← BACK TO THE BLOG DATA 13 SEP 2026

From the board's spreadsheet to a dashboard people actually use

BY MIGUEL·10 MIN READ

Printed sheet with line, bar and pie charts on a desk, next to a laptop and a notebook with a pen

Almost every company that asks us for a dashboard already has one. It's called "the board spreadsheet", one person maintains it, it takes two or three days to close each month, and it lands after the decisions that depended on it have already been made. The tool isn't the problem: nobody ever decided what should be watched or who owns each number.

That's what this guide is about. What an executive dashboard is, how to choose the metrics (and above all how many), how to structure it in levels so leadership and operations aren't fighting over the same screen, and why 80 % of the work happens before you open the visualisation tool. No vendor recipes, and with examples you'll recognise.

What is an executive dashboard, and how is it different from a report?

An executive dashboard is a small set of agreed indicators, each with an owner, that shows the state of the business across several dimensions at once and refreshes itself. The "balanced" idea comes from Kaplan and Norton's balanced scorecard: looking only at the P&L tells you about the past, so you balance it with what explains the future (customers, internal processes, team capabilities).

Three things separate it from a report:

That third point sinks the most projects. A beautiful tool with definitions nobody owns decays within six months: someone changes a criterion in the ERP, the number shifts, nobody notices, and the board stops believing it. From there it's back to the spreadsheet. If this sounds familiar, your problem is data governance, not dashboards.

A dashboard doesn't fail because it's ugly. It fails because someone looks at a figure, doesn't believe it, and nobody is responsible for clearing it up.

Which metrics should it include (and how many)?

Short answer: between 8 and 15 on the leadership view. The useful answer is how you get to those.

Start from decisions, not from data. Ask each member of the leadership team: which decisions do you make every month, and what number would change your mind? If a metric changes no decision, it's informational decoration. It leaves the dashboard and, if someone wants it, it lives in a separate report.

Then spread them across the four classic perspectives so you don't end up with a purely financial panel:

| Perspective | What it answers | Concrete examples | |---|---|---| | Financial | Are we making money, and at what margin? | Gross margin per line, recurring revenue, days sales outstanding, variance to budget | | Customer | Do they choose us and come back? | Repeat rate, net gains and losses, response time, complaints per 100 orders | | Internal processes | Do we deliver well and at cost? | Average lead time, orders shipped complete, rework, cost per case processed | | People and capabilities | Will we still be able to do this? | Turnover in key roles, open vacancies, training hours, single-person dependency in critical processes |

Two practical rules when choosing:

And a warning about the 40-metric mistake: it always arrives the same way. To avoid offending anyone, one metric per department goes in and the panel becomes a mural. Then nobody reads it, each area looks only at its slice, and the "balanced" effect disappears. Better to leave out an indicator one area will complain about than to make it unreadable for the managing director.

How do you structure it so both leadership and operations use it?

With three levels, not one screen for everyone. This is the structure that works best in companies of 50 to 500 people:

That third level gets skipped almost every time and it's the most profitable one. Trust in a dashboard isn't won with explanations; it's won by being able to click through to the invoice line.

Two design details that change actual usage:

How much work is actually behind it?

Less than some people fear, and in a different place than most imagine. The usual split for a first dashboard:

| Phase | What it includes | Share of effort | |---|---|---| | Definition | Leadership interviews, metric list, a written definition for each, owner assignment | 15-20 % | | Data | Connecting ERP, CRM and files; unifying customers and products; history | 40-50 % | | Business logic | Implementing each definition: what an active customer is, how a discount is allocated, what gets excluded | 20-25 % | | Visualisation | Building screens, filters, access | 10-15 % | | Adoption | Short training, review routine, tweaks in the first weeks | 5-10 % |

When a vendor shows spectacular dashboards in the first meeting, they're showing that final 10-15 %, which is the cheap part. The expensive part — and the one that decides whether the project survives — is unifying data and writing the logic. If that layer interests you, we explain it in what business intelligence is, and platform choice in business intelligence tools.

Realistic timelines: a first leadership dashboard with two or three sources connected takes 4 to 8 weeks, as long as someone from the business is available to settle definitions. The bottleneck is never technical; it's getting three people to agree on what "margin" means.

When is it worth building, and when should you wait?

Worth building when at least two of these are true:

Worth waiting — or doing something smaller first — if:

In that case the prior step is usually sorting out priorities and sources, which is exactly what a data strategy covers before you touch any tool.

What kills a dashboard (and how to avoid it)?

The five we see most often, each with its antidote:

If you've already attempted a project like this and it stalled halfway, it's worth reading why AI pilots fail: the reasons are almost identical, and none of them are technological.

Frequently asked questions

How many metrics should an executive dashboard have?

Between 8 and 15 on the leadership view. Below 8 you're usually missing a perspective of the business; above 15 it stops being readable at a glance and each area looks only at its own part. The detail doesn't disappear: it moves down to the departmental views.

What's the difference between a dashboard and a balanced scorecard?

A dashboard can be purely financial or operational. A balanced scorecard balances four perspectives — financial, customer, processes and people — so you don't decide by looking only at past results. In practice, "balanced" forces you to include metrics that anticipate.

Can you build a dashboard in a spreadsheet?

Yes, and for a first version it's legitimate if the data load is automated. The limit shows up with several sources, large history and multiple users editing: at that point the spreadsheet stops being a reporting tool and becomes a versioning risk.

How long until it's live?

A first leadership dashboard with two or three sources usually takes 4 to 8 weeks, with business availability to agree definitions as the big constraint. The technical part is rarely the bottleneck.

If you're at the "we have data but we decide blind" stage, the cheapest start is a diagnosis: which sources exist, which metrics actually matter and what can be built in a month. That's what we do in the audit, and if you'd rather talk it through in half an hour and have us tell you whether it's worth it, let's talk.

Shall we apply it to your case?

The 360° AI Audit turns these ideas into a concrete plan for your company: three weeks, fixed price and the full picture of your AI before spending a euro.

See the 360° Audit→ Let's talk↗