Why dashboards don’t improve decisions on their own

Most businesses that invest in dashboards get better visibility. Fewer get better decisions. That gap usually has little to do with the dashboard itself. It comes from the assumptions every dashboard quietly relies on.

The research has noticed the gap too. A review of studies on dashboards in performance management found that, despite their popularity, “little is known about the extent of their effectiveness”.

The three assumptions

A dashboard only changes anything if three things happen:

• Someone looks at it, at the right time
• They notice the thing that matters, among everything else on the screen
• They act on it, and someone follows up

Each of those is less likely than it sounds.

People look less often than you’d think

Dashboards often get heavy use in the first weeks after launch. Then the novelty wears off. People check them before a meeting, or when someone asks a question. A problem that started on the third of the month may not be seen until the monthly review, by which time it has been costing money for three weeks.

The important change is easy to miss

A typical management dashboard has dozens of numbers, most of them moving a little every week. The one that matters looks much like the ones that don’t. Spotting it takes experience, attention and time, and the person with all three is usually busy with something else.

The same review found dashboards are more likely to succeed when they deal with that information load, for example by letting people drill down from a summary into the detail.

Often the real issue sits across two screens. Margin looks fine overall. Freight looks fine overall. The problem is one customer whose freight cost rose while their pricing stayed put, and no single chart shows that.

Seeing a number is the easy part

Even when the right person notices the right number, nothing has happened yet. Someone has to dig into the cause, decide what to do, assign it and check it got done. In plenty of businesses, the monthly meeting notes the same variance three months running.

Everyone sees the number move. Nobody owns what happens next.

What helps

• Fewer numbers on the screen, each chosen because someone will act on it
• Alerts when something moves outside its normal range, so nobody has to go looking
• An explanation alongside the number: which customers, sites or products drove the change
• A named owner and a due date for every action that comes out of a review
• A record of what was decided, and whether it worked

Some of this is discipline and some of it is tooling. Newer approaches, often called decision intelligence, do the watching and the first round of explaining automatically, so people can spend their time on the decision itself.

Dashboards still matter

None of this is an argument against dashboards. A shared, trusted view of performance is the foundation. The point is that the dashboard is where the work starts, and most of the value comes from what happens after someone has looked at it.