How to Label a Pie Chart Properly

Labels are where pie charts are won and lost. The geometry can be perfect and the colors tasteful, but if a reader has to guess which number belongs to which slice, the chart has failed at its one job. This guide covers the label system in our editor and the judgment calls it cannot make for you.

What should a label say?

Three pieces of information compete for each slice: the category name, the raw value, and the percentage. All three toggle independently, and the right combination depends on where the chart will live.

A chart inside a report, sitting next to explanatory text and a legend, can carry percentages alone. A chart that must stand on its own, on a slide, in a social post, printed on a handout, should carry names on the slices, because legends make readers do a color-matching exercise that direct labels eliminate.

Add raw values when the absolute numbers matter: budgets, donations, headcounts. “Programs · 78%” persuades; “Programs · $312,000 · 78%” informs. Skip values when they would just be noise, like survey counts nobody will act on.

Inside, outside, and the auto mode

An inside label is the cleanest look, but it needs a slice wide enough to hold it. As slices thin out, inside labels start overflowing their wedges, and tools that force them anyway produce the classic unreadable pie.

Outside labels solve the space problem and create a new one: distance. A label floating near the rim could belong to either of two neighboring slices. Leader lines fix that, drawing a short connector from slice to label so ownership is explicit. Turn them on whenever labels sit outside; there is a toggle in the label settings, and it defaults to on for a reason.

The auto placement mode makes this decision per slice: labels sit inside where they fit and move outside with leaders where they do not. It is the right choice for most charts because real datasets mix fat and thin slices. The force-inside and force-outside modes exist for when you want visual consistency and know your slices can afford it. There is also a none option, for charts that lean entirely on a legend or on annotation added later in a design tool.

Fonts, sizes, and contrast

Label typography has its own controls: font family, size, and color. Two practical rules cover most cases.

First, size for the final viewing distance, not for your screen. A chart pasted into a document gets read at arm’s length; 12 to 14px labels are fine. The same chart projected in a meeting room needs 16 to 20px or more. It should look slightly oversized in the editor preview.

Second, check contrast against the slices, not the background. Inside labels sit on slice colors, so a dark gray label that is perfect on pale pastels disappears on a navy slice. If your palette mixes light and dark slices, outside labels sidestep the problem entirely, or pick a label color that survives your darkest slice.

The legend question

Legends and slice labels overlap in purpose, and running both at full strength is redundant. A clean way to decide:

• Slices show names: hide the legend.
• Slices show only percentages or values: keep the legend, positioned wherever your layout has room. Top, bottom, left, and right are all available.
• Slices too small and numerous to label: legend only, and consider whether the data wanted a bar chart.

When labels will not fit, the data is the problem

Every labeling technique has a breaking point, and it arrives around the seventh or eighth slice. If the auto placement is scattering leader lines everywhere and the chart looks like a maypole, no font setting will save it. Group the small slices into a single Other slice with the one-click grouping tool, and suddenly five labels fit beautifully where eleven fought for space.

Label the Other slice honestly, and if the breakdown matters, list its contents in a caption. Readers forgive aggregation; they do not forgive guesswork.