Why fingers miss
A finger is not a precise pointer. It covers the thing it is aiming at, and the fingertip contacts the glass over an area, not a point; the touchscreen then works out a single position from that area. Add to this a visitor who is holding a bag, wearing gloves, standing at an awkward angle, moving slightly or has a tremor, and the chance of landing a little off target rises.
A well-designed touch target allows for this. The aim is that a visitor who is reasonably close to the right place gets the right result, and that a near miss does not trigger something else.
Size
Bigger targets are easier to hit. Published design guidance for handheld devices commonly suggests minimum target sizes on the order of several millimeters per side, often cited as roughly 9 to 10 millimeters (a little over three-eighths of an inch) for phones, and some accessibility guidance sets a minimum in pixels that depends on screen density. Use the figures in the guideline you follow rather than a number from a general article, and treat any minimum as a floor, not a goal.
Public exhibit screens usually deserve more generous sizes than a phone, for several reasons:
- Visitors stand at arm’s length and may be at an angle.
- Many are first-time users with no practice.
- The audience includes children, older adults and people with limited dexterity or tremor, and visitors wearing gloves in cold seasons.
- Visitors may be moving, jostled by crowds or distracted.
As a practical approach, make important buttons as large as the layout reasonably allows, and make the most common or most consequential actions the largest. A primary action such as “Start” or “Next” can be markedly bigger than secondary ones.
Always specify size in physical units (millimeters or inches) rather than only in pixels, because the same pixel size looks very different on a phone and a big wall display. When you design on a monitor, check a physical prototype at the real display size.
Spacing
Size alone does not help if the buttons are crowded. Gaps between targets reduce the chance that a slightly misplaced touch lands on a neighbor. A few points:
- Leave clear space between buttons. The gap should be large enough that a fingertip aimed at one button is unlikely to touch another. A larger gap is safer when the consequences of a wrong tap are significant.
- The touchable area can be larger than the visible shape. A small icon can sit inside a larger invisible hit area. This can make controls easier to hit without enlarging the graphics. Be careful that hit areas do not overlap.
- Group related controls, separate unrelated ones. Spacing can also express structure, and distant controls are less likely to be confused.
- Keep destructive actions apart. Do not place “Delete” or “Reset” next to “Save” or “Next”. Put them far away or ask for confirmation.
Edges and corners
Touch accuracy often drops near the edges of a screen, where bezels, frames and enclosures can interfere with reach and sensing. Targets pushed into the corner are harder to hit, and may be blocked by a mounting frame. Keep important controls a comfortable distance from the edge, and check how the enclosure sits against the glass.
Position and reach
Where the control sits matters as much as its size. A large button at the top of a tall screen is no use to a seated visitor or a child who cannot reach it. Keep frequent controls in a band that is reachable from both standing and seated positions, consulting reach ranges in the relevant guidelines. Large vertical screens in particular may need controls clustered in a lower region.
Feedback
Because visitors cannot always tell whether a touch registered, show a response immediately:
- A visible change, such as a color shift or press effect, on contact.
- A sound, where appropriate and not disruptive.
- A clear result, such as the next screen appearing quickly.
Without feedback, people tap again, which can cause double actions. Consider ignoring repeated taps within a short window and avoid requiring a long press or a double tap. Those gestures are hard for many people.
Also consider acting on release rather than on first contact. This lets a visitor who touches the wrong place slide away without triggering it, which forgives errors.
Gestures
Small targets and tricky gestures overlap. Swipes, drags and pinches demand precise movement. Where possible, pair a gesture with a simple alternative, such as a button for next and previous, in addition to a swipe. Sliders need especially large handles and a way to tap a position. A control that can only be operated by dragging will exclude some visitors.
Hardware matters
Even a well-designed layout fails on a badly calibrated screen. If the touch point is offset from where the finger lands, targets are effectively shifted. Parallax from thick glass or an angled screen makes this worse, since the visitor sees the button slightly displaced from the sensing surface. Regular touch calibration and a choice of display with the glass close to the image reduce these errors. Dirty or scratched screens, wear in busy spots and shared touch hygiene routines also affect reliability, and so does glare, which can make it hard to see what you are touching.
Test with real hands
The best check is to watch people use the screen. Prototype testing with children, older adults, people with limited dexterity and visitors in winter clothing will show where targets are too small or too close. Look for repeated attempts, accidental touches and hesitation. Visitor observation on opening days, and again after months of use, will show whether the design holds up. Adjust sizes based on what you see rather than on a calculation alone.
A quick checklist
- Are the main controls as large as the layout allows, and sized in physical units?
- Is there generous space between buttons, especially for risky actions?
- Are controls away from the edges and within reach for seated visitors and children?
- Does every touch produce immediate feedback?
- Is there a simple alternative to any swipe or drag?
- Is the screen calibrated, clean and free of glare?
- Has the design been tried with the full range of visitors?