to use the express elevator if possible. This will actually worsen the problem, by making elevator 1 more desirable an less available, increasing congestion. Jul 23, 2021 at 9:40
• Agreed, what I wrote initially doesn't solve the disabled issue. Jul 23, 2021 at 12:37

It is also slightly larger in size than the other 3 elevators, to accommodate wheelchair-using residents and strollers.

Based on the above statement (emphasis on slightly is mine) I take it that the other three elevators can be used by wheelchair-using residents and strollers.
So, on floors other than rooftop and basement, we can dedicate elevator 1 for travel to rooftop and basement only by the panel A having "rooftop" and "basement" buttons only. The internal panel in elevator A should have buttons for all floors (more on this later) but only the "rooftop" and "basement" buttons must be enabled. The only way someone can exit the lift in an intermediate floor is when someone else is entering in that floor (which is why it stops there) but this is more a matter of chance than choice.
On the rooftop & basement floors, however, the panel A can have only one button: "down" and "up" respectively. Additionally, once the lift stops in the rooftop or basement floor, all floor buttons must be enabled on the inside panel, but the intermediate floor buttons must be disabled once the lift stops at one of the intermediate floors (because someone else is entering there).
This way, only users entering from one of the dedicated floors will get to exit at the floor of their choice. The other users (entering at intermediate floors) have the choice to exit at rooftop or basement only (they already made that choice while pressing the respective button on panel A, so there's nothing to complain).

• A user could call elevator 1, ride it to either end of its' range of travel, and then push the internal button for their correct floor. Slower than the direct route the other elevators can take, but if those are crowded, it's a viable option. Jul 21, 2021 at 12:19
• @Psychemaster yes, that thought occurred to me. But as you said, that may happen only when the other three elevators are crowded. It is somewhat similar to people entering an elevator in the reverse direction just so that they can exit at their intended floor on the way back (and beyond) - when it's overcrowded while travelling in the intended direction. Jul 21, 2021 at 13:28
• Your assumption is likely incorrect in that this "slight" width increase may be the difference between a 36" door that comfortably accommodates a person rolling themselves in in a wheelchair and a 32" door that they'd scrape their knuckles on wheeling in and out. Jul 23, 2021 at 17:53

Key-switches. Replace the call buttons with key-switches at all levels except Garage, Lobby and Roof. Yes, there's a loophole, in that people can summon it to the lobby and use it as a normal elevator to ascend to any floor. But it won't be stopping empty at all intermediate floors unless somebody with the appropriate key called it there. You could prevent that by also using key-switches inside the elevator car for all floors except these three.

I propose disconnecting the panel B from elevator 1. This would at least reduce the duplicated floors from panel B.

1. When there are a lot of people waiting, as a common user I would most likely see that both panels are pressed and has higher chance of using elevator 2,3,4
2. In the minor case of going to the rooftop, basement and using the wheelchair. I would be directed to elevator 1 (either by a sign or visual cues). Since more people will be using the other 3 elevator, elevator A might have empty space. At this point elevator A users can:
• Skip the empty pressed floors, as there are less people.
• If there are people in elevator A actually need to go to the floors in between, I would say they have the fair travel time as everyone else in the elevator. In case there are wheelchair users, they may negotiate with others to go to their floor first.
• I suspect this will only exacerbate the issue that people press the buttons on both panels: "Why wait for one of three elevators to not be busy, when I could wait for one of four?" Jul 23, 2021 at 17:47

1. My building has a similar set of elevators. The #1 elevator is on the north side of the building and the #2-4 elevators are on the south side. No one presses all 4 buttons.
2. I recall a building I visited that charged \$0.10 for using the elevator. If there was a per call charge - even a nominal amount - then people would probably stop doing that. I think you would have to make trips to the ground floor free, but you could charge for other trips. It seems like you could package charging and requesting into a mobile app so no one would have to touch the elevator controls. I would be happy to pay a small fee to never again have to push elevator buttons.
• Adding an accessible elevator in a separate location can make people who need it feel segregated and "othered". Often times those are in less-convenient locations, and they're used for freight, so generally they're not as nice to be in. Jul 22, 2021 at 17:07
• @Izquierdo I agree. To be clear in my building all elevators are fully accessible. The north elevator is the freight elevator and does not service the two garage floors. If we need to take something heavy to the garage we can drive it in a truck. But if the elevators were all in a central location we would have the same problem. Some people want to go to the lobby (any elevator will do) and others want to go to the garage (only some elevators) so there would have to be two separate controls. Jul 23, 2021 at 13:14

If any solution goes

Make all four elevators equal in size and floors they attend. Then you only need a single button and everyone will be treated equally.

More realistic

Similar to a disabled toilet, which we'd all use if it's free and no other toilets are available, you would refrain if there was a disabled person inside or in line.

If you can find a way to indicate on every floor that elevator 1 is being used for roof, garage or disabled access, you would have more chance to discourage people from pressing the button.

I suspect that the other difference elevator one has is that its load limit is 1000 pounds greater than the other three. (There should be a small plaque somewhere, or a notation on the certificate.)

In short, it is the freight elevator. Large deliveries, movers, etc use this elevator (and usually get the key to lock it out). The loading dock is in the garage, so it must go down that far. Similarly, rooftop deliveries require it go to the roof.

I suspect that the real failure in the UX is that there should not even be panel A. Panel B should call all elevators any time, and any elevator not locked out will answer. When a large deliver occurs, they will get the key and call elevator 1 from either the loading dock, a special call button in the office, or a key switch on the office level. (Failing that, one can keep calling elevators and sending the wrong ones away until the freight elevator shows up.)

• Panel A is needed if you want to go to the roof. Without Panel A, I guess you would just keep summoning elevators until you get the one you needed. Jul 23, 2021 at 10:23
• Unless it's a large commercial facility that receives many deliveries (this was not my impression), all the carriers (DHL, FedEx, UPS, Amazon, USPS) are not going to be willing to add an extra key to every driver's ring on that route. Additionally, visitors with disabilities would be locked out of the very elevator that's intended for their use. Jul 23, 2021 at 17:46
• @emory I seem to recall living in such a building. If one wanted the rooftop pool, one did a normal call button, and if you didn't get the special elevator, took the stairs the last level. Jul 23, 2021 at 17:56
• For clarity: there is a freight/service elevator physically separate from these four residential elevators that is used for "freight"/etc. purposes. None of the four described above are "freight" elevators in name or practice. Jul 23, 2021 at 17:59
• @DoktorJ I think it is more likely that the carriers you refer to will just leave stuff in the office. I can't see them visiting a dozen floors to place 30 packages in the hallways. Can you? The large deliveries I was referring to is somebody getting a bed, or a soft and chairs. Large bulky objects. These delivery people will check in at the office, and can borrow a key, or even be escorted (by the resident if nothing else). Jul 23, 2021 at 18:01

Is it really necessary to take Elevator 1 directly from any floor if your eventual destination is the carpark or rooftop?

What about simply changing elevators on the ground or top floor? So there's no "call rooftop" or "call carpark" buttons on the floors.

Instead every elevator has "carpark" and "rooftop" buttons, but if it doesn't actually go there, the system will instead arrange a rendezvous with elevator 1 on any suitable floor - either one that both elevators are already going to stop at, otherwise the last reachable floor in that direction of travel.

And play an announcement "For access to the carpark please disembark on this floor and transfer to elevator 1".

A back-of-the-envelope calculation suggests that this might actually be the faster way to get to the carpark or roottop on average, assuming the building has at least 4× as many floors as elevators.

In a sufficiently tall building, this would also be the fastest way to get between floors generally.

(Of course, the "disability" access buttons would still call elevator 1, but now they could be signposted as exclusively for that purpose.)

(edit) Footnote: to make this a near-optimal solution, the system would need to schedule the elevators so that there's at least one running in the same direction as elevator 1 slightly before it. By which I mean, short enough that a fast person transferring would find elevator 1 arriving just as they step up to it, and long enough that a slow person transferring would get to elevator 1 before it decides to depart, and that minimal additional waiting time would be required.

Also, you would want to block the "close door" button in E1 while it's waiting for people to transfer.

(edit2) Clarification:

There's only 2 ordinary call buttons on each floor ("up" and "down"), and the "wheelchair" call buttons won't bring you an "extra" elevator. Rather, if you press one of those, you'll wait until either the larger elevator #1 is available, or an empty normal elevator is available. So you'll wait longer, but since the normal elevators can only take 1 wheelchair, you'll get an express ride when it does arrive. If you get elevator #1 then it might stop for other passengers, but fewer.

• Interesting idea, but this doesn't help the root problem that Elevator 1 will very often need to stop unnecessarily at floors that have already been served by Elevators 2-4. This change won't result in people pressing the Elevator 1 button any less than they do now, so you'll usually have a long wait to go the last floor. You could make Elevator 1 go directly to the top/bottom when Roof or Garage are pressed in the other elevators, but that'll be very weird behavior for anyone else on Elevator 1 who's not going to the Roof or Garage. Jul 23, 2021 at 12:33
• @NuclearHoagie Sorry, I should have made clearer, there's only 2 call buttons on each floor, "up" and "down". Jul 24, 2021 at 18:56

I'm going to add yet another option, to address a separate aspect of the problem. (If you think this doesn't cover something, please read my other two answers before commenting to that effect.)

Part of the reason people hit both call buttons is because they can and without thinking.

Even if the buttons have different labels, they have the same affordance, requiring the same arm & hand movement, and are just a few seconds apart as one walks along the corridor.

I would fix this by changing one of the pairs of call buttons to something that requires a different physical action; for example:

• a lever that you raise or lower against a pneumatic arrestor so that it takes 2 seconds to move the lever far enough; or
• a knob that you rotate 60°, with a speed limiter (about 5 rpm or 30°/s) so you can't simply spin it; or
• a kick-plate at knee level that you have to hold for 2 seconds; or
• two pressure pads on the floor, each 250mm wide and 550mm apart, that you have stand on simultaneously (or park your wheels on); or
• a smart pressure mat that recognizes the difference between rolling and stepping;
• or even just a button that you have to hold for 5 seconds.

There are lots of options, but the important part is that you want something that you cannot activate while continuing to walk, and which requires a different muscular action from tapping the regular buttons (a long hold is sufficiently different from a sharp tap for this purpose).

Of course this won't be perfect, but the mental discordance of having to take two actions with a bodily stop should encourage most people to think about what they're doing and why.

Have only one panel (or 2 or more panels with same functionality) and buttons:

ROOF
UP
DOWN
BASEMENT
WHEELCHAIR

Using appropriate symbols (see other answers).

All elevators are numbered and elevator 1 marked with symbols for wheelchair, basement and roof (on the door and the floor directly below it - so as not to get confused with the panels).

Only if roof, basement or wheelchair is pressed, will evelevator 1 stop. Otherwise elevator 2, 3 or 4 stops.

Additionally, people with wheelchairs who live or work in the building get access to the extra service elevator.

Apart from that (and slightly off-topic): Walking and climbing stairs is healthy. Encourage people to use the stairs. Using stairs instead of elevator may effectively save time, especially if you compare it with time wasted waiting for and standing in elevator and time wasted sitting in the car and driving to the gym where you get your workout on a stepper. Absurdistan. This will free the elevator for people who actually need it.

Building on @eskew 's excellent answer, as long as privacy is not a concern, why not simply move the full floor selector panel entirely to the outside of the elevator, instead of having multiple directional buttons on the outside and floor buttons on the inside? That would reduce input to a single interaction.

Pros:

• Every panel on every floor would have every floor being stopped at illuminated. If yours is already illuminated, you immediately know an elevator is already on the way, and you may not need to touch the panel at all.

• There's no way to "game" the system by pressing both sets of buttons, potentially reducing wasted energy of having two elevators show up unnecessarily. The computer will automatically bring any closest bay to your floor.

Cons:

• Privacy could be a big concern. In case you don't want that creepy dude behind you knowing which floor you're going to, there's not really any way to hide which floor you selected. Then again, he may just as well jump in the bay with you regardless of how the elevator is designed.

• I guess you'll also need a panel on the inside, because Forgetful Joe will invariably forget which floor he needs to go to, or which level he parked on.

• If a special elevator is required, you would still need some sort of way to request that one specifically. Perhaps with a specialized button to call for that one specifically.

• This is a good idea, but say there is a large crowd in the lobby waiting to go to multiple floors and I want to go to the 45th floor. Nearly simultaneously a bunch of elevators arrive. How will I know which elevator to enter? I assume one has been pre-assigned to service my request and the others not. Jul 23, 2021 at 10:50
• Great observation @emory! What if we put the bay number on the button of the floor you choose. So there would be a hard number next to the button for the floor like usual, but on the face of the button, it would show the bay number of the elevator going to that floor. If there's a duplicate panel on the inside still, you could always hit the floor button there also in case you discover that you've accidentally walked into the wrong bay. Jul 23, 2021 at 14:33
• The hotel I'm staying at has an external floor selector. When you select a floor, it'll confirm your floor selection and assign you an elevator. The side of the elevator has all stops listed, so you can double check your floor when boarding. The inside panel only contains emergency buttons. These elevators integrate with RFID room keys. Perhaps it would be possible to save a "home floor" and disability preference on your key so that that buttons wouldn't be needed for normal operations (you could have quick selects for parking/rooftop, etc. and a full floor list buried a bit deeper). Aug 1, 2021 at 8:55

This cannot just be solved with UX as your users already know. If you kicked everyone in the building out, then some of the other answers would work for sure. To alter your knowledgeable users you need to change the UX and the functionality and reprogram them.

My suggestion is a little bit simplistic.

It would include a brief bulletin or plaque outside the elevator saying the elevator's priority is servicing roof, garage, and handicap access.

In the elevator (most have this ability) you would need one button turned to a handicap button. This would allow someone in a wheelchair to press the button and then their floor.

Here is the special part. You have elevator 1 reprogrammed to give absolute priority to the handicap floors, roof, and garage. So if I come on the elevator and press floor 12 and we are on floor 3, and then another person gets on after and presses "Roof"... we go to the roof.

And if someone gets on, on the roof and presses "Garage", it then goes to the garage, even though it has passed floor 12 twice! If this happens to a person 1-2 times they will surely never get on that elevator again unless it is for intended use.

We have done this similar arrange for our "moving/freight" elevators.

The catch - anyone could press the handicap button then their floor. I don't really see this as an issue as most people would be embarrassed to do this. Also if it were referenced that the elevators were monitored I would think it would discourage almost everyone.

• Looking at the many possible solutions, I think "this cannot be solved with UX" is wrong. The problem is that the users DON'T KNOW that their behavior brings them no benefit, they believe it helps, like pressing the tv remote buttons really hard, or pressing a button a hundred times to make the elevator come faster. And this can in fact be addressed by UX. Jul 23, 2021 at 9:44