A dashboard is not a collection of charts. It is an argument about what matters, laid out so someone can take it in without being walked through it. That constraint decides almost every choice you will make.
Start with the question, not the data
Write the question at the top of the sheet in plain words before you build anything: "Are we selling more this quarter than last, and where is the change coming from?" Everything that does not help answer it comes off the page. This one habit removes most of the clutter that makes dashboards ignored.
A headline number that answers the question, a breakdown that explains it, and the detail table underneath for anyone who wants to check. If a reader has to scroll to find the answer, the layers are the wrong way round.
Separate the three sheets
Keep raw data, calculations and the display on different sheets. It makes the file easier to refresh, far easier to audit, and means you can rebuild the front page without touching anything that computes.
- Data — the imported table, untouched, converted to a named Table.
- Working — pivots, lookups and helper columns. It can be ugly; nobody sees it.
- Dashboard — nothing but references, charts and labels.
Choose the chart by the comparison
| You are showing | Use |
|---|---|
| Change over time | A line chart. Time runs left to right, always. |
| Ranking of categories | A horizontal bar chart, sorted longest first. Labels stay readable. |
| Share of a whole | A stacked bar, or a single number with a percentage. A pie chart works only with two or three slices. |
| One number against a target | The number itself, large, with the target as a small line of context beneath it. |
Avoid the second axis. A chart with two scales lets a reader draw whatever conclusion they arrived with, because the crossing point is set by your axis choice rather than by the data.
Make it readable at a glance
- One accent colour for the thing you want noticed, grey for everything else.
- Round the numbers. Nobody needs revenue to the paisa on a summary tile.
- Label the units once, in the title, rather than on every axis tick.
- Delete gridlines, chart borders and legends you can replace with a direct label.
- Date-stamp the sheet so a screenshot passed around next month is not mistaken for current.
Make it refresh in one click
If updating the dashboard is a fifteen-step ritual, it will go stale. Point every pivot at a named Table, put all filters into slicers rather than manual filters, and check that Refresh All produces the finished sheet with no manual step after it.
Key takeaways
- Write the question on the sheet, then delete everything that does not answer it.
- Keep data, working and display on separate sheets.
- Pick the chart from the comparison you are making, and skip the second axis.
- If Refresh All does not finish the job, the dashboard will go stale.
Turn this into a career, not a bookmark.
Data Science & Analytics with AI takes you from spreadsheets to dashboards a business runs on.
Explore the course