Stacked bar vs grouped bar: which question are you answering?
These two charts hold the same numbers and answer different questions, which is why teams argue about them without realising they want different things.
Stacked bars give you the total for free
Because segments sit on top of each other, the full bar height is the total. That makes stacking right when the total is the headline and the composition is context — revenue by quarter split by product, for example.
But only the bottom segment is easy to read
Every segment above the first starts at a different baseline, so comparing the third segment across bars is genuinely hard. If that comparison is the point, stacking is the wrong chart.
Grouped bars compare cleanly and lose the total
Side by side on a shared baseline, every series is comparable. The cost is that the total no longer exists visually — the reader has to add.
Five segments is the practical limit
Beyond five, stacked bars turn into colour soup and grouped bars turn into a picket fence. That is the point to switch to small multiples: one small chart per series, same axis.
- Total matters most → stacked.
- Series comparison matters most → grouped.
- Both matter → stacked for the total plus a small chart for the comparison.
- More than five series → small multiples.
One rule that saves time: whatever you put at the bottom of a stack is the series people will actually compare, so put the important one there.
Frequently asked
When should I use a stacked bar chart?
When the total is the main message and the breakdown is supporting detail, with roughly five segments or fewer.
Why are stacked bars hard to compare?
Only the bottom segment shares a common baseline. Segments above it start at varying heights, so the eye cannot compare their lengths reliably.