Where a thumb can actually go · Design, Before Any Software

Phones outgrew hands

A phone with a 4-inch screen could be operated entirely with one thumb. A phone with a 6.7-inch screen cannot: the top corner opposite the holding hand is genuinely out of reach without regripping, and the whole top third requires a stretch that makes the phone unstable.

The reach map, for a right-handed one-thumb grip:

Easy: the bottom half, and the lower-right region.
Stretch: the middle, and the upper-right.
Hard or impossible: the top-left corner and the top edge.

Mirror it for a left hand. Which means the top-left corner — historically where the back button, the menu and the logo all live — is the hardest point on the screen for roughly nine people in ten.

What the industry did about it

This explains a decade of interface changes that otherwise look like fashion:

Bottom navigation bars replacing top tab bars in most mobile apps.
Safari moving its address bar to the bottom of the screen on iPhone.
Reachability on iOS and one-handed mode on Android, which slide the whole screen down on a gesture.
Bottom sheets replacing dialogues that used to appear centred or at the top.
Primary actions pinned to the bottom of forms and flows rather than sitting at the end of the content.

Each of those is the same finding applied.

The numbers, honestly

The widely cited figure is that about half of phone use is one-handed with the thumb. It comes from an observational study published in 2013 that watched 1,333 people using phones in public. That work is genuinely useful and it is also twelve years old and predates phones of the current size.

What has held up in later work is the more important finding: people switch grips constantly, within a single session — one-handed thumb, one-handed with the other hand steadying, two-handed with both thumbs, phone on a table. Nobody uses one grip.

So the design conclusion is not optimise for the thumb zone. It is:

Put frequent actions within easy reach, because that is where they are cheapest.
Do not put anything essential where it cannot be reached, because some share of uses will be one-handed.
Do not assume a grip, which means avoiding gestures that only work from one side and layouts that only make sense held one particular way.

Consequence, not recessive

The reach map interacts with the destructive-action rule from the touch-target lesson. Easy to reach is also easy to hit by accident, especially while walking.

So the allocation is two-dimensional:

frequent + safe        → bottom centre, large
frequent + destructive → bottom, but separated and confirmed
rare + safe            → top, or behind a menu
rare + destructive     → top or in a submenu, and confirmed

A delete button in the easiest-to-hit position on the screen is a design that will generate support tickets, however well it scores on reachability.

Gestures are not targets

One useful distinction. A pull-to-refresh gesture starts at the top of the screen and is perfectly comfortable one-handed, because a swipe can begin anywhere the thumb can touch and does not require precision. Reach constraints apply to targets, which require accuracy, far more than to gestures, which do not.

This is why the top of the screen is a reasonable place for a scrollable content area and a poor place for a small button.

The caveat is that gestures are invisible. A swipe-to-delete that nobody discovers is not a feature, and every gesture needs a discoverable equivalent somewhere — which is usually a button, which then needs a reachable position, which brings you back to the map.

The layout that follows

Put together, the reach map and the grip variation produce a fairly specific shape for a phone screen:

Bottom: primary navigation, the main action of whatever screen you are on, and anything used repeatedly.
Middle: content, which is scrolled rather than tapped, and where any gesture works comfortably.
Top: the title, contextual information, and at most one control — which should have an equivalent lower down if it matters.

This is close to an inversion of the desktop convention, where navigation sits along the top because a mouse reaches everywhere equally and the top is where the eye starts. Carrying the desktop arrangement onto a phone is the commonest cause of an app that feels awkward without anybody being able to say why.

One caveat about the bottom edge itself. The lowest strip of the screen belongs to the operating system on both platforms — the home indicator on iOS, the gesture bar on Android — so a control flush against the bottom edge competes with a system swipe. The safe-area padding from earlier in this block is what keeps them apart.

Testing it

The check costs nothing and is skipped almost universally: use your own design on a phone, standing up, one-handed, while holding something in the other hand.

Sitting at a desk with the phone in both hands is not a test of reach, and neither is a simulator. Ten minutes of genuinely one-handed use will tell you more about your interface's bottom half than any amount of annotating a heat map.