Choosing the Right Chart for a Mixed Audience
A logistics company must present quarterly performance to two audiences in one meeting: regional managers who need to compare 14 depots on on-time delivery, and the executive team who want the trend over eight quarters and the single worst-performing depot. The analyst has prepared a spreadsheet with on-time percentage per depot per quarter, plus delivery volume per depot. A colleague proposes a 3D pie chart of all 14 depots for the latest quarter, with a line chart of the eight-quarter national average placed behind it. The room is small, the projector is dim, and several attendees will view the slides on phones afterward.
1. Evaluate the colleague's proposal and explain what each chart type fails to communicate.
Read the worked answer
A 3D pie with 14 slices fails on two counts: humans cannot reliably compare 14 similar angles, and 3D perspective distorts slice sizes so the visual area no longer matches the value. It also cannot show volume, so a depot with 2 deliveries at 50% looks the same size as one with 2,000 at 98%. Placing a line chart behind it compounds the problem - overlapping marks in a dim room, and on a phone the detail becomes unreadable. The pie answers 'share of what?' when the real question is 'which depots are underperforming and is it getting better?'
2. Propose a replacement visual set and justify the encoding choices.
Read the worked answer
Use a sorted horizontal bar chart of on-time percentage by depot for the latest quarter, sorted ascending so the worst depot is at the top - position is the most accurate encoding for comparison, and horizontal bars accommodate 14 labels. Add volume as a secondary annotation or a paired bar so small-sample depots are visibly flagged. For the trend, a single line of the national average over eight quarters answers the executive question directly. If depot-level trend matters, a small-multiples grid beats 14 lines on one axis. Keep it 2D, high-contrast, and legible at phone size.
3. How should the analyst handle the small-volume depots in the ranking?
Read the worked answer
A depot with very few deliveries will swing wildly and can top or bottom the ranking by chance. The analyst should show the volume alongside the percentage, or use a confidence interval / minimum-sample threshold and grey out depots below it. Presenting a 50%-on-2-deliveries depot as 'worst' would mislead the meeting and damage trust in the whole deck.