Open your BI tool and count the dashboards. Then count the ones someone opened last month. For most companies the second number is much smaller.
Ably’s product operations team audited theirs: “Over half of our Metabase dashboards had zero views. The ones people did use often disagreed with each other.” They also found twelve more dashboards hiding in a library of Claude skills, each one a script that built a new dashboard on request.
That is dashboard sprawl. Every new question gets a new dashboard. Old ones never go away. After a year, nobody can say which “Revenue overview” is the real one, so people ask the data team, and the data team builds another.
Why dashboards multiply
- Copying is easier than changing. If a dashboard is almost right, the fastest path is to duplicate it and change one filter. Now there are two.
- Each dashboard owns its formulas. A copy takes its calculated fields with it. When the original is fixed, the copy is not.
- Nothing says which one is official. A dashboard the data team checked looks the same as one an intern made in an afternoon.
- Nobody sees what is unused. Without a view of owners and usage, deleting anything feels risky, so nothing is deleted.
- AI makes more of them. When an agent can build a dashboard in minutes, the count goes up faster than ever.
The fix: fewer places, shared definitions, clear ownership
One dashboard per workspace
In Kater, each workspace has exactly one dashboard, drawn from one dashboard.yaml file. The file cannot be deleted, and there is no “duplicate dashboard” button that makes a quiet copy. A workspace is a topic, such as sales or support. Its dashboard is the one place to look for that topic.
To combine workspaces, you build an app, which is a separate, named object.
Every chart reads one definition
The charts on a dashboard do not carry their own formulas. They read measures from a shared model in Kater SQL. Two dashboards that show revenue show the same revenue. When a definition is fixed, every dashboard that reads it is fixed too. So even when there are many dashboards, they cannot quietly drift apart.
Verified marks say what the data team stands behind
A field is verified only when a shared model defines it, and the verified mark goes with it into queries and apps. A number computed in one workspace has no mark until a promotion moves it into a shared model. The data team can also add a curated tag to the models, queries and workspaces it stands behind. Neither changes permissions. So a reader can tell the checked number from the experiment.
The Objects page shows what is stale
Admins get an Objects page in the admin console. It shows every object’s owner, what uses it, and what is stale or has no owner. You can reassign objects in bulk and export the list as CSV. That turns “I am afraid to delete anything” into a list you can work through. Read more in the Promotions post.
You can search every dashboard
In Kater, everything is a file: models, queries, charts, dashboards and apps. You or your agent can search every dashboard for a metric, see which ones use it and read the definition behind any number. That is very hard to do when dashboards live in a vendor’s database.
Many dashboards should be a digest
A lot of dashboards exist so someone can check, once a week, whether anything changed. A digest does that job better. A dashboard shows you every number. A digest tells you which one to look at. Your agent writes it on a schedule: the one thing that changed, why it moved and what to do.
When the weekly check is a digest, the ten dashboards behind it can be one.
Frequently asked questions
What is dashboard sprawl?
Dashboard sprawl is when a company has far more dashboards than anyone can keep track of. Many are copies with small changes, many are never opened, and nobody knows which one is current. People stop trusting all of them.
How does Kater prevent dashboard sprawl?
Each Kater workspace has exactly one dashboard, drawn from one dashboard.yaml file. Every chart on it reads shared metric definitions. The data team marks what it stands behind with a curated tag, and the admin Objects page shows what each object uses, and what is stale or has no owner.
How do I find which dashboards use a metric?
Everything in Kater is a file, so you or your agent can search every dashboard for a metric, the same way you search a codebase. You can also read the definition behind any number.
What should replace the dashboards nobody opens?
Often a digest. A dashboard shows every number. A digest is a page your agent writes on a schedule that tells you which number moved and why, so people read one page instead of browsing ten dashboards.
Want to see how few dashboards you really need? Book a demo.