Skip to content

Dashboards

The dashboard problem is almost never a charting problem. It’s a trust problem. People come to me because their numbers live in five places, no two of them agree, and every important meeting starts with someone exporting spreadsheets and arguing about whose figure is right. They think they need a dashboard. What they actually need is a single, trustworthy version of their own data — and the dashboard is just the part you can see. That’s the work I do here.

The problems I solve here

The most common: a business running on exported spreadsheets and gut feel. The data exists — in the CRM, the accounting tool, the product database, the ad platforms — but it’s scattered, and pulling it together is a manual ritual that eats hours and produces numbers nobody fully believes. The job is to bring those sources into one place the business owns and surface the handful of numbers that actually drive decisions.

The second: the dashboard that’s a wall of charts and tells you nothing. Someone built or bought a reporting tool, switched on every metric it offered, and now there’s a screen so busy that no one looks at it. A dashboard with forty numbers on it has no point of view, which means it can’t support a decision. The fix is subtraction — finding the five numbers that matter and cutting the rest.

The third: the per-seat BI tool that’s quietly become a tax. A business paying steadily climbing fees for a hosted analytics platform that holds their own data hostage and charges per user to look at it. The honest answer is often to build a lean reporting layer on their own database, own it outright, and stop paying rent to see numbers they already generated.

How I think about dashboards

A dashboard is only as good as the data underneath it and the question it’s built to answer. Get either wrong and prettier charts won’t save it. So I start at the bottom — what data exists, where, how reliable it is, and whether it can be trusted to agree — long before anything gets drawn. Most failed dashboards failed at the data layer and tried to fix it at the presentation layer. You cannot chart your way out of bad data.

The second principle is that a dashboard must have a point of view. Its job is to answer specific questions a specific person needs to act on — is the business healthy, where’s the bottleneck, what changed this week. If a number on the screen wouldn’t change a decision, it shouldn’t be there. The instinct to show everything because the data exists is the enemy of a useful dashboard. Restraint is the skill.

And underneath all of it: this is your data, and a dashboard is the clearest argument I know for owning it. When your numbers live in your own database, you can answer any question you think of next, cross-reference sources a vendor would never connect, and build whatever view you need without asking permission or paying per seat. Rent your analytics and you can only see what the tool decided to show you. Own the data and the dashboard becomes a starting point, not a ceiling.

How I’d actually do it

I’d build to the one decision first. Rather than a grand unified reporting platform, I’d take the single question that’s currently most expensive to answer — say, true profitability per client, or which part of the pipeline is leaking — and build just the data plumbing and the one view to answer it well. That ships fast, proves the value immediately, and gives a foundation to add the next question onto. A dashboard built one real decision at a time stays useful; one built to show everything from day one becomes the wall of charts nobody reads.

Technically that usually means consolidating the relevant sources into a database the business owns, a clean layer to keep it current, and a deliberately simple front end — often a few well-chosen views rather than a sprawling app. The restraint is the point. The aim is a small number of numbers the business trusts enough to act on without re-checking, on data they fully control.

Where this connects

This is one of the specialisms I’m hired for. Pulling scattered sources into one trustworthy place is, underneath, an API integrations problem — that’s how the data gets there — and it rests on getting the databases and data model right. The engagement this most often powers is an MVP development: the lean, owned reporting layer that replaces the spreadsheet ritual and the per-seat tool.

When you need a team instead

A large reporting platform feeding a whole organisation, maintained and extended continuously, is more than one person’s job. That’s Digital Royalty, my agency, with the team to build and run it at scale. My lane is the senior call — what to measure, what to ignore, how to architect it so it stays honest — and the lean first version. I’ll tell you which you need.

Who this isn’t for

If you want every metric your tools can produce displayed on one screen, I’m not your person — I’ll argue for fewer numbers, not more. I’m worth hiring when you want a dashboard that actually changes what you decide.

Talk to me

If you don’t trust your own numbers, or you’re paying per seat to look at data you already own, tell me about your problem. The fix is usually simpler and cheaper than the tool you were about to buy.