What it is
A 100% stacked chart divides each column into the share each part contributes to that column’s total. Every column has the same height, so differences in the totals disappear and differences in the composition stand out.
✓ Use it when
- You care about proportions within each group: device used by age group, answer given by country, energy mix by year.
- Your data is already in percentages that add up to 100 per row — Chart Buffet notices and suggests this chart.
- The groups have very different totals, but you want to compare them fairly.
✗ Avoid it when
- The totals themselves are important — a 100% chart hides them. Use a regular stacked chart, or show the totals separately.
- There is only one group — a doughnut or a single bar does the job.
Tips for a better chart
- Order the segments meaningfully: for agree–disagree scales, keep the scale’s order and use colours that read as opposites.
- Show the percentage inside each segment when it is wide enough, and let the tooltip carry the rest.
- Mention the totals in the subtitle or source line if readers will wonder about them.
The maths
Each segment is the part’s share of its column: pij = 100 · vij / Σm vim. The shares in a column always sum to 100, which is why the totals vanish.
Because the first segment is anchored at 0% and the last at 100%, both the bottom and the top segments can be compared accurately across columns; only the middle ones float.
A little history
Percentage bars were a favourite of early-twentieth-century statistical handbooks, which recommended them for showing how a whole divides (Brinton 1914). Management consultants made them a fixture of strategy presentations, where they often replace a row of pie charts (Zelazny 1985) — a substitution perception research supports, since aligned lengths are easier to compare than angles (Cleveland & McGill 1984).
Questions
My values don’t add up to 100. Can I still use it?
Yes. Chart Buffet converts each column to shares of its own total automatically.
Is a 100% stacked chart better than several pie charts?
Usually. The segments line up against common edges, so comparing groups is far easier than comparing angles across separate pies.
References
- Brinton, W. C. (1914). Graphic Methods for Presenting Facts. New York: The Engineering Magazine Company.
- Cleveland, W. S. & McGill, R. (1984). Graphical perception: theory, experimentation, and application to the development of graphical methods. Journal of the American Statistical Association 79(387), 531–554.
- Zelazny, G. (1985). Say It with Charts: The Executive’s Guide to Successful Presentations. Homewood, IL: Dow Jones-Irwin.