A dashboard is the last step, not the first
"We need a dashboard" is one of the most common requests we hear, and one of the most understandable. Owners want to see the business at a glance, without waiting for someone to pull a report.
It is also one of the easiest things to get wrong. Plenty of businesses have a dashboard nobody opens. It usually looked good on launch day. What failed was everything underneath it.
The chart is the easy part
Drawing a chart takes an afternoon. The hard part is getting numbers into it that are complete, current and defined the same way every time. Most of a dashboard project, done properly, is plumbing:
- Getting the data out. Every system stores things differently and exposes them differently. Some have good connections, some need workarounds, and some still only export a file.
- Joining it up. The same customer or product has to be recognisable across systems, or nothing lines up.
- Keeping it fresh. A dashboard that someone updates by hand on Mondays is a report with extra steps. Data should arrive on its own, on a schedule that matches how often decisions are made.
- Handling the gaps. Missing values, late records and one-off corrections all need a rule, or the numbers will wobble and people will stop trusting them.
When that work is skipped, the dashboard shows the problem rather than the business. People notice the first wrong number, stop looking, and go back to asking someone to pull a report.
Start from the decision
The better starting point isn't "what should we measure?" It's "what do you decide each week, and what would you need to see to decide it faster?"
A logistics operation might decide each morning which routes need another vehicle. A practice might decide each week which work is at risk of missing a deadline. A store might decide which lines to reorder. Each of those decisions points to a handful of numbers, not fifty.
A dashboard built around three real decisions gets opened. One built around everything that could be counted gets admired once.
Fewer numbers, defined once
Every number on the screen should have a one-line definition that anyone in the business can read. If two people would explain a figure differently, it isn't ready to be on the dashboard yet. That definition becomes part of the build, not a footnote.
It also helps to show where each number comes from and when it was last updated. Trust grows quickly when people can check a figure for themselves, and collapses quickly when they can't.
Then, and only then, the dashboard
Once the data arrives on its own, joins cleanly and means one thing, the dashboard is quick to build and easy to change. Adding a new view becomes an afternoon's work instead of a project, because the foundation is already doing the heavy lifting.
That is the order we work in: plumbing first, definitions second, charts last. It is less exciting to describe, and far more likely to still be in use a year later.