Showing how one total became another.

The form for a question no other chart answers: not what changed, but what the change was made of.

Drawn as a waterfall chart, bridge chart, margin bridge, depending on the data.

Two large changes, cancelling out

Net −$4,000 across changes totalling $974,000

Rendered from the project Budget v1 vs v4, through the same compiler the product uses. Not a picture of one.

What it is for

A waterfall starts at one total, applies signed contributions in sequence, and lands on another. It is the right form whenever a number moved and the interesting part is the decomposition: revenue from January to July, gross margin down to contribution, a budget from one version to the next.

Its value is that it makes offsetting movements visible. A total that barely moved can conceal a large increase and a large decrease, and every other chart form reports that as nothing happening. In the worked comparison example, two large changes cancel almost exactly and the headline total moves by less than a percent.

Reach for it when

A total changed and the question is what drove it
The components genuinely add up to the difference with nothing left over
Offsetting movements are suspected inside a total that looks flat
You are bridging one definition to another, such as gross margin to contribution

Use something else when

The components do not sum to the difference, which makes the chart a false accounting
There are many small steps, where the bridge becomes unreadable and a sorted bar chart of contributions is better
The change is over time in regular intervals, where a trend chart reads more directly

How this form misleads

Steps that do not reconcile

A waterfall asserts that the start plus the steps equals the end. A residual quietly absorbed into the last bar makes the chart wrong in the one way it promises not to be.

Show any unexplained residual as its own labelled step. An honest residual is information; a hidden one invalidates the whole chart.

Double counting across steps

Price and mix effects overlap: a shift toward expensive products shows up in both unless the decomposition is done carefully, and the bridge then over-explains the change.

Decompose in a stated order, hold everything else fixed at each step, and write the order down since a different order gives different sizes to the same steps.

Ordering the steps to make an argument

The order determines how much each factor appears to contribute when factors interact. Choosing the order after seeing the result is a way of choosing the conclusion.

Fix the order by a rule that is stated on the chart, such as volume before price before mix, and keep it across periods.

What this will not draw, whatever you ask for

The agent picks an intent from a closed list and a single compiler renders it. These forms are rejected at the schema layer rather than discouraged in a style guide, which is the difference between a rule and a preference.

3D anything: Perspective makes two equal values look different. There is no case where the third dimension carries data.
Gauges and speedometers: An enormous amount of ink for one number, and the dial implies a range that is usually invented.
Radar charts: Area scales with the square of the values and the shape changes entirely if you reorder the axes.
Word clouds: Size encodes frequency, which is the least interesting thing about a word, and nothing is comparable.
Pie charts above five slices: Angles are hard to compare, and beyond five slices the labels leave the chart anyway.
Per-series arbitrary colour: Colour means something specific in this product. A series that picks its own hue breaks the meaning everywhere else.

Make one from your own data

You do not choose the form. The analysis works out which question your data can answer and picks the chart that answers it, then tells you why.

No account needed to start. You only pay when you like what you see.