Business intelligence: from scattered data to usable dashboards

Business intelligence: van losse data naar bruikbare dashboards
Date
November 19, 2026
Author
Isatis Group
Category
Blog
Read time
8 min read

Learn how business intelligence turns scattered data into dashboards people use. Five practical steps, from the first decision to reliable numbers.

It's Monday morning and the management meeting starts in ten minutes. The finance employee has a revenue figure from the accounting package, sales pulls a different number from the CRM, and the planner has a third version in an Excel file. Nobody thinks their own figure is wrong, and yet they don't match. So the meeting starts with the question of which number counts, and only afterwards, if there's time left, with what to do about it.

Business intelligence is the collecting, combining and presenting of company data so that people can make better decisions with it. In practice it comes down to three things: sources that come together in one place, agreements on what a figure means, and dashboards that show the right people what they need to know. The technology is rarely the problem, but the agreements around it often are.

Business intelligence in practice: what it is and what it isn't

Many companies already work with it. In 2023, 67 percent of companies in the sectors surveyed used at least one of the software types ERP, CRM or business intelligence, according to Statistics Netherlands (CBS). That figure covers all three types together. So it says nothing about how many companies use BI alone, nor about whether it helps them make better decisions. What it mainly shows is that this is no longer an exotic topic.

It helps to know what it isn't:

  • Not a software package you buy and switch on. A tool only shows what you put into it. If the source is messy, you get messy charts.
  • Not a nice-looking report. A report nobody opens helps nobody, however good it looks.
  • Not an AI project. AI can add value later, but first the basic figures have to be right.
  • Not a one-off project. Sources change, questions change, and so your dashboard has to change with them.

Step 1: start with the decision, not the data

The most common mistake is to start by unlocking all the data and only then think about what you want to do with it. That produces dashboards with forty charts that nobody looks at. Turn it around. Ask each team which decision they make regularly and which figure is missing or arrives too late. A few examples:

  • Planning: do we need to put extra people on next week or not?
  • Finance: which invoices are past their payment term, and for which customers?
  • Operations: where is work stuck, and for how long?
  • Management: are we still earning what we think we earn on a particular product?

Pick two or three to start with. For each decision, write down who makes it, how often, and which figures go with it. That's your first design.

Step 2: bring the sources together

You need one shared source when the same question gets a different answer depending on the system you consult. That's the moment when separate systems, spreadsheets and databases have to come together in a data platform that people can actually work with. You don't automatically choose the biggest model. A data warehouse is a central database set up specifically for reporting. A data lake stores raw data in all kinds of formats. Sometimes something simpler is enough, matched to where your organization is right now. At Isatis we build this kind of data and AI solutions, including the platforms underneath.

Before you begin, answer a few practical questions:

  • Which systems supply data? Make a list, including the Excel files someone keeps on the quiet.
  • How fresh does it need to be? For monthly reporting, a nightly refresh is more than enough. On the shop floor it may be different.
  • Who manages the connections? A connection that breaks without anyone noticing gives you wrong figures with a neat look.
  • What do you do with the spreadsheets? If a process currently lives in Excel, it may be smarter to capture it in an application first. Read how to replace Excel with a web application if you're considering that.

Step 3: define each figure once

Many discussions about figures are really about definitions. Is revenue based on invoice date or delivery date? Do you count with or without VAT? When is a customer active? As long as that isn't written down, you build dashboards that each show their own truth.

Make a short list of the most important figures. For each one, note which source it comes from, how it's calculated and who owns it. One page is enough. When someone doubts a number later, that's where you look first.

Step 4: design for the people who use it

Data is only useful if the right people can see it and act on it. So work with the users to figure out what they need to see, how often and in what context. A team leader on the shop floor has different questions than a finance director. A few design rules that work in practice:

  • One screen, one purpose. If you want to show several things, build several dashboards.
  • The figure on top, the cause underneath. Anyone who sees the total wants to be able to click through to what drives it.
  • Use color only for deviations. If everything is red, yellow and green, nothing stands out anymore.
  • Show the date of the last refresh. That way everyone knows whether they're looking at current figures.
  • Allow drill-down to the underlying records. This builds trust, because people can check for themselves where a number comes from.

Production environments have their own pitfalls. You can read more about them in why an OEE dashboard often fails and what works instead.

Step 5: test the figures and keep them reliable

A dashboard that shows a wrong number even once quickly loses its users' trust, and that's hard to win back. So treat the figures like software that gets tested:

  • Compare with the source. Take spot checks and see whether the dashboard says the same as the system the data comes from.
  • Build in automatic checks. Think of an alert when a source hasn't been delivered or when a count suddenly drops to zero.
  • Agree on a change process. When a definition changes, everyone needs to know.
  • Keep watching after go-live. Who actually opens the dashboard, and which questions remain? Adjust it accordingly.

The temptation to stop looking at a dashboard after delivery is strong. But that's exactly when you find out whether it fits into the work.

Where AI fits with your dashboards

AI is an addition to a good foundation, not a replacement for it. Once the data comes together reliably, you can look at forecasting, anomaly detection and automatic document classification. Think of an alert when a figure deviates from what you normally see, instead of someone scrolling through charts every morning. Choose AI only where it saves time or catches mistakes. Make sure the outcome is explainable and testable, and that your own team can maintain it. The decision stays with a person.

Frequently asked questions

What does business intelligence mean?

It means bringing company data from different sources together and presenting it so that people can make decisions with it. Usually that happens in dashboards and reports. So it's about the whole path from source to decision, not just the chart at the end.

What is the main goal of business intelligence?

Better decisions based on figures that everyone trusts. When teams no longer have to correct each other about which number is right, the goal has largely been reached.

How much data do you need to get started?

Usually less than you think. Most organizations already have more data than they realize. Start with one decision and the two or three sources that go with it, and only expand once that works.

Should I use AI right away?

No. First make sure the basic figures are right and that people use the dashboard. AI only makes sense when there's reliable data underneath and you have a concrete problem, such as detecting deviations or sorting documents.

Not sure whether your data is ready for dashboards, or don't know where to begin? Get in touch for an introductory call of 15 minutes. You tell us what you want to achieve and we'll tell you honestly whether we can help. We respond to a message within 24 hours.


About the author: Jeroen Penninx, Director and co-founder of Isatis.

Share this article

Jack van Poll

Jack van Poll

Co-Founder, Isatis

Writes about nearshore engineering,
software partnerships and building teams that last.

Get in touch

Read next.

All articles