Skip to content

How to kill your dashboards

  • by

Good Read –

I am a passionate dashboard developer – but this is a plea for deleting dashboards

Every knows dead dashboards. They may be the dashboard that everyone knows exists that no one can find. It may be everyones dream, but so detailed and huge, that it takes half an hour for each query. Everyone praises this dashboard for what it can do, but no one uses it. A classic is the amazing dashboard with the bad data, outdated data or simply incorrect data. Also widely known is the dashboard that has been unfinished and unusable for at least two years, along with its little siblings the backups, beta versions, predecessors, and alternate versions.

There are two things that a dead dashboard always has: very few people use it on very few occasions, or even no one uses it. But there is aways one reason to hold on:d

“It just needs one or two more visuals and it will be ready!”
“But I am sure Betsy from B.T. has said she used it for that presentation six months ago and it was super useful!”
“But if the performance was better, more people would use and it would be a great dashboard!”

People hang onto dead dashboards, like to an ugly coffee mug with a crack that they just can’t throw away because five years ago it was their absolute favourite.

So today, I am here to make you let go of old baggage, to free you from fixing old dashboards instead of making new ones, to keep you from reparing rotten data pipelines instead of using the new ones, to help you let go of that coffee mug.

I have thought of two ways to kill a dashboard: the “early grave and resurrection” method and the “death with a warning” method.

But first of all: how do you determine, if a dashboard is dead? I use this formula:

  • How many people are this dashboard’s target audience?
  • How many of them have actually used it in the last 60 days?
  • Have people used it more than twice in this time period?
  • Are there scenarios, when a new group of people may become the target audience?
  •  And how much effort would it take, to make a newer, better dashboard?

These questions never failed me yet. They are like the last nail in the coffin. How to kill the dashboards? Read this:

  1. Early grave and resurrection
    For the new year 2023 the company X cancelled all regular meetings of three or more people, no more dailys, weeklys, monthlys– no exceptions. After a cleansing period of two weeks only those meetings that people wanted back were allowed to be resumed. Lets do that, but with dashboards. Lets delete all dashboards and see, who wants what back. Of course, we have a back up in place and are happy to resurrect those, who died too young.
  2. Death with a warning
    In a well monitored analytics environment a team of behind the scenes data stewards monitor when and how often dashboards are accessed. And if there has been no activity for two months, they send an email, warning you that they will delete the dashboard if you don’t object. Why two months? To pick up on a monthly turnus, if there is one.

There is one more method that I wouldn’t want to hide from you: create an assassin. I will go through any analytics platform and flag dashboards for termination. And if there is no one paying me ransom, i.e. convincing me to spare the dashboard, it will be incinerated.  

Throwing away your old mugs from your college days is hard. There are so many memories. So much love that was poured into every cup. But when the deed is done, all the free space in your kitchen cupboard will feel like a breath of fresh air, the weight of the old days gone and the new, fancy and sparkling mugs displayed in their full glory.