How Do Users Really Hold Mobile Devices?

By Steven Hoober

Published: February 18, 2013

“People can use mobile devices when they’re standing, walking, riding a bus, or doing just about anything. Users have to hold a device in a way that lets them view its screen, while providing input.”

As UX professionals, we all pay a lot of attention to users’ needs. When designing for mobile devices, we’re aware that there are some additional things that we must consider—such as how the context in which users employ their devices changes their interactions or usage patterns. [1] However, some time ago, I noticed a gap in our understanding: How do people actually carry and hold their mobile devices? These devices are not like computers that sit on people’s tables or desks. Instead, people can use mobile devices when they’re standing, walking, riding a bus, or doing just about anything. Users have to hold a device in a way that lets them view its screen, while providing input.

In the past year or so, there have been many discussions about how users hold their mobile devices—most notably Josh Clark’s. [2] But I suspect that some of what we’ve been reading may not be on track. First, we see a lot of assumptions—for example, that all people hold mobile devices with one hand because they’re the right size for that—well, at least the iPhone is. [3] Many of these discussions have assumed that people are all the same and do not adapt to different situations, which is not my experience in any area involving real people—much less with the unexpected ways in which people use mobile devices.

For years, I’ve been referring to my own research and observations on mobile device use, which indicate that people grasp their mobile phones in many ways—not always one handed. But some of my data was getting very old, so included a lot of information about hardware input methods using keyboard- and keypad-driven devices that accommodate the limited reach of fingers or thumbs. These old mobile phones differ greatly from the touchscreen devices that many are now using.

Modern Mobile Phones Are Different

“I’ve carried out a fresh study of the way people naturally hold and interact with their mobile devices.”

Everything changes with touchscreens. On today’s smartphones, almost the entire front surface is a screen. Users need to be able to see the whole screen, and may also need to touch any part of it to provide input. Since my old data was mostly from observations of users in the lab—using keyboard-centric devices in too many cases—I needed to do some new research on current devices. My data needed to be more unimpeachable, both in terms of its scale and the testing environment of my research.

So, I’ve carried out a fresh study of the way people naturally hold and interact with their mobile devices. For two months, ending on January 8, 2013, I—and a few other researchers—made 1,333 observations of people using mobile devices on the street, in airports, at bus stops, in cafes, on trains and busses—wherever we might see them. Of these people, 780 were touching the screen to scroll or to type, tap, or use other gestures to enter data. The rest were just listening to, looking at, or talking on their mobile devices.

What My Data Does Not Tell You

Before I get too far, I want to emphasize what the data from this study is not. I did not record what individuals were doing because that would have been too intrusive. Similarly, there is no demographic data about the users, and I did not try to identify their devices.

Most important, there is no count of the total number of people that we encountered. Please do not take the total number of our observations and surmise that n% of people are typing on their phone at any one moment. While we can assume that a huge percentage of all people have a mobile device, many of these devices were not visible and people weren’t interacting with them during our observations, so we could not capture this data.

Since we made our observations in public, we encountered very few tablets, so these are not part of the data set. The largest device that we captured in the data set was the Samsung Galaxy Note 2.

What We Do Know

“In over 40% of our observations, a user was interacting with a mobile phone without inputting any data via key or screen.”

In over 40% of our observations, a user was interacting with a mobile phone without inputting any data via key or screen. Figure 1 provides a visual breakdown of the data from our observations.

Figure 1—Summary of how people hold and interact with mobile phones

Summary of how people hold and interact with mobile phones

To see the complete data set:

Voice calls occupied 22% of the users, while 18.9% were engaged in passive activities—most listening to audio and some watching a video. We considered interactions to be voice calls only if users were holding their phone to their ear, so we undoubtedly counted some calls as apparent passive use.

The users who we observed touching their phone’s screens or buttons held their phones in three basic ways:

  • one handed—49%
  • cradled—36%
  • two handed—15%

While most of the people that we observed touching their screen used one hand, very large numbers also used other methods. Even the least-used case, two-handed use, is large enough that you should consider it during design.

In the following sections, I’ll describe and show a diagram of each of these methods of holding a mobile phone, along with providing some more detailed data and general observations about why I believe people hold a mobile phone in a particular way.

In Figures 2–4, the diagrams that appear on the mobile phones’ screens are approximate reach charts, in which the colors indicate what areas a user can reach with the finger or thumb to interact with the screen. Green indicates the area a user can reach easily; yellow, an area that requires a stretch; and red, an area that requires users to shift the way in which they’re holding a device. Of course, these areas are only approximate and vary for different individuals, as well as according to the specific way in which a user is holding a phone and the phone’s size.

Users Switch How They Hold a Mobile Phone

“The way in which users hold their phone is not a static state. Users change the way they’re holding their phone very often—sometimes every few seconds.”

Before I get to the details, I want to point out one more limitation of the data-gathering method that we used. The way in which users hold their phone is not a static state. Users change the way they’re holding their phone very often—sometimes every few seconds. Users’ changing the way they held their phone seemed to relate to their switching tasks. While I couldn’t always tell exactly what users were doing when they shifted the way they were holding their phone, I sometimes could look over their shoulder or see the types of gestures they were performing. Tapping, scrolling, and typing behaviors look very different from one another, so were easy to differentiate.

I have repeatedly observed cases such as individuals casually scrolling with one hand, then using their other hand to get additional reach, then switching to two-handed use to type, switching back to cradling the phone with two hands—just by not using their left hand to type anymore—tapping a few more keys, then going back to one-handed use and scrolling. Similar interactions are common.

One-Handed Use

“The 49% of users who use just one hand typically hold their phone in a variety of positions.”

While I originally expected holding and using a mobile phone with one hand to be a simple case, the 49% of users who use just one hand typically hold their phone in a variety of positions. Two of these are illustrated in Figure 2, but other positions and ways of holding a mobile phone with one hand are possible. Left-handers do the opposite.

Figure 2—Two methods of holding a touchscreen phone with one hand

Two methods of holding a touchscreen phone with one hand

Note—The thumb joint is higher in the image on the right. Some users seemed to position their hand by considering the reach they would need. For example, they would hold the phone so they could easily reach the top of the screen rather than the bottom.

One-handed use—with the

  • right thumb on the screen—67%
  • left thumb on the screen—33%

I am not sure what to make of these handedness figures. The rate of left-handedness for one-handed use doesn’t seem to correlate with the rate of left-handedness in the general population—about 10%—especially in comparison to the very different left-handed rate for cradling—21%. Other needs such as using the dominant hand—or, more specifically, the right hand—for other tasks may drive handedness. [4]

One-handed use seems to be highly correlated with users’ simultaneously performing other tasks. Many of those using one hand to hold their phone were carrying out other tasks such as carrying bags, steadying themselves when in transit, climbing stairs, opening doors, holding babies, and so on.

Cradling in Two Hands

Cradling is my term for using two hands to hold a mobile phone, but using only one hand to touch the screen or buttons.”

Cradling is my term for using two hands to hold a mobile phone, but using only one hand to touch the screen or buttons, as shown in Figure 3. The 36% of users who cradle their mobile phone use it in two different ways: with their thumb or finger. Cradling a phone in two hands gives more support than one-handed use and allows users to interact freely with their phone using either their thumb or finger.

Figure 3—The two methods of cradling a mobile phone

The two methods of cradling a mobile phone

Cradling—with a

  • thumb on the screen—72%
  • finger on the screen—28%

With thumb usage, users merely added a hand to stabilize the phone for one-handed use. A smaller percentage of users employed a second type of cradling, in which they held the phone with one hand and used a finger to interact with the screen. This is similar to the way people use pens with their mobile devices. (We observed so few people using pens with their mobile devices—only about six—that I have not included them as a separate category in the data set.)

Cradling—in the

  • left hand—79%
  • right hand—21%

Anecdotally, people often switched between one-handed use and cradling. I believe this was sometimes for situational security—such as while stepping off a curb or when being jostled by passersby—but sometimes to gain extra reach for on-screen controls outside the normal reach.

Two-Handed Use

“We traditionally associate two-handed use with typing on the QWERTY thumbboards of devices like the classic Blackberry or on slide-out keyboards.”

We traditionally associate two-handed use with typing on the QWERTY thumbboards of devices like the classic Blackberry or on slide-out keyboards. Two-handed use is prevalent among 15% of mobile phone users. In two-handed use, as shown in Figure 4, users cradle their mobile phone in their fingers and use both thumbs to provide input—much as they would on a desktop keyboard.

Two-handed use—when holding a phone

  • vertically, in portrait mode—90%
  • horizontally, in landscape mode—10%

Figure 4—Two-handed use when holding a phone vertically or horizontally

Two-handed use when holding a phone vertically or horizontally

People often switched between two-handed use and cradling, with users typing with both thumbs, then simply no longer using one hand for input and reverting to using just one of the thumbs consistently for interacting with the screen.

However, not all thumb use was for typing. Some users seemed to be adept at tapping the screen with both thumbs or just one thumb. For example, a user might scroll with the right thumb, then tap a link with the left thumb moments later.

Also notable is the overwhelming use of devices in their vertical orientation, or portrait mode—despite theories about the ease of typing with a larger keyboard area. However, a large percentage of slide-out keyboards force landscape use. [5] All ways of holding a phone typically orient the device vertically, but for two-handed use, use of landscape mode was unexpectedly low. Though several of my clients have received numerous customer complaints in app store reviews for not supporting landscape mode.

What Do These Findings Mean?

“Some designers may interpret charts of one-handed use to mean that they should place low-priority or dangerous functions in the hard to reach area in the upper-left corner of the screen. But I wouldn’t recommend that.”

I expect some to argue that one-handed use is the ideal—and that assuming one-handed use is a safe bet when designing for almost half of all users. But I see more complexity.

Some designers may interpret charts of one-handed use to mean that they should place low-priority or dangerous functions in the hard to reach area in the upper-left corner of the screen. [6] But I wouldn’t recommend that. What if a user sees buttons at the top, so switches to cradling his phone to more easily reach all functionality on the screen—or just prefers holding it that way all the time?

Even if we don’t understand why there are such large percentages for handedness, we cannot assume that people will hold their phone in their right or left hand. When targeting browsers or mobile-device operating systems, I am always uncomfortable ignoring anything with a market share over 5%. That’s a general baseline for me, though I adjust it for individual clients or products. But I would never, ever ignore 20 to 30% of my user base. While I am personally very right handed, now that I have these numbers, I am spending a lot more time paying attention to how interactions might work when using the left hand.

Another factor that I had not adequately considered until putting together these diagrams is how much of the screen a finger may obscure when holding a mobile phone in any of these ways. With the display occupying so much of the device’s surface, this may explain part of the reason for a user’s shifting of his or her grasp. As designers, we should always be aware of what content a person’s fingers might obscure anywhere across the whole screen. Just remembering that a tapping finger or thumb hides a button’s label is not enough.

Now, my inclination to test my user interface designs on devices is stronger than ever. Whether I’ve created a working prototype, screen images, or just a paper prototype that I’ve printed at scale, I put it on a mobile device or an object with similar dimensions and hold it in all of the ways that users would be likely to hold it to ensure that my fingers don’t obscure essential content and that buttons users would need to reach aren’t difficult to reach.

Next Steps

“With clear correlations between tasks and ways of holding a phone, we could surmise likely ways of holding devices for particular types of interactions rather than making possibly false assumptions….”

I don’t consider this the ultimate study on how users hold mobile devices, and I would like to see someone do more work on it, even if I’m not the one to carry it out. It would be very helpful to get some solid figures on how much people switch the ways they’re holding their mobile phone—from one-handed use to cradling to two-handed use. Having accurate percentages for how many users prefer each way of holding a phone would be useful. Do all users hold their phones in all three of these ways at different times? This is not entirely clear. It would also be helpful to determine which ways of holding a mobile phone are appropriate for specific tasks. With clear correlations between tasks and ways of holding a phone, we could surmise likely ways of holding devices for particular types of interactions rather than making possibly false assumptions based on our own behavior and preferences.


[1] Sottek, T.C. “Transit Wireless to Expand Cell Coverage to 30 NYC Subway Stations by 2013 with Support from T-Mobile and AT&T.” The Verge, November 19, 2012. Retrieved January 18, 2013. “Customers use only 20 percent of voice traffic underground compared to an above-ground cell site, but 5 times more data.”

[2] Clark, Josh. “Designing For Touch.” .net Magazine, February 1, 2012. Retrieved January 18, 2013.

[3] Diaz, Jesus. “This Is Why the iPhone’s Screen Will Always Be 3.5 Inches.” Gizmodo, October 8, 2011. Retrieved January 18, 2013.

[4] Wikipedia. “Handedness.” Wikipedia. Retrieved January 19, 2013.

[5] Hoober, Steven. “Mobile Input Methods.” UXmatters, November 1, 2012. Retrieved January 18, 2013.

[6] Wroblewski, Luke. “Responsive Navigation: Optimizing for Touch Across Devices.” LukeW Ideation + Design, November 2, 2012. Retrieved January 18, 2013.


Hi Steven,

Very interesting and insightful article and project. I believe you didn’t take into account in your study cases like when people sit in a cafe or are at home. You examined only public and dynamic environments. I believe there is a lot to learn from these cases, too. However, I don’t know if these observations would be very different from the ones you mentioned. Just a thought.

Fantastic resource. Thanks for putting this together and sharing!

You left out of your study smartphones with a slide-out keyboard like the Droid and Samsung Stratosphere. I think you would find more often that not—in data entry mode—that the phone is almost always held in landscape mode—even when the app doesn’t support landscape like E*Trade’s and PayPal’s apps.

There is a segment of the population out there that still likes the visceral feel of buttons.

I think your odd findings on single-handed use correlate to the telemarketer effect, as I call it. This is where someone whose dominant hand is the right hand uses their left hand on the phone, so their right hand can write messages on paper. I and many others like me, who have answered phones for a living without headsets, use their phones with their left hand. I think this creates a skewing in the numbers for use with a single hand, making the left-hand numbers higher than the left-handed percentage of the general population—estimated at 8-15%. Your 33% is almost double those figures.

Thanks for sharing with everyone. It seems there may be a slow, steady migration toward doing more entertainment, work, and communicating in other ways than only talking. The ideal, hand-size phone that Apple came up with may be becoming obsolete. The activities are changing I suppose.

Ioannis Karlis: Yes. I hoped to make that clear, but there was no obvious way to observe people in offices or their homes. Not just for this study, but for any study. People become very self-conscious of what they are doing when they think about it, so things like a diary study won’t help here. (It was hard to get the reference photos for the drawings from coworkers; they suddenly could not hold the phone in their natural way!)

Based on the variations, I think at most you’d see some changes in percentages. But I think that might also correlate more closely to the types of actions they are performing on the device, instead of their environment. Anecdotally, around the tech offices where I work, people seem to hold their phones as the field study showed. I did observe a few people in cafes and so forth; these are many of the sitting individuals. I didn’t discuss this in this summary, but if you look at the raw data, there’s a little more information in there that you may enjoy trying to sort through.

Martin Middelmann: I completely agree with you about the prevalence of hardware keyboards. So much so that I’ve already written about it. Check out my column “Mobile Input Methods,” where I even try to quote some stats on QWERTY-keyboard phone usage rates, but that data is hard to find.

I also did sort of fuzz the data. The vast majority of devices were touch only or mostly touch, but I counted anything that was a mobile phone or mobile Internet device (MID). I am sure I got a few iPod Touches and game consoles, but I also certainly got a few 10-key feature phones, and some slide-out or fixed (BlackBerry style) QWERTY devices. Despite the stats I’ve provided, there were never enough of them to justify my adding another column to the spreadsheet for those cases when I could reliably tell what the device was.

Combining the slide-out QWERTY keyboard and the 2-handed landscape touch keyboard, there aren’t a lot of those people out there. Though it is my favorite layout still. I keep a Galaxy Epic in my bag as a test device, and loved it when it was a current phone.

Alvin Chao: I had math logic problems, so I think I decided that dominance flipping would equalize, but I guess not—now that I say it. The right-handers would flip to left-handedness more. That seems plausible, because when there was what I perceived to be a clear reason for switching, the right hand was often occupied with something specific for which a person could not change hands easily: opening a door where the approach is wrong for the other hand, using keys or keypads or an ATM, or holding hands with a girlfriend. It’s okay to text at the same time, but organizing a hand switch is too much. :)

Steve Grieshaber: This is exactly what I was trying to get to at the end, but which I just don’t have the data to support yet. I’d love to figure out a way to snoop on what people are doing with their devices, but even if it’s technically possible, there are privacy issues!

I’ve seen a few bits of data float by in the last month or two about more and deeper creation tasks being done on mobiles. And, of course, we’ve always had this huge fraction of human communications originate from mobile, even with 10-key pads. SMS is a lot of typing, so as compelling services continue to be built that make entry easy, foolproof, and useful, I expect to see more true interactivity and information origination on handsets.

Someone will figure out how to get us more data on this, sometime.

Super interesting article. Thank you. I also think that sometimes users read the interface design and position their hands to be efficient for that specific interface. There are so many kinds of interfaces these days that users start to have a usability culture—even non-tech-oriented ones. Reading an app is almost intuitive for them now. So, maybe some small percentage of your observations are a result of this and not only about comfort or habit, and therefore, one subject may prefer one-handed use, cradling, or two-handed use depending on the interface.

Great article, Steven. Thanks for publishing your findings. UX designers should always look to understand the users’ natural behavior when designing experiences. I’d be interested in knowing behaviors around phone handling and cradling in relation to pockets. Because mobile phones are quite personal, do we handle our devices in our pockets as something valuable and sacred with a protective grasp? Or perhaps people hold their phone in a manner that affords quick interaction with the physical buttons because they cannot see the screen.

This article made me reflect on my personal behaviors. I find that when putting my phone into my pocket, my natural behavior is to place my phone upside down, screen facing out. When I want to remove or even handle my phone in my pocket, my right index finger seems to naturally locate and lightly press the physical on/off button, while my thumb rests on the middle part of the left edge. Perhaps this gesture is based on my habit and need for knowing the current time. I often push the on/off button inside my pocket to illuminates the screen, then take the phone out to view the display.

Pedro Lança: I completely agree that not only do people seem to be selecting their method of holding and interacting based on the interface layout, but that a lot of them are conscious of it and have evolved their way of using the phone deliberately.

Eric Berkman’s comments on how he pockets and grabs his phone are, I suspect, very typical of everyday users who balance access and the security of their device, more than with anything else they routinely carry.

This can be seen in other fields as well—for example, when people reject phone apps that are stretched to fit a tablet. Everyone rejected the Facebook hybrid app—which was proven when everyone came back once they launched a native app. Users are picky, expect a lot from us, and are aware of what they are doing.

Physical buttons are going away fast and are so much an increasingly smaller subset as to not be relevant anymore.

Thanks for the article, Steven. Here’s something related we did last year.

Sketches of how people use mobile devices at airports and train stations.

Thanks for an interesting study.

There are two points you mention briefly that deserve more research:

  1. You state that an iPhone should be usable with one hand—and as the cited article mentions, it’s primarily because of screen size. But then you fail to take screen size into consideration in your study. Clearly, a 5-inch screen requires two hands for anyone other than King Kong. :-) So that skews your results, depending on how many phones were larger than, say, 4 inches.
  2. At the end, you talk about different tasks potentially requiring different ways of holding a phone. That would be very useful information to have. If a user is typing on the keyboard, two fingers are faster than one, and so a user would be much more likely to use two hands regardless of screen size. Alternatively, a user scrolling or selecting buttons, menus, or items in a list could do that just as quickly with one hand as two. So that would skew the results depending on the task at hand.

So while preliminarily useful, your data should not really be used to argue for users holding a particular phone in a particular way—or not doing so. For one thing, iOS and Android UX designers could potentially come to opposite conclusions based on phone size data. Also, different screens might be designed in different ways depending on whether a user is inputting text or doing some other input on that screen.

Two details that you may find interesting: First, when trying to reach the upper part of the screen while holding the phone in my right hand, the base of my thumb sometimes touches the lower part of the screen, leading to undesired results. Second, when waking up and lying in bed, my head is turned sideways. If I hold the phone sideways, however, it switches to horizontal mode, not allowing me to read. I have to hold the phone vertically, which, in that position, is unnatural to me.

Monitoring this on the commute fast becomes an obsession! I know an associate of mine, Orde Saunders, was a bit of a watcher, and I recall Steph Rieger giving it a tweet or two last year.

Another important piece to consider in the jigsaw. :)

Depending on your region and audience, this is not true. The vast majority of handsets in the world are feature phones, essentially all of which still have keys, even if some also are touch. If our designs need to address them—and I do projects where we indeed get good-to-overwhelming audiences on feature phones—we need to think about those or maybe even design for them specifically.

There is also some evidence that maybe half of Android devices that are used for Web browsing have a hardware keyboard. Stats are few and far between, but any that I can find that mention physical characteristics say this.

I can’t wait for the BB10 devices to be fully launched, so we can get solid numbers on the X10/Q10 split. Will be interesting to see, and it should be easier to get the data on how many choose the hardware keyboard, finally.

Neat. Totally tweeted this to the world.

I take photos sometimes, especially when I have a very long lens because people don’t like you taking photos. And I sort of wonder what I was taking notes about when I gathered the data in this study. Sketching takes a while—so did no one notice you were drawing them?

I agree more research is needed, and indeed have said so. But I would disagree with your first point that the results are skewed. They are absolutely true. But maybe not relatively true, if you care about a specific subset.

It was simply impossible to identify devices. A previous observational study that I did proved this to me. I recorded very few devices because I had to be sure what class they were. If I had thought it necessary to record device type with that level of confidence, I would have been able to record only a quarter to a tenth the number of observations.

Additionally, iPhone is not the only small device. I made a deliberate choice not to identify device class for this reason as well. It would have meant lumping iPhone together with the other small devices and implying that they were one size class, which is very much not the case. There are—and we observed—devices that are smaller than iPhone, the same size, a bit bigger, as well as much bigger.

And anecdotally, I saw a lot of cradling and two-handed use of iPhones—including plenty of the noted vertical keyboard use on iPhones. It would be interesting to see if grasping changes based on device size, but I get the very strong impression these would be subtle changes. There is not a magic size that makes everyone hold a device with just one hand.

To the second point, yes indeed it would be nice to find out what people are doing, but this seems almost impossible without doing something like recruiting for a study and installing snooping software. If someone has a better idea, let’s carry it out.

Also, people vary. A lot of people do not touch type with two hands on desktop or notebook computers, so I would not be surprised if these figures extend to mobile devices. It could be argued—or even proven through observation and math, if you believe some researchers—that gestural input such as swypes and so on could be at least as effective as two-finger, individual-key typing. This is another aspect to take into consideration. Not every goal or task can be assumed to result in a particular interaction.

You stated that: “a user scrolling or selecting buttons, menus, or items in a list could do that just as quickly with one hand as two.” But—again somewhat anecdotally—I’ve observed more than one user employing a two-handed grasp and using both thumbs alternately to interact with elements on the screen. The left thumb selected items and the right scrolled, for example. This is interesting behavior, and it would be enlightening to find out why this happens and how common it is becoming.

There are a few other issues I could take up—including your repeated use of the word skew. But in response to your saying, “For one thing, iOS and Android UX designers could potentially come to opposite conclusions based on phone size data,” I hope that there are increasingly few designers who design exclusively for iOS, Android, BlackBerry OS, S60, or Windows. We must recognize that the market is multifaceted, and every device needs to be addressed. Branch to implement OS-compliant details, sure. But a lot of decisions about basic architecture and layout can be common across platforms, so generalized and flexible solutions are good.

Thanks for a great article, Steven!

I especially appreciate that you shared your data file, so other people in the community can do analysis!

I have a question about your data. Were there 1330 different people? Or did the observations include the same person, but switching their way of holding a device? You mentioned that people did this frequently, but it didn’t look like that was captured in the data file.

I totally agree with Lou about how ways of holding a device should divide into different situations to research, which should relate to the size of the phone screen.

So, when we consider the user experience design, we should think about these situations, too.

What do you think about, for example, the Groupon Web site? Why is their call-to-action button in the upper-right area? It is not a responsive Web site. Does anyone have the answer?

Hi Steven,

Thanks for the article.

You might find some research I’ve done relevant. I designed and evaluated a set of interaction techniques that use auxiliary finger input in a one-handed mobile scenario. It’s very closely positioned to the theme of your article.

Video: Unifone: Designing for Auxiliary Finger Input in One-Handed Mobile Interactions

Paper (PDF): Unifone: Designing for Auxiliary Finger Input in One-Handed Mobile Interactions

Also, you might find this paper by Karlson, Bederson, and Contreras-Vidal relevant (PDF): “Understanding Single-Handed Mobile Device Interaction

Cheers, David

Great points!

You and Greg Nudelman (greg AT designcaffeine DOT com) really should talk. He has done a lot of work on this and just gave a presentation at World IA day in Los Angeles on this very subject. I’ll point him to this article as well.

Great observations!

I wonder how these findings could apply to using mobile devices whilst driving.

Yes, 1333 individual observations. I hope they are actually individuals, too, but it’s possible I may have encountered some of the same people at the bus stop on subsequent days, because I revisited the same ground.

But that’s not all you asked. I did not expect such speedy and frequent changes in grasping, and many people did not change, so I didn’t account for it in the research design. And I didn’t even notice it was happening until we’d already logged hundreds of observations, then couldn’t figure out a great way to note it with the observational capabilities I had. The best way by far to get lots of observations was to sit or walk during rush hour or at a mall. When making such observations, people were visible for just a few seconds, so we took all such observations from the initial few seconds that we observed a person.

When I was on a train with people, and they switched from one to two hands, I did not log the second interaction in the primary data. One variation, though: if a person was sitting there without a phone, then pulled it out and interacted with it, I counted the first observed interaction. Similarly, I purged boring interactions—that is passive use or phone calls—when logging touch interactions.

If any of that is suspect or does not make sense, everyone please tell me, and I’ll try to explain further.

Interesting design stuff. I do recall reading that paper in the distant past, around when I did my own research to which I referred in the article. While interesting, both the limits of keypad or keyboard location and the differences in problems with the finger or thumb obfuscating the user interface make this less applicable to full-screen, touch-screen interactions.

But yes, I probably should have made reference to a few of these to make this point, so thanks for including the information so everyone has the background.

I am generally aware of Greg, but have missed any discussion like this. If he posted a slideshow or video or blog entry about it—I can’t find one—please provide the link to us all here.

Thank you so much for sharing this info!

I’m an Android phone user with an iPod and iPad, and one thing that bothers me about iOS is that many buttons in apps are at the top. Like Back and Forward buttons. It’s never made sense to me from an ergonomic or a usability point of view. (You end up at the bottom of the screen usually, so why would I want to reach back up?)

Anyway, I wonder if the hand-position switching occurs with iOS users when they are trying to access those buttons.

Greg presented at World IA Day in Los Angeles (2/9/13) on this very subject.

I will be posting videos and decks of the day later this weekend and will definitely drop a link here as well.

Steven and all, Greg’s presentation “Designing Magic Mobile Moments” is available on YouTube.


I know what you mean! I live in a college town, and I think it’s fun to watch what people do with their phones.

It would be awesome if we could develop an app for a naturalistic study on how people are actually holding their phones. I imagine it would run noninvasively in the background and would collect orientation of the device, time—it would be neat to see when people are actually using the device—and touch data—where on the screen people are actually touching. It would be troublesome to tell whether the phone is being held with one or two hands, though.

Today’s phones are too big to be handled with only one hand. If they were smaller, like the iPhone, people would use only one hand. So phone size is the deciding factor.

Regarding left-handed versus right-handed use, it largely depends on the country people live in. The percentage of left-handed people is atypically larger in certain countries, like the US.

Thanks Steven, that clarifies things. I am taking a look at the data now myself. I will let you know if I find anything else interesting!

This is an incredibly interesting study!

Regarding handedness, I would venture to say that this could be an entire study in and of itself, based on some small-sample studies I have conducted on my own.

In the same fashion as skateboarding, snowboarding, or surfing, where an individual has better balance and greater comfort based on foot dominance (best foot forward), many individuals have an ear preference when using the telephone. Thus, you will often see individuals performing one-handed phone gestures with the same hand they would use to answer a call with their dominant ear (left hand: left ear).

This actually creates a new set of scenarios when coupled with your dominant-hand theory.

A user is right-handed and right-ear dominant, but has one hand pre-occupied—coffee, bag, book, umbrella—which hand is holding the phone?

If he is left-ear dominant, but right-handed, I would hypothesize he holds an umbrella in his right and the phone in his left.

Just some food for thought.

I’ve noticed that typing with an Android tablet, using the swipe (‘swype’) method, can be incredibly fast. Actually, faster than using a conventional full-size keyboard. In comparison to the arduous business of two- or one-finger typing with a touch keypad, it is worlds apart. In order to do it you need to hold the device with one hand and swipe with a finger of the other hand.

I think this will be a strong trend in the future. Having tried it, it is clearly much, much better than the alternatives, and I think may eventually replace keyboards on desktop computers and laptops as well.

Clarissa Stiles:

Yes, exactly my thoughts. It would be great to get this much data, but I cannot figure out how to correlate how the device is held with what is gathered digitally.

Some of this, like two handed use, could be surmised from time between taps on distant parts of the screen. But clearly a passive hand or the difference between a thumb and finger would be hard to determine.

But I’ll be happy to work with anyone who wants to try to build a test protocol or a bit of software for this.


That the iPhone is perfect with one hand appears to be a myth, and one I addressed just a bit, early in the article.

People’s hands are of widely different sizes. If, for the sake of argument, you assume the iPhone 1-4 is the perfect size for a 50th percentile male, the math is pretty easy to do. So these new 5-inch phones are just fine for about—I didn’t do the math in detail—20% of the population.

Anecdotally, while gathering this data, I saw no obvious preference for one-handed use among iPhone users and, likewise, no obvious preference for cradling or two-handed use among larger devices. I saw individuals entering information into Galaxy Notes with one hand and a thumb, and two-thumb typing on tiny 3-inch phones.

J.R. Hottle:

In visually intense applications—using night vision in aircraft, combat shooting—there is a phenomenon called cross-eye dominance where the user may be right handed, but their left eye is dominant. They have trouble with conventional setups and sometimes cannot fight it adequately.

It sounds like the same phenomenon is true of footedness and so on, based on what you describe, and this is very interesting to think of—though obvious if I had stopped to think about it. I agree this would be interesting to study, and coupled with rather fuzzy data on how many people are really right handed among a specific population, I’d love to see it.

Personally, I have a slightly worse hearing loss on the right side, but after 40 years, I have to very consciously move the handset to the left side if it’s a dodgy signal. Lots of weird, specific cases pop up aside from the immediate environmental concerns I had originally ascribed some of the change to.

Thank you for taking the time to do this. I know it’s only a simple study but it’s enough to make people wonder more and delve into real-life behavioral interactions. I agree that we could have improved the study by ticking off the activities to correlate with the specific behavior you observed, but it’s just a limitation at this point—that doesn’t mean you couldn’t conduct another study at a later date.

I do have a few inquiries on how you did the study. For example, I noticed that you created informal heatmaps to designate where touch interactions happen—that is, as shown by the illustrations. By what method were you able to estimate these distances? In conjunction with this, how long did you spend time observing one subject for his or her mobile phone behavior?

Rodge: Sorry I missed your comment. There’s no system to subscribe, so hopefully you’ll find this still.

Methodology first: There was no fixed time to observe, but anything long enough to clearly understand the device, context, grasp, and use was several seconds, so it was clear that it was not a momentary state, and I counted it. That meant a lot of the observations were people going the other way on a city street. As they passed, I ticked off boxes on a sheet. I can go into other specifics if it is important.

Ideally, the observer would have more time. I can imagine something like a video camera on a stick over a sidewalk, for example; then viewing the video in fast-forward and slowing and pausing to see more detail. I don’t have the facility, permission, or equipment and don’t do enough ethnogragraphy to know whether this is even legal or ethical.

The heatmaps are based on Josh Clark’s original ideals, which I referred to above. I used a pretty coarse measure of reach based on a small and informal sampling of coworkers. (I have weird thumbs, so it’s never me.) The drawings are based on actual photos of coworkers holding handsets, then I made the overlays of the sweep angle from that. These are coarse and vague because they are guidelines. Human hands and device sizes and configurations vary too much for anything more specific.

I’d love to know what people were doing on the handset. I still have no good way to come up with this observationally. Probably the best I can imagine is a large, but controlled cohort, like a set of students who could be observed by stalking them or by using embedded equipment or video surveillance. The on-device behavior could be monitored by an installed app. These quickly get out of hand and are too complex and/or too expensive to consider. But, I’ll work with anyone on design or data gathering for any other experiments.

Hi Steven,

Firstly, thank you very for collecting this user data and writing the article. It’s great to see some states like this.

In your opinion, for a mobile gaming app that is in fixed landscape mode, should the user be able to play it using only one hand? (A basic card game, for example.)

I’m having a discussion at work about this and feel that, for long periods (1-2 hrs) of use, the user could get lazy and start playing the game one handed if possible (in landscape).

My thoughts: People sometimes get lazy when doing things over long periods of time and, if they can do something one handed, they will—for example, driving with one hand on the wheel.

I would be really interested to get your thoughts on this.

Many thanks, Steve

Interesting question. I’ll preface my response by saying that I haven’t ever done a game that I recall. I do lots of stuff, so that’s weird, but there it is.

On the other hand, last I checked, gamers were still mostly human, so most attributes apply generally. I think you are on the right track. Let’s say that we aren’t allowed to talk about mobile context (distraction, etc.) because that’s not cool anymore. But when’s the last time you read a book, magazine, or iPad article without moving. Of course, people shift and use one hand and so on.

And, this seems easy to test. Now that school’s out, use the kids that get brought to the office. Assuming your game is age appropriate, have a couple of them play it for as long as they can stand, while recording them. When you’ve finished, fast forward and record how they held, for how long, etc. Proof!

Then tell us all about it.

Dood? Why do you make things so hard?

Human beings can be balanced or unbalanced. Unbalanced human beings can be either unbalanced toward the Yin part of the body or unbalanced toward the Yang part of the body.

What your study reveals is the percentage of the populace that is very unbalanced, sort of unbalanced, and balanced. The way the people hold the phone is only incidental to the overall balance of their body.

One-handers are unbalanced people with a majority of Yang or Yin as stated above.

People who cradle a phone have some kind of development in the less dominant hand that allows them to use that hand as a cradle.

People who use both hands are relatively balanced between Yin and Yang. Their percentage is so low because the percentage of balanced human beings in the West is low.

Once the West decides to get around to teaching Yin Yang Theory, the percentages you found will change. Although who knows when that will be, since the West “opened up China” in 1970 and 40 years later Yin Yang Theory still hasn’t made it to the mainstream in the West.

The percentage of one-handers will go down as people become more Yin and Yang balanced, and the percentage of people using both hands will go up.

Did you ever think of doing your study in a foreign country? Or maybe on each continent to see how the people on a continent generally react?

Have you considered the implications of your findings if they do vary from country to country?

I am the inventor of HandAble, a hand-held cell phone holder. The information developed in your report is right on from our perspective.

Hi Steven

Thanks for sharing these interesting observations with us.

Like your previous correspondent, Aaron, I too have invented a device that I am hoping will revolutionise the manner in which hand-held devices are manipulated. (More detail to follow at a later stage.)

Would you have any objections to me quoting any of your findings? And if not, would you mind giving me your credentials in order for them to assess the credibility of your data?

Regards from Johannesburg, in sunny South Africa.


Thanks for such an in-depth analysis! This awareness of the user’s interactions with smartphones is very important as we all are constantly interacting with them. This is also the heart of what we have been doing with a useful new product that we market.

I’d like to let you know of our Mobigrip grip assist device. This simple disc mounts to the center back of your phone and, with a built-in silicone loop, it flexibly connects one finger to your device. All movements are secured by the loop that moves and stretches as the device is manipulated by the user. Mobigrip enhances the way a user holds or can hold their device, because the device is now holding you, and cuts down substantially on accidental drops and slips. It really is a comfortable and cost-saving enhancement to any phone. We would be happy to send samples for review if you wish.

The response we get is verging on 100% universal appreciation for the natural feel, cost savings, and simplicity to the global mobile marketplace. Thanks again!


Sorry that I haven’t been paying attention to comments. Catching up:

Jalon, I did some of the survey in Canada, and if you look at the spreadsheet, it does list locale and context. I don’t travel enough for overseas surveying, but anyone is free to take the form and add to it. I’d love to see if it varies notably by region or culture.

Terry, this info gets quoted a lot, and I know people at big companies like Disney who have used it to justify improvements and changes to their bosses. Use it if it can make your work better.

Thank you for a great article. Are we allowed to use the images in our blogs if we share your learnings and link to this article?

Great resource! Thank you so much!

Join the Discussion

Asterisks (*) indicate required information.