What a dashboard is for
A dashboard is a single page that shows the handful of figures someone needs to understand how something is going. It might summarize a small business’s weekly orders, a project’s progress or a household’s monthly spending. The aim is a quick read: a person should be able to look at it for ten seconds and know whether things are on track, and where to look if they are not.
Most dashboards fail not because the data is wrong but because there is too much of it. The goal of a simple one is to leave things out deliberately.
Start with the questions
Before choosing any chart, write down the two to five questions the page should answer. For example, a hypothetical one-person online shop might ask:
- How many orders came in this week compared with last week?
- Is the total sales figure going up or down over the past few months?
- Which product categories account for most of the sales?
- How many orders are still waiting to be shipped?
Each question suggests a figure or a chart. If you cannot name a question a figure answers, it probably does not belong. It also helps to know who will read the page and how often. A page checked daily by one person needs different content from a monthly summary for a group.
Choose a few key figures
Pick the smallest set of measures that answers your questions. A good measure is one a reader can act on or at least understand without a long explanation. Some practical guidelines:
- Prefer a few headline numbers at the top: the current value of each key measure, in large type.
- Add context. A number alone is hard to interpret. Show it next to last week’s value, a target or a simple change, so the reader can tell whether it is good, bad or normal.
- Use consistent definitions. If “orders” means placed orders in one place and shipped orders in another, readers will be confused. Write the definition in a small note.
- Be careful with averages. If a few extreme values could distort the picture, consider showing the median average alongside.
Lay it out clearly
Readers tend to scan from the top left. Put the most important figure there and follow with supporting detail. A few simple layout habits go a long way:
- Group related items together and leave white space between groups.
- Align things. Items lined up in rows and columns look calmer than scattered boxes.
- Keep one chart per question and give each a plain-language title that states what it shows, such as “Weekly orders, last 12 weeks”.
- Label axes and units, and say what period the figures cover and when they were last updated.
- Use color sparingly. One or two colors, with a highlight color reserved for something that needs attention, work better than a rainbow. Remember that some people cannot tell certain colors apart, so do not rely on color alone to carry meaning.
The choice of chart matters too. A line usually suits change over time, bars suit comparisons between categories, and a table is often better than a chart when readers need exact values. The basics of chart types are worth reviewing before you build, and the wider field of data visualization offers further principles.
Avoid clutter
Clutter is anything that makes the reader work without helping them understand. Common culprits include:
- Decorative effects such as heavy shading, three-dimensional bars and background images.
- Too many charts squeezed onto one screen. If a page needs scrolling through twelve visuals, it is probably two pages.
- Gridlines, borders and legends that repeat what is already obvious from the labels.
- Long tables of raw data that belong on a separate sheet.
- Many decimal places where whole numbers would do.
A useful test is to remove an element and ask whether anyone would miss it. If not, leave it out.
Keep the data and the dashboard apart
In a spreadsheet, a tidy approach is to keep three layers on separate sheets:
- Raw data, with one row per record and a clear header row.
- Calculations, where totals, counts and comparisons are worked out with formulas, or with pivot tables for grouped summaries.
- The dashboard, which only displays results from the calculation layer.
This separation keeps the dashboard readable and makes errors easier to trace. If a number looks wrong, you can follow it back through the layers. It also means you can add new rows of data without redesigning the page. Check ranges when you do, since a formula that stops at the old last row is a classic source of silently missing data, so check the ranges after every update.
Maintaining it
A dashboard that is never updated, or whose data nobody trusts, is worse than none. Decide who updates it and how often, and add a visible “last updated” date. A short review habit helps: tie it to a weekly review so you notice when a figure has stopped being useful. If a measure never prompts a decision, consider replacing it.
If other people will view the page, think about sharing permissions. Viewing access is usually enough for readers, and sensitive underlying data may need to stay on a sheet they cannot see. When the dashboard draws on website analytics or other sources, remember that figures from different sources may be defined differently, so note where each came from.
A short checklist
- Can a reader state the main message within ten seconds?
- Does every figure answer a named question?
- Is there a comparison or target next to each key number?
- Are titles, units, dates and definitions clear?
- Is anything on the page purely decorative?
- Can the numbers be traced back to the data?
A modest, readable dashboard that is trusted and used is more valuable than an elaborate one nobody checks.