What a dashboard is
Business intelligence is the practice of collecting, organizing and presenting an organization’s data so that people can make better-informed decisions. A dashboard is its most visible product: a single screen of numbers, charts and tables that summarizes how something is going. A sales dashboard might show revenue this month, orders by region and the number of open leads. A support dashboard might show waiting times and unresolved tickets.
Behind the screen there is usually a chain of steps. Data is pulled from operational systems, cleaned, combined and stored, often in a data warehouse. Rules then turn it into measures, and a reporting tool displays them. Each step involves choices, and each choice shapes what the viewer sees.
What dashboards do well
- Speed. A glance can reveal that something has changed without anyone building a report.
- Shared view. When a team looks at the same figures, discussions can begin from the same facts.
- Monitoring. Regular measures make it easier to notice a trend, a spike or a drop.
- Self-service. Many tools let non-specialists filter and explore, which reduces the queue for analysts.
- Accountability. Visible targets can focus attention on what matters, though this has side effects discussed below.
What dashboards hide
Definitions
Every number depends on a definition. “Active user” may mean someone who logged in once this month or someone who took a meaningful action. If the definition is not visible on the dashboard, two teams can read the same figure differently. Agreeing on such definitions is part of data governance.
Averages and totals
A summary can conceal the spread underneath. An average response time can look fine while a minority of customers wait very long. A total can rise while one important segment declines. A single headline figure rarely shows these differences unless the dashboard also breaks the numbers down.
Missing context
A figure without a comparison is hard to interpret. Is 5,000 orders good? That depends on last month, last year, the season and the target. Dashboards that show only the current value invite guesses.
Data quality problems
A dashboard looks authoritative whether or not the data is sound. Duplicate records, missing entries or a broken data feed can all produce tidy-looking charts. Problems in the underlying records carry through to the display, a risk explored in the article on data quality. A sudden jump or drop is sometimes a real event and sometimes a fault in the pipeline.
Timing
Dashboards may be refreshed hourly, daily or in real time, and viewers do not always check. Decisions made on stale data can be wrong without anyone noticing. A visible “last updated” note is a small but valuable feature.
What is not measured
Dashboards show what is measured, which is not the same as what matters. Customer trust, employee morale or long-term risk are hard to count, so they are often absent. When a team is judged by the numbers on screen, the unmeasured parts can quietly be neglected.
Design choices
How charts are drawn can change the impression they give: where an axis starts, which colors draw the eye, which items are placed at the top. These topics are covered in the article on data visualization.
The target problem
When a measure becomes a target, people may begin to optimize the measure itself rather than the goal it was meant to represent. A team measured on tickets closed per day might close tickets quickly without solving the problem. This is a general pattern in organizations, not a certainty, but it is a good reason to pair any target with a check on quality.
Correlation and causation
Dashboards often place two lines side by side. They may move together without one causing the other. Seasonal changes, marketing campaigns and outside events can all affect several measures at once. Deeper data analytics can test such ideas more carefully than a chart can, although even careful analysis has limits.
Questions worth asking
- Where does this number come from, and who defines it?
- When was it last updated?
- What is it compared with?
- Who or what is left out of this measure?
- Could the underlying records be incomplete or wrong?
- What decision is this dashboard supposed to support?
If a dashboard cannot answer the last question, it may be showing numbers because they are available, not because they are useful.
Building and reading with care
Good dashboards tend to be modest: few measures, clear labels, visible definitions and comparisons, and a named owner who can explain them. Readers also need some digital literacy with numbers, such as knowing the difference between a rate and a count. No dashboard replaces judgment. It is a prompt for questions rather than a final answer, and the people closest to the work often know things the numbers do not capture.