Redesigning the smartphone dial pad
uxdesign.cc
uxdesign.cc
I don't like a redesign because it's simply too ingrained. It's a sailed ship.
Note that our screens are big enough to waste space to nothing, at least for a phone dial screen. So we can redesign it to:
___________________________
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| 1 2 3 |
| |
| 4 5 6 |
| |
| 7 8 9 |
| |
| 0 |
| |
|_________________________|As a side note, we make up ~10 percent of all users. If you design to the thumb reach of right handed users by default, you're fucking over a lot of people.
Add on that ~19 percent of the population has some form of disability, and a solid chunk of that involves hand/limb usages, making the head-up-ass assumption that a right handed thumb map is the "ideal" approach is... well you've shoved your head up your ass.
----
edit: I want to add, I'm not attacking you so much as the general premise of the article. I come off as fairly hostile on this since it's such a common day to day inconvenience for a broad swath of the population.
Don't even get me started on various other items designed for right-handed people. Like this popcorn scooper I found at a nearby convenience store: http://a.co/jjanj7H I've even run in to butter spreading knives that are impossible to use left-handed.
Being left-handed is already a UX nightmare. Please don't make it worse.
Aside: I discovered touching a navbar button often doesn't work for a left hander, whereas it does work for a right handed person.
This left vs right handed UX flaw occurred with the navbar button at top left on an iPhone (e.g. back or hamburger).
A left finger comes in at an angle that makes the capacitive touchscreen register the touch as occurring at the extreme left edge of the screen. A right index finger registers the touch more correctly within the navbar button. Some navbar buttons don't extend the touch hit area to the edge, so a left finger fails to click it, whereas a right finger works...
But still, the thumbs are probably the least "handed" element of the body. I mix right and left thumb use based on whatever I need to be doing with my other hand.
Assuming the hearmap is the mirror image for left handed folks.
You want eventually the navbar mirrored, but you surely don’t want to read 321, 654, 987 on your number pad. Being lefty doesn’t make your brain read rtl.
There is no quick wins there.
Hold your horses!
Just use the left side of the keyboard as a toggle button...
when you click it once... the keyboard switches over to a left handed position and stays there until someone else switches it back to the right handed position!
My daily driver is an iPhone SE, I have average / medium sized hands and can reach the whole screen well enough.
1. http://androidaustralia-zippykid.netdna-ssl.com/wp-content/u...
I think you're probably right, but i'm glad somebody is taking the time to experiment anyways.
We already use the keyboard to type numbers and it works much better than a full screen dial pad.
Part of the reason the current layout works is because it is well chunked. You have 123 on top, 456 in the middle, and 789 on the bottom. 0 and functional keys, which often serve their own purpose, get their own row. Four sets of three, nicely like the way our minds like to remember numerals and sets of numerals. Shape-wise, 1-9 are in a square, in numerical order from left to right just like a paragraph, then the others continue the pattern to a lesser extent. This creates a very easy and forgiving mental map because it is a simple shape, the full width (more or less) of the display or device, with a familiar pattern.
In contrast, 4 of these designs have different numbers of numbers on each row, some with several different widths. You'd have to remember, was it three numbers in the first row, or two? How many in the second? Was that one with an extra button on the side?
So to start these designs don't seem to consider why the original design is successful.
Next they don't seem to include the context of the thumb thing. We often hold our phone in one hand because precision is not required for most apps, for example Instagram where all functionality is within that bottom easy zone. But other apps require precision, many games for instance, and it's likely people change their grip frequently for those contexts. Is there data on that specifically? Anecdotally I seem to remember most people I know holding with one hand and dialing with the other. So the reach of the thumb isn't as big a consideration as the designers might think.
And as others have pointed out it won't save much time, since hardly anyone uses the dialpad except the occasional adding of a contact or dialing of a conference line.
I don't mean to rag on this (rereading, I sounded more critical than I am), I think it's fun to try to redesign stuff. But it reminds me of when people try to remake the ordinary wall outlet. They start modifying it without really considering how deceptively well designed it is.
We can recognize QR codes, heck, even do AR on a phone. Surely identifying phone numbers via the camera and inputting them that way would be easy and a measurable improvement over the current method of mashing buttons.
Since dialing a number by hand is such an infrequent operation, inventing some wonkball new standard for it is a super bad plan. There's a lot of other things you could innovate on and come up with real improvements, like how you might make a better audio playback control system, or re-visiting mixing tools (equalizer, cross-fader, effects filters) for music to make them easier to understand on an intuitive level.
The dial pad is fine. Leave it alone.
Not many people really dial phone numbers these days though - dialing via contacts, history, a link on a website or Google Maps are far more common in my experience.
The issue is already pretty much solved without the need for any OCR - of course Google Goggles does do that too if you need it.
i did
https://github.com/ransomw/simp-phone/blob/master/client/app...
... and it still turned out to be a convoluted story of code design.
perhaps the results will be useful to anyone considering pursing the UI design end of things.
What if, worse, they're only familar with your pad and are confronted with an actual phone that uses a totally different system?
What if this isn't a dial pad, but an interface for a car or heavy equipment? What if, in a moment of panic, someone's muscle memory kicks in and they do what they've always done to avoid disaster, only in your system it does the opposite?
Switching from rotary dial to a dial pad was a big deal, and whatever system they chose for the dial pad out of all those kooky options would be the standard. If they had them in random order for whatever reason, then leave it alone.
I see this all the time when people think they're "helping" by showing a keyboard in alphabetical order. No, not helping.
It used to be that every single country would have one or more power plug types. Over the years some have faded out, sometimes reluctantly, sometimes stubbornly, but having fewer standards is almost always better.
The long-tail of UX is misunderstood indeed. Just because you can redesign something doesn't mean you should.
What if they're handicapped, and only have one finger and the new layout allows them to dial faster? The current design would have failed them. Every design works within the constraints that are given. Creating cases where a particular design works/doesn't work isn't hard or interesting, unless you actually do something about it (which is what this article is proposing).
>The long-tail of UX is misunderstood indeed.
Then I would invite you to study it some more.
>Just because you can redesign something doesn't mean you should.
No reading of what I said, or the linked article has anything even remotely close to suggesting that. Maybe you could have asked for clarification since your reply seems very confused to me.
I am very much a UX enthusiast if not long-time student of that field. I am a constant proponent for usability features.
These patently awful dial pads help nobody but the authors of the blog post fishing for clicks. They're junk. They're worse than useless.
Apple made a "one handed keyboard" for iOS 11 (http://appleinsider.com/articles/17/09/21/inside-ios-11-appl...) and it shows how it can be done. They scrunched it down a bit to make it more convenient to use. They did not arbitrarily rearrange everything into an unfamiliar layout and force that layout on everyone who wanted to use that feature.
Another example is the Matias "half" keyboard (http://matias.ca/halfkeyboard/) which comes in two varieties: Minimal and full-sized. Both offer similar features and are intended to be used with one hand. Note they didn't jumble up the keys, either, they made it as conventional and familiar as possible while still achieving usability goals.
However, the consistency of rows I think only makes sense with physical buttons where you can feel your way across them. With a flat panel, we don't really have the feedback anyway, so you can't touch your way down to the next row and know where you are.
Having said all that, I'm not a fan of these designs, but I am a fan of the designer taking on the challenge.
The future is more likely dialing by voice than it is a re-arranged button layout.
My opinion, what a monumental waste of time. I'm damn happy that I read tbgvi's comment before looking at the article, as the few minutes I spent browsing over it seem a waste.
What's next, let's design coat buttons to allow people to get dressed 0.5 seconds faster(after of course speeding hours relearning their muscle memory to actually do it -and don't forget the patented new button design giving them 20% of all sales of course)
I can't remember the last time I used a phone number for anything. Contacting friends is mostly done via social media. Talking on the phone is done via voice or video via same social media. Contacting companies? Via e-mail or their own webpage with some kind of chat support.
I only have a phone number, so that people without smart phones (my grandmother) can contact me.
And 2FA!
Daily.
> Contacting friends is mostly done via social media. Talking on the phone is done via voice or video via same social media.
My friends aren't reliably on social media (and some of them aren't at all). SMS is quickest for small things, an actual phone call is quickest for a slightly longer discussion. E-mail for a group discussion or long-term planning.
> Contacting companies? Via e-mail or their own webpage with some kind of chat support.
I start with their online help or ordering system. For a tech company, e-mail might be the next option. For food, or something where I'm contacting someone local, it's almost always a phone call.
I'd have to think about the last time I actually dialed a number by hand, though.
If I meet a new person I enter 10 digits and that is the extent of my usage. Not to say that the design in use is justified by limited use, but my point is that dialing a number is already antiquated. Most of the time I dial something from a saved/shared contact, maps, or from a link in an email or on a website - never actually typing in a number.
For example, as others commented, many users primarily 'dial' their smartphone nowadays by tapping on somebody's face or going through their contact information. The cost of changing the experience may not be worth the added benefit. Along those same lines, it may well be that most dial pad users are not the owner of the phone, but somebody who borrows it to make a call or enter their own info. In that case, the experience should be familiar to them, regardless of whether the user has grown accustomed to a new dial pad. These are the kinds of factors I'd like to see discussed in researching a new dial pad experience.
Resistance to change on something like this is real. In developing a web-based phone app, people who were using computers, with keyboards, would still prefer to click on the buttons in the style of a traditional dial pad. To me, investigating those challenges are much more interesting than "Here are some designs; let's see which you say you like."
I wouldn't say dialing a new number is a common use case for me. But that doesn't mean people don't do it often enough to warrant research.
These generic "solved" UX issues are still things that millions of people do each day, given a billion phones, so it's well worth a modern UX study IMO.
The dial-pad may be too difficult to reach on our big screens now, but the momentum required to change it will never exist because it simply doesn't get enough use to be a pain point anymore.
Excellent write up and use of graphics though. The Bell labs part was really interesting.
Replacing the dial pad with the numpad keyboard would solve the problem and not require custom layout. It would also be more consistent across applications that do calls instead of having a special dial pad for the phone app.
The article glosses over the fact that the layout we use now was not not the best in the Bell study. As you can see from the image in the article [1] the traditional circular layout was the only one that had both significantly shorter keying time and significantly lower error rate. The other source they link to [2] says that "Performance and preference differences, though, were deemed to be fairly small overall across the five finalists, so for engineering reasons Bell went with the layout we know today."
[1] https://archive.org/stream/bstj39-4-995#page/n3/mode/2up
[2] https://99percentinvisible.org/article/squaring-circle-seven...
The image in the article shows nothing of the kind, as you might notice if you ask "lower than what?" On your analysis, layout IV-A had a "significantly shorter keying time" than layout VI-A, but that's nonsense -- those are the same layout.
The answer is that "significantly lower error rate" and "significantly shorter keying time" refer to statistics taken within a comparison group of three layouts, and don't apply to comparisons between a layout from one group and another layout (or the same layout!) from a different group.
You are right in that the comparison was only between the group of three layouts in a row. From the image no conclusion about the performance of the layouts in different rows can be drawn.
However, in the same study, they did a direct comparison between the grid and and the circular layouts and here are the numbers:
KEYING TIME PER CENT
(SECONDS) ERRORS
THREE-BY-THREE PLUS ONE 6.01 2.5
TELEPHONE 5.90 2.0
SPEEDOMETER 5.97 3.0
The oppprtunities for improvement here are minuscule, and nonexistent if you take into account familiarity and expectation. We should save this kind of blue sky redesign for new or less solved design problems and stop making people’s experience worse in pursuit of some noble but misguided strive for constant and relentless disruption.
I would probably still center it though, since we don't know which hand is holding the phone.
Because the "register a number as dialed" mechanism was different, and didn't require people moving their fingers in a circular motion -- so they've made the buttons equidistant.
>He suggests that most of us hold the phone from the bottom such that the base of our thumb is at the bottom right of the screen (considering 90% of population is right handed).
49% is not most of us -- most of us use something else (according to the diagram). It's just the way with the most users, not the way most people use (subtle difference: the other 2 popular ways amount for 51% of users).
What's more important, the other 2 ways are not some outliers, or severely fragmented styles, but 2 holding styles will major followings. So those should be catered too as well.
One-handed: clicking audio controls, scrolling through news or similar content
One hand holding, one hand tapping: more precise input of things like phone numbers, browsing web pages, etc.
Two hands holding: typing long messages, some games.
Maybe it's because I have a larger phone (5.5"), but I wouldn't do any input task more than a single button press or scroll using my thumb. I always bust out the other hand for that.
I know I'm overlooking existing infrastructure and countries with fewer smartphones but we need to leave this behind at some point.
Without each side being unique, how would voice://John_Smith-00003475 be better than a phone number? We already have address abstraction in the form of contact avatars on our phones. Click John's face, then talk. And since this is done locally, I know that a wrong number is just my fault and easy to fix. Choose any reason why a central or even distributed resolution contains an error and suddenly you're back to keying phone numbers just to make a call while we wait for corrections to be issued by, whom, exactly? My parents don't even know the difference between a cellular connection and a WiFi connection on their phones, are they supposed to be responsible for their own online contact records too?
They have tried almost all the concepts from that bell study.
Especially interesting is 7610 optimized for thumb reach. Other funny ones are 7600, 3200, 3650 and Ngage QD
I'm more or less right-handed - fairly ambidextrous for many things - but I write with my right hand.
I always hold a phone or similar device with my left hand and tap on it with my right hand, like the middle example in the three illustrations on the blue background midway through the article (the one that it says 36% of users do it this way).
It never once occurred to me to hold a phone with my right hand and tap on it with my right thumb, like the leftmost illustration in that image (49% of users).
If I do tap with my thumb, I still have the phone in my left hand and use my left thumb.
I think this comes from my earlier days with pads of paper and a pen or pencil. I wanted to write with my right hand, so naturally I held the pad in my left hand and the pen in my right hand.
Of course I've always assumed everyone does it this way: hold the device in your non-dominant hand and tap on it (or write/draw on it with a stylus) with the dominant hand.
It's a total surprise to me that people would actually hold a phone or device with their dominant hand. I guess it makes sense if you're always using it in a one-handed manner, but for me it is just something that never even came to mind as something to try. Since I'm holding it in my left hand anyway, if I do use a thumb it will be my left thumb.
Not saying one way or the other is right or wrong, of course. I wonder if this is one of those things where people fall into one of two (or more) groups, and don't even know the other groups exist?
Also I find that doing standard dialing is pretty rare these days, and so low mental load, that I think I wasted more brain cycles reading this than I have ever just dialing a number.
It seems to be that a better way to measure results would be to create an Android only app (iOS doesn't support swapping primary dialer). and getting participants to use your dialer for a couple months, once you notice a new baseline per user compare results between the various treatments.
Tl;dr unless I missed something! these results aren't going to prove anything, other than familiarity rules.
All that aside -- I LOVE this! In the spirit of experimenting and trying to improve on a very old concept, and creating an app to whip up some quick results, very, very cool!
The telephone dial pad is as familiar as the QWERTY keyboard, and attempting to change that -- especially for something that people use less than ever -- seems futile to me. That said, the underlying point of increasing screen sizes leading to hard-to-reach dial pads is valid.
What I'm missing though is that to me there's a painfully obvious solution: make the existing dial pad smaller. Nothing says it has to take up the entire physical screen space. Take the existing, familiar layout, and resize and reposition it to be within the "natural thumb arc" area. This does mean different size screens need different layouts, but that seems like a trivial detail.
I'd take a couple of thoughts.
Most of my dailing is done via voice, phone URL or stored contact.
Most of my time in dialpad is spent entering digits for automated systems (e.g. "press 1 to join meeting" or "dial meeting code then #" or "press 0 to talk to a receptionist").
What dedicated call vs. hangup buttons?
For example, speaking to your cell phone company, you could see:
1 [Billing] 2 [Account] 3 [Add a line]
4 [Problem] 5 [Option5] 6 [Option6]
7 [Option7] 8 [Option8] 9 [Option9]
0 [Operator]
And heck, let's get rid of phone queues while we're at it - just have the agent trigger a callback rather than making them listen to an awful hold music loop for potentially upwards of 15 minutes.
Here, though, I'm very suspicious. I worry that this is in the same category as the designers who come up with yet another novel way of making a hotel shower work. Is it really better? Probably not. And if it were, it's still different than every other shower out there, meaning that most people are very unlikely to gain the level of proficiency at which the benefits would kick in.
Pre-smartphone, I used the dial pad a lot, from 3 to dozens of times per day. Now, I might use it a few times per week. So is it really worth the effort to rework all these interfaces and retrain everybody? I'm skeptical.
I'd be soooo happy though if phones and keyboard numpads had the same layout (rather than being inverted horizontally. No preference for either, just want them to be the same.)
Alas.
Some do (the Galaxy Note series). Having used a stylus with a PDA in the past, I can also see the disadvantages: the stylus is easy to lose, and can't be used one-handed.
And then I referenced the graphical representation [1] of the study's results.
Y'all. 49% is not "most". 49% * 90% ~= 45%, so even less than that.
[1] https://cdn-images-1.medium.com/max/2000/1*RoVzWTb90D3fiEjlZ...
A few user-selectable preset configurations + strictly user-defined option is the only sensible direction. That this wasn't implemented 10 years ago suggests a bit about its pragmatic value.
I receive more calls than I make, and when I do place a call I almost always do so through my address book or favorites. So while I think this is a great experiment I'm not sure its actually that useful.
I am used to the placement of the buttons too so having them in a square pattern is important. Moving them around in any way without moving the entire square cluster is going to feel awkward.
Interested to see UX where the send button is in the middle :)
That raises some questions.
First, why is there a difference? For any given sequence of numbers both of those layouts requires the same amount of movement, just mirrored around the 456 line. There should be no physical reason for one to work better than the other, so it would seem it is something mental. What is that?
My guess is that it has something to do with our left to right then top to bottom reading order. If someone speaks out loud the digits from 1 through 9 in order and you are asked to write them down in a 3 x 3 matrix as they say them you will probably fill them in reading order, giving 123/456/789.
Second, it would be interesting to see what would happen if the order within the rows were reversed, 321/654/987 vs. 987/654/321. My guess is that the latter would beat the former. It gives a less complicated sequence when accessed in reading order.
Third, I wonder which layout would have won for calculators if the calculator makers had tested like Bell Labs did. Let's assume it is a given that we have to go with either 123/456/789 or 789/456/123, and just need to determine which works best in a calculator.
With the phone all you need to operate it are the digits (the first touch tone phones did not have the '#' and '*' keys).
With the calculator you also need the decimal point, the operator keys, and '=' [1]. The layout of the digit keys needs to fit well with the layout of the non-digit keys.
I'd guess that '=' is the most used key, with '+' the second most used. Third is probably '-', although I'd not be shocked if it was 'x' because it will see a lot of use in percentage calculations which might be common enough to put it ahead of '-'.
Because '=' and '+' are so important, it makes sense to put them somewhere that can be identified by feel, and the best place for that is probably one of the bottom corners.
If that's where we put '=' and '+', then I expect calculator order in the 3x3 digit grid would win, because of Benford's law [2]. Calculator order would put the smaller digits closer to '=' and '+' than telephone order would.
[1] or 'ENTER' if you are civilized.
Left to right and top to bottom does make logical sense, but is 3x3 the best possible solution? Even if it is, should we not reconsider placing those buttons in a more easy to reach position without compromising the muscle memory that users have developed using this old layout?
The experiment so far is showing that it is possible, but I would wait and watch where it goes :)
Thanks for your thoughtful comments.
Edit: it seems to work fine on iPhone but not iPad.
But what I'm really talking about is not myself but the thumb diagram in the article. It seems extraordinary to design a consumer product that is known to be uncomfortable. It's almost like the South Park It bike was released. I get that there was optimisation for eyes at the expense of hands (and portability, in a mobile device!) but the point of design is to balance constraints - it's a definition of bad design throw half of them out of the window.
these phones have monster battery life, and the same flagship specs of their bigger brothers, less the high-res display. but at 4.6", 720p is plenty good and also great for battery life.
[1] https://www.gsmarena.com/sony_xperia_z5_compact-7535.php
[2] https://www.gsmarena.com/sony_xperia_xz1_compact-8610.php
[1] https://forum.xda-developers.com/z5-compact/development/rom-...
http://scotthurff.com/posts/how-to-design-for-thumbs-in-the-...
though even there I cannot find any explanation for "ow-space".
Maybe a reference to some game?
> This is not because it is working.
Err...