January 6 2025

What is the dashboard for?

Not always an easy answer

Data in a dashboard serves lots of different purposes. The dashboard in your car ideally tells you how fast you’re going, to a fairly high level of accuracy at a glance. If the fuel is starting to get a bit too low, or the engine is a bit hot, it’ll light up a warning. It doesn’t tell you the amount of fuel you have in the tank to two decimal places, because that is never going to be useful to you, nor does it tell you speed in 20 km/h increments. Half a tank of fuel is good to know about, less than a quarter, probably best to fill up, particularly if you’re driving in remote areas in Australia. You mightn’t get another chance. Speed needs to be fairly accurate for a host of reasons. Engine temperature, probably just tell me when it is a problem.

Dashboards for a business are pretty similar, really. They should ideally be built to answer a some specific questions for a user, at a practical level of granularity. That isn’t the only case though, and trying to fit everything into the best practice model will ultimately leave your users frustrated. There will be some users who just want a big pivot table that they can compare a whole host of different metrics. I actually quite like making these. Sure they tend to be a bit ugly, but more often than not, these are the types of things that actually get a lot of use from the users, and that’s what I see as being a successful dashboard.

It is very common for someone to spend a lot of time making a dashboard look attractive and highly interactive, only for the users to download the data out of it and put it into excel. It can be disheartening for the BI developer, and annoying for the user. Save yourself the heartache and give the user what they are after, and in time, add on the things like the ‘low fuel light’ equivelant. Having your users trust and rely on the data is the first step. You can build things up from there.

But when you are building out a dashboard and dealing with metrics, think about what level of granularity is actually important to your users. If you aren’t sure, ask them. If you’re dealing in millions, or hundreds of millions of dollars, it is probably fair to say you can do away with the decimal places and round things a bit. Always keep the scale in mind. Take it back to the car dashboard. Is the metric like your speed, fuel, or engine temp?