Author: Jochen Wolters

Another person and me get on a hotel elevator. I press the button for my floor, she presses the one for hers. When the elevator stops at the floor that she selected, a very puzzled look appears on her face. Then she presses the button for one floor higher up. Apparently, she was unsure what button she had to press to get to the desired floor.

Here’s the control panel from that elevator:

Look at it in its entirety, and you can easily see that the button for a certain floor is positioned to the right of its label.

The curse of a regular grid

If you focus in on any buttons that are located near the center of the panel, however, it becomes difficult to tell whether the button on the left or on the right of a label is the one to press to get to the desired floor.

Both labels and buttons are positioned on a more or less regular grid, so you cannot easily make out the pairings between the two.

If, once you press a button, the number on a label would light up, you could at least quickly see whether you had made the appropriate choice. Instead, rings around the buttons light up to confirm your selection, as seen is the top-most photograph.

Even more challenging for vision-impaired users

The labels feature braille numbers underneath the digits, which explains the digits’ odd vertical offset. Despite this welcome feature, I would expect blind users to have even more problems with using this panel, because there are no tangible connections between the labels and the buttons.

So, to understand which button belongs to which label, blind users would have to first find the edges of the panel, understand that the panel starts with label plaques on the left and ends in buttons on the right, and, keeping this knowledge in mind, find the right label/button pair on the panel.

Let’s closure this case

Let’s assume that the designers chose the regular grid layout on purpose. In that case, adding more horizontal space between label/button columns to make the panel more user-friendly probably is not an option. Placing the labels right on top of the buttons would also improve on this design, but how about another approach.

With a thankful nod to the Gestalt Law of Closure, I would change the shape of the number labels like this:

This firmly groups paired labels and buttons together visually. A grid of these labels and buttons with the same spacing as the original would look like this:

By adding a bezel to the label plaque, or by embossing the entire plaque, blind users could also feel the connection between label and button.

Placing the braille dots between the number and the button would achieve two things: the vertical position of the number is more visually pleasing, and after reading the braille pattern, blind users’ fingers would automatically end up properly on the neighboring button.

Update 2011-02-08: This building sports a total of six guest elevators, but they are not created equal: one of the six can take its passengers to an additional floor that remains hidden from view in the other five elevators.

After using the elevators for a few days, my mind had developed a mental model of them, part of which postulated: “To go to the lobby, press the single button that is located, horizontally centered, right below the large panel of buttons.”

One day I got on the elevator and, relying on this mental model, intuitively reached for the 1st-floor button and pushed it as I would normally do. Only this time, I realized that nothing was happening: out of the corner of my eye, I noticed that the ring around the button did not light up even after I had pushed the button repeatedly.

Taking a closer look at the button panel, I now saw that, on this one elevator, there was an additional button labeled B1. Not only was this button placed on the same line as the one for the first floor: in an attempt to preserve the panel’s more or less symmetrical appearance, the 1st-floor button had also been moved one column to the right.

As I have written previously, moving controls in a user interface oftentimes creates annoying usability problems. Although the effects in this case are not quite as severe as the ones explained in that earlier blog post, the button for the 1st floor should be found in the same location on all six elevators.

If the majority of users doesn’t have access to a certain feature, hide it

As for the B1 button, I assume that hotel guests ususally cannot even go to that floor. Otherwise, why did the button refuse to accept my (erroneous) request to go to that floor? If it is necessary to somehow “unlock” the B1 floor before being able to use the associated button, why implement it as a button in the first place?

By looking and generally behaving like any other floor button, users expect that it can also be used just like the other ones.

If using a key in one of the locks seen on the panel is required to successfully use that button, one option would be to implement its functionality as part of that very lock. Assuming that this lock controls more than just the B1 floor button, it could have three positions: 1 – special access locked, 2 – special access unlocked, 3 – Go to floor B1. Position three would operate just like the “start engine” position on a car’s ignition lock, i.e., after selecting it, the key would be pushed back to position number 2 by a spring.

This implementation would remove that ominous B1 floor from the users’ view and mind, and prevent accidentally hitting this button when all you want to do is go to the lobby.

Update 2011-02-28: Different hotel, different elevator — different button labels!

This elevator sports much fewer buttons, making it possible to grasp the entire button panel at a single glance. It is much less challenging to make out the pairings between buttons and labels in this case.

And yet, whoever designed this, seems to share my point of view that labels should provide a strong visual cue about which button they belong to:

What I find surprising, is that the buttons have been crammed together like this. Laying out the floor buttons in a six-high/one-wide column would have made for better mapping; adding more space between the floor, door, and emergency buttons would have created a stronger division into the three functional groups.

In his seminal book, “The Design of Everyday Things,” Donald Norman introduces the term “affordance” which …

… refers to the perceived and actual properties of the thing, primarily those fundamental properties that determine just how the thing could possibly be used. A chair affords (“is for”) support and, therefore, affords sitting. […]

Affordances provide strong clues to the operations of things. Plates are for pushing, Knobs are for turning. Slots are for inserting things into. Balls are for throwing and bouncing. When affordances are taken advantage of, the user knows what to do just by looking: no picture, label, or instruction required. […] When simple things need pictures, labels, or instructions, the design has failed.

How bad design makes a simple device hard to understand

A coffee pump pot, the likes of which can be found at hotel breakfast buffets the world over, easily falls into the “simple things” category: all you need to do to use it is to place a container under the spout and press a button to dispense the breakfast beverage of choice into said container.

Why, then, did someone feel the need to apply a sticker screaming “PUSH HERE” to this pot’s pump button? Is this an indicator that “the design has failed,” as per Norman’s quote above? Let’s see…

When I tried to use this pot, I first pushed on the location of the “Regular [Coffee]” label. When that didn’t work, I tried the indented area just above it, only to realize that doing so unlocks the lid.

It took a while for me to even see the “PUSH HERE” sticker and understand that it denotes the pump button.

The real-life implications of “affordances”

This pot has a total of three UI elements that afford pushing: the indented area and the brown dot on the unlock lever, and the pump button. So why did I assume that, of these three, the brown dot with the “Regular” label on it was the control I had to use to get some coffee out of the pot? There are several explanations:

• Stickers aside, the brown dot is the only element of the pot that stands out visually, whereas the actual pump button blends in so well that it almost looks like a lid. Therefore, the perceived affordance of the brown dot to operate like a push button is stronger than that of the pump button.

• The brown dot is closer to the spout than the pump button. Therefore, its logical link to the spout is stronger than that of the pump button.¹

• The placement of the “Regular” label further strengthens the visual cue that the brown dot needs to be pressed to get coffee: In every desktop or mobile computer UI, a control’s label is placed so close to the control itself that the term “intimate relationship” comes to mind.

  In the case of, e.g., menus, menu items, and buttons, the label appears right on top of the widget. Consequently, the label on top of the brown dot creates the same association as a similar label on a button in a software application: “[Press here to get] Regular [coffee]!”

• When placing the cup underneath the spout, you tend to focus your view on that area, so it is easy to overlook the “push here” sticker on the pump button that is a bit farther away.

Usable without explicit instructions

Of all the other guests I observed using this coffee pump pot, a few immediately knew how to operate it. I assume that these users simply were already familiar with this specific model. Most others, however, did struggle until they, too, eventually worked out the intricacies of this device.

If I were tasked with improving this design to make it more accessible to experienced and inexperienced users alike, I’d try this:

• Make the pump button stand out by placing a large brown dot and an easy-to-read label on the button’s top stating the type of coffee.

• Make the unlock lever blend in and, thus, easy to overlook, by removing the brown dot and the “Regular” sticker. Maybe even reduce its size or move it to a slightly less prominent location.

In other words, guide the user’s gaze towards the pump button and away from the unlock lever to fine-tune the perceived push affordance of either control.

During recent travels, I noticed a novel design for shower curtains. Whereas most curtains are mounted to their rod with eyelet hangers made from metal wire or plastic, these curtains utilize plastic disks that are sewn into the curtain’s fabric.

These disks have a horizontal slot at one side; slots in neighboring disks are pointing towards each other and the fabric between the slots is cut as well. In other words, the center holes in two neighboring disks are “connected” by these cuts and slots.

Pro: easy-on, easy-off

Thanks to this nifty design, removing the curtain from its rod is very easy: open the curtain all the way, grab a stack of disks and bend them to open the slots, and then slide them off the rod.

Standard eyelet hangers usually need to be taken off the rod one hanger at a time. In contrast, with this new design, several hangers can be removed at once, saving hotel staff quite a bit of time and effort when exchanging the curtains during cleanup.

Con: difficult-to-close

When you try to close the curtain, however, you realize that the design is flawed.

You can easily pull the curtain along from one end until it is about two thirds down the rod. At that point, the disks will start to lie flat against the rod, which generates a surprising amount of friction between them.

In fact, the friction will increase to the point where you cannot move the curtain any further along unless you start pulling at the curtain’s center section to move the remaining folds along.

Optimized for the less-common use case

There are two groups of “users” for a hotel shower curtain: the guests who will open and close the curtain when taking a shower, and the service staff who will exchange the curtain for cleaning.

Compared to curtains that are mounted the “old-fashioned” way with eyelet hangers, this disk-hanger design makes it more difficult for the guests to close the curtain while making it easier for the service staff to remove and replace it.

But how often will a curtain like this be opened or closed, and how often will it be replaced?

Assuming that most hotel guests will take a daily shower or bath, and that curtains are only exchanged when new guests move into the room, this curtain will be closed and opened many more times than it is removed from its mounting rod. This is even more true for rooms that are occupied by more than one person.

I think it is fair to say that, while this new design shows off an interesting and inspiring fresh approach, it effectively has been optimized for the less-common use at the expense of its main usage.

What’s more, as much as I appreciate making things easier for service personnel: annoying hotel guests due to having optimized the design of an everyday thing like a shower curtain for non-guest user groups, is a very risky decision.