Chapter 5 — Practice-on-your-own solutions

The screen against the Big Idea. Every visual gets held against the Big Idea. A visual earns its place if it either states the Big Idea, decomposes it, or contextualises it in a way the audience needs to act. Everything else is decoration — and decoration on a decision page is not neutral; it costs the eye budget that should have landed on the visuals that carry the argument.

Visuals that earn their place (two).

1. A plan-variance bar chart of all regions, sorted by variance ascending. NA lands near zero (on plan); APAC lands at -12% accented in a warning colour; the other regions render in muted gray. This chart states the Big Idea in a single visual: two data points and two accents.
2. An APAC drill-down — either a small-multiple line per APAC product family showing where the miss originated, or a plan-variance decomposition of APAC by sub-region. This chart earns its place because the audience’s next question after seeing the miss is «where did it come from?» The drill-down answers it inside the same reading frame.

Visuals that do not earn their place (twenty).

• Twelve KPI cards. Twelve numbers with no comparison shape are twelve numbers Copilot picked because it could, not because the audience needs them. A KPI card without a comparison is Priya’s $42.7M card from Chapter 1: a number that asks the audience to supply meaning.
• Four line charts. None of them tracks the plan-vs-actual comparison the Big Idea is about. Trend framing dilutes the argument.
• Three bar charts (other than the region plan-variance). Any bar chart that is not the plan-variance chart is competing with the plan-variance chart for attention.
• The map. Maps are geography-framing; the Big Idea is a variance-framing. Different question shape, different chart family.
• Two tables. Tables serve reference lookup, not argument. If the audience is meant to look up a specific value, the tables can live on a second page; on the summary page they compete with the plan-variance chart.

The one-sentence Refine direction. «Delete all visuals except the region plan-variance bar chart (sorted ascending, APAC accented) and an APAC drill-down by product family; move the twelve KPI cards, four line charts, other bar charts, the map, and the tables to a second reference page or remove entirely.»

Why the twenty-visual page failed on its own terms. The audience arrived to answer one question («how did the plan land?») and left with twenty candidate answers, none of which reads as the argument. The page said everything and argued nothing. Chapter 5 walks the discipline in What a chart earns its place through.

The colour rewrite. «Colour Product X in the brand accent (the strongest single colour in the palette); render all four other bars in a single medium gray. Do not use any additional palette colours anywhere on the chart.»

Why the current five-colour version fails. Colour is a pre-attentive attribute — the human visual system detects it before the reader has begun to consciously parse the chart. Pre-attentive attributes work by contrast against a neutral baseline: one salient element against a uniform background pops. When every bar is a different colour, every bar is trying to be the emphasised bar, and the pre-attentive channel saturates. The reader’s eye lands on whichever colour their visual system happens to detect fastest (often warm reds and oranges regardless of position), not on the bar the Big Idea named.

The specific failure mode. The chart is using colour as a distinguishing attribute («each bar gets its own colour so the reader can tell them apart») when the design task called for colour as an emphasising attribute («one bar carries the argument; the others provide context»). Distinguishing is the job of the axis label — the reader already knows which bar is Product X because it’s labelled. Colour is a scarce channel; spend it on the emphasis, not on the distinguishing.

The colleague’s reflex to defend. The five-colour version usually gets defended as «on-brand.» Brand palettes are decoration standards, not visual-encoding standards; a chart that uses the whole palette because «we have to be on-brand» is treating the palette as required output when it’s a curated set of options. Using one brand colour with intent beats using five brand colours by default. Named in the Chapter 5 discipline as the one-thing-loud rule.