An executive dashboard that 'feels misleading'
You inherit a monthly executive dashboard built by a departing analyst. Slide 1 is a pie chart of revenue across 11 product categories, the smallest of which is 1.2% of revenue. Slide 2 is a line chart titled 'explosive growth in active users'; the y-axis runs from 48,000 to 51,000 with no zero and no annotation. Slide 3 is a 3D-shaded bar chart of regional sales whose title reads 'Sales by Region' with no units, no date range, and an unlabelled colour legend. The CFO says the pack 'feels misleading' but cannot say why, and asks you to fix it before the board meeting in two days.
1. Diagnose each slide. What specifically is wrong, and what is the consequence for the reader?
Read the worked answer
- Slide 1 (pie, 11 categories). A pie encodes value as angle, which people judge poorly; with 11 slices, most are within a few degrees of each other and the 1.2% slice is invisible. The reader cannot rank categories or compare close ones. The same data as a sorted horizontal bar chart is instantly readable and adds no ink.
- Slide 2 (truncated axis). Starting at 48,000 magnifies a 6% move into a near-vertical climb. That may be legitimate if the point is short-term variance, but the title 'explosive growth' asserts a magnitude the axis hides. A reader who does not check the axis will overestimate the change by an order of magnitude.
- Slide 3 (3D, unlabelled). 3D shading distorts bar length through perspective, so the ranking of regions may simply be wrong. Missing units, date range and legend mean the reader cannot tell whether the numbers are currency or units, which period, or what the colours encode. A chart that cannot be interpreted without its author is not a chart.
In all three cases the fault is that the visual choice was made for decoration rather than for the comparison the reader needs to make.
2. Redesign slide 2. The underlying fact is that active users moved from 48,000 to 51,000 over twelve months. What should the board actually see?
Read the worked answer
First decide what the reader must judge: the level (are we at 51k?) or the rate of change (are we growing faster or slower than before?). They need different charts.
Recommended:
- Title the chart with the finding, not an adjective: 'Active users: 48,000 to 51,000 (+6.2% over 12 months)'.
- For level and trajectory, use a line chart from zero (or a clearly labelled break with an explicit note) so the reader sees that 51k is a modest move on a 48k base.
- If the real story is the recent pace, plot monthly net additions as a column chart instead - that shows acceleration or plateau honestly, without exaggerating the cumulative total.
- Annotate the axis break if one is kept: 'axis starts at 48,000'.
If the analyst's intent was to flag a genuine recent acceleration, prove it with the additions chart rather than a stretched axis. 'Explosive' is a conclusion the reader should be able to reach from the chart, not one the chart has to force.
3. Beyond fixing the three slides, what would you add to the pack so the board can trust it?
Read the worked answer
- Consistent, declared scales. Fix the axis range across periods and across the region chart so month-to-month comparisons are valid and a scale change is never silent.
- Direct labels instead of legends. Label bars and lines in place; sorted horizontal bars for categories.
- Units, period and source on every slide - a footnote with the extract date, currency, and whether figures are gross or net.
- Accessibility. Avoid red/green as the only distinction, keep contrast high, and add alt text for the published version. 3D and shadow effects go entirely.
- A 'what changed since last month' note on the front page, so a reader knows which movement matters.
- Definitions. State how 'active user' and 'revenue' are defined; a board metric that shifts definition between months is worse than no metric.
With two days, prioritise slides 2 and 3 - a truncated axis and an uninterpretable chart are trust-destroying, while the pie is merely inefficient. Fix, label, and add one line saying what changed and why.