What We Learned from Building 10,000+ Dashboards

At Luzmo, we've helped teams build over 10,000 embedded dashboards. Different industries, different data, different users… but the same patterns keep showing up.

Some of those patterns are mistakes we've watched customers make again and again. 

Others are surprises: things we didn't expect about how people actually interact with dashboards once they're live. A few challenged assumptions we held ourselves.

We asked our delivery team (the engineers, solutions architects, and product people who work closest to our customers) three simple questions: What mistakes do you see repeated? What surprised you? And what's the one piece of advice you'd give someone before they build their first dashboard?

Here's what came back.

#1 Most dashboards fail because they show too much, not too little

The instinct when building a dashboard is to include everything. Every metric, every breakdown, every possible filter. The logic feels sound: more data means more value.

In practice, the opposite is true. Overloaded dashboards don't empower users — they paralyze them. When everything is presented at once, nothing stands out, and the dashboard becomes a wall of numbers that nobody actually reads.

Carmen Vandeloo, Delivery Manager at Luzmo, has seen this pattern play out across hundreds of implementations.

"Too many widgets, many tables, and not investigating the insights actually sought after by users."

Juan C. Tello, Solution Engineer at Luzmo, echoes the same observation from the technical side.

"Dashboards that are way too long and cluttered to be engaging."

The best dashboards don't show more data; they show the right data. That means starting not from what's available, but from what someone needs to act on.

#2 Users don't explore dashboards

There's an assumption baked into most dashboard design: users will explore. They'll hover, click, filter, and discover insights on their own.

That's not what happens.

Radu Pop, Support Engineer at Luzmo, observed this firsthand during conversations with end users.

"On the few occasions where I talked to end users, they were following a formula and not taking in the dashboard as a whole, but only following a template of where they should click. This led to them being blocked easily when one step of their usual process was not working as expected."

This has real design implications. If users navigate dashboards like checklists, then breaking their expected flow (moving a widget, renaming a filter, changing a layout) doesn't just confuse them. It blocks them entirely. Dashboard design isn't just about what data to show but understanding the habits your users have already built around it.

#3 Most teams underuse the tools they're already paying for

When teams build dashboards, they tend to fall back on what they know: bar charts, line charts, number widgets, and tables. It's safe. It's familiar. It also leaves most of the platform's capability untouched.

Joost Stessens, Solutions Engineer at Luzmo, sees this repeatedly.

"Only using column/bar charts, line charts, and number widgets in dashboards, instead of all those great widget types we provide. And talking about 'reporting,' but only providing tables with too much data. I read such tables as 'Analyze it yourself.'"

The table problem is especially common. A dashboard filled with raw tables isn't a dashboard. It's a spreadsheet with a nicer frame. It pushes the cognitive load back onto the user and defeats the purpose of visualization altogether.

#4 Interactivity is the most underutilized feature in dashboards

Most modern dashboard platforms support click-through interactions out of the box — clicking a chart segment to drill down, filtering by selection, linking between views. But most end users have no idea these capabilities exist.

Joost explains:

"I would imagine that not so many end-users are actually aware of the interactivity that most of our charts provide out of the box. They're not used to clicking on chart segments, as that often isn't available in other implementations."

Carmen adds that the most compelling use cases go a step further.

"The more interesting ones for me have always been the ones where they enable the users to take actions from the dashboard."

The gap between what a dashboard can do and what users think it can do is one of the biggest missed opportunities in embedded analytics. If interactivity isn't surfaced or guided, it might as well not exist.

#5 Don't wait for a perfect dashboard: ship it and iterate!

There's a strong temptation to hold a dashboard back until it's complete. Every metric accounted for, every edge case handled, every filter in place. The problem is, users rarely know what they need until they see something in front of them.

Jasper van Nistelrooy, VP of Product at Luzmo, puts it directly.

"Trying to make the first release perfect, instead of releasing early and iterating based on customer feedback. Most dashboard users find it much easier to share feedback on an existing dashboard than explaining upfront what they need or want. Especially with Luzmo Studio, it's super easy to respond to customer feedback quick, and I see often too little of that."

Perfectionism at launch is one of the most common reasons dashboards end up misaligned with user needs. The faster a dashboard reaches real users, the faster it becomes genuinely useful.

#6 The best dashboards are built for decisions, not for data

When asked for one piece of advice before building a first dashboard, the answers from the Luzmo team converge on the same principle: start from the decision, not the dataset.

Joost puts it this way:

"Understand which decisions your customers want to make on their data, and prepare your dashboard for that."

Juan frames it through impact:

"Invest time in properly identifying which are the insights that would have the biggest positive impact on your end users."

Jasper adds a critical nuance — those decisions aren't the same for everyone.

"Different users have different needs. A 'one-size-fits-all' dashboard rarely works, so try to understand varying user needs upfront and think about how you want to cater to the differences — for example, different dashboards, self-serve through embedded editing, or AI."

A dashboard that answers "what happened" is useful. A dashboard that answers "what should I do next" is indispensable.

#7 Users will export your data, and that's not a problem

Dashboard builders often assume that if the dashboard is good enough, users won't need to leave it. But real user behavior tells a different story.

Jasper observed:

"Many end-users export data to do further analysis or combine it with other data they have. That's not a bad thing. That means you should optimize a dashboard to provide a quick glance of the most important things and allow for easy data exports, or recurring scheduled exports."

Rather than fighting this behavior, the best approach is to design for it. A dashboard that serves as both a summary view and a reliable data source (with clean export options) will always outperform one that tries to be the only place users interact with their data.

#8 A dashboard is never done

The biggest risk after launch is an abandoned dashboard. User needs shift, data sources change, business questions evolve. A dashboard that was perfect six months ago can quietly become irrelevant.

Radu highlights why ongoing iteration matters:

"We give the right advice before building dashboards, but it's super important to keep following up with end users in order to keep adapting the dashboards to match their changing needs. This is of course not easy to do and takes time."

Jasper reinforces this from a different angle, it's not just about improving what exists, but about recognizing when users outgrow it.

The teams that treat dashboards as living products, rather than one-time deliverables, are the ones whose analytics actually get used.

Build dashboards your users actually want to open

Ten thousand dashboards taught us this: the gap between a dashboard that exists and a dashboard that gets used comes down to intent. Intent to understand what users need. Intent to ship early and learn fast. Intent to keep improving after launch.

Luzmo makes it easy to embed interactive, user-friendly dashboards into any product — and to iterate on them as your users' needs evolve. 

If you're ready to build dashboards people actually use, try Luzmo today. 

‍