KPI dashboards that Swiss executives actually use
Why most executive dashboards go unread, what belongs on one, and the data-residency questions a Swiss company has to settle before choosing a tool.
Most executive dashboards are built, admired once, and then quietly abandoned. The build was not the problem. The problem is that nothing in the organisation depends on anyone looking at it.
A dashboard is not a reporting artefact. It is the input to a recurring decision, and if no decision is scheduled against it, it is a screensaver with good typography.
Why they go unread
Four causes, in rough order of frequency.
No decision is attached. The dashboard shows the state of the business. Nobody has said what they would do differently at any particular value, so nothing follows from reading it.
Too many numbers. Thirty metrics is an archive, not a dashboard. An executive can hold five to seven in working memory; beyond that, attention distributes randomly.
Numbers nobody trusts. If two departments produce different revenue figures, the dashboard inherits the disagreement, and the meeting becomes about the data rather than the decision. That is a master-data problem showing up downstream, covered in AI architecture: why data governance decides the outcome.
Stale by the time it is read. A monthly refresh for a weekly decision is worse than no dashboard, because it invites confident decisions on old information.
What belongs on an executive dashboard
A working rule: one screen, five to seven metrics, each attached to a person and a decision.
Four groups usually cover a mid-market business:
- Financial health — revenue against plan, gross margin, cash runway. Not the full P&L; the two or three that move first when something is wrong.
- Customer — new customers, churn or retention, and one satisfaction signal. Pipeline value if the sales cycle is long.
- Operational — the throughput measure that defines your capacity, whatever it is: orders shipped, cases closed, projects delivered.
- Risk — the one that Swiss boards ask about and most dashboards omit: open compliance items, security incidents, key-person concentration.
For each, three things must be defined and written down: who owns it, what value triggers action, and what that action is. A metric without a trigger is decoration.
The trend matters more than the value
A single number tells you almost nothing. Revenue of CHF 2.4 million is good or bad depending on the last four periods and the plan.
Every metric needs a comparison — against the previous period, the plan, or a threshold. The most common dashboard design error is a large number with no context beside it, which is why they generate questions rather than answers.
Data residency and privacy, before the tool is chosen
This is where Swiss companies need to think slightly differently from generic BI advice.
An executive dashboard often aggregates personal data — customer records, employee metrics, sometimes health or financial detail depending on your sector. Under revDSG/nDSG, and GDPR wherever EU data is involved, that carries obligations that constrain the tool choice:
- Where is the data processed? Many BI tools process in the vendor’s cloud, which may or may not be in a jurisdiction you can accept.
- Who can reach it? Vendor support access is the question that gets missed. Storage location is in the marketing material; access control is in the contract.
- Does the dashboard need personal data at all? Aggregates and pseudonymised data usually answer the executive question just as well, and carry a fraction of the obligation. This is the cheapest fix available and it is routinely skipped.
- Retention. Dashboards accumulate history by design. Some of that history you are obliged to delete.
For FINMA-supervised businesses, add the usual outsourcing requirements: documented risk assessment, audit rights, and a demonstrable exit.
Settle these before selecting a tool. Reversing a BI platform decision after it holds two years of history is expensive.
Tool choice, briefly
The tool matters less than most selection processes assume. What actually differentiates:
- Where it processes data, per the section above
- Whether it connects to your sources without a nightly manual export
- Whether non-technical people can read it without training — if a dashboard needs a briefing, it will be read by the person who built it
Anything else — visualisation breadth, AI features, marketplace size — is secondary for an executive dashboard with seven metrics on it.
Designing one that survives
Five practices that separate dashboards still in use after a year from the rest:
- Start from the decision, not from the available data. Write the decision down first.
- One screen. If it scrolls, it is a report.
- Put the owner’s name on each metric, visibly. Ownership that is not written down does not exist.
- Schedule the reading. Fifteen minutes in an existing recurring meeting beats a beautiful dashboard nobody opens.
- Remove a metric when you add one. Dashboards only grow unless something resists it, and growth is how they die.
Point 5 is the one that matters most over time and the one nobody enforces.
Related reading
If you want help defining the seven metrics that matter for your business, book a call. No obligation.