The executive dashboard that actually supports decisions

In many SMEs and mid-sized companies, the executive dashboard is a twenty to thirty page document, produced around the tenth day of the month, discussed for an hour and rarely followed by a decision. It is not wrong. It is simply hard to act on. In our approach to finance and performance, the starting point is plain: a dashboard is judged by the decisions it prompts, not by what it displays. Here is a method in six steps.

Reporting or dashboard: the difference that matters

Reporting answers the question "what happened". Banks, shareholders and auditors need it. An executive dashboard answers a different one: "what do we do this week or this month". It is useful when, for every figure shown, you know who acts, at what threshold and with which options.

Three symptoms show that a dashboard is not doing that job: more than twenty indicators, figures disputed in the meeting, and no link to the actual calendar of decisions.

Step 1: start from decisions, not data

In a two-hour workshop, list the five to eight decisions leadership makes repeatedly: adjusting prices, committing capital expenditure, hiring, arbitrating inventory, accepting a large order, financing working capital. For each, ask two questions: what information would trigger it, and how often is it needed? Any indicator that connects to no decision leaves the scope.

Step 2: keep ten to twelve indicators

Aim for a balance between outcome indicators (margin, invoiced revenue) and forward-looking ones (order backlog, weighted pipeline, 13-week cash forecast). The first tell you where you are, the second where you are heading. A common split:

• growth: order backlog in months of activity, weighted pipeline;
• profitability: gross margin by product line, capacity utilisation;
• cash: forecast cash position, days sales outstanding (DSO), inventory level;
• organisation: staff turnover, workload of key teams.

A practical rule: if one indicator comes in, another goes out.

Step 3: a threshold, an owner, an action

This is the step most often skipped, and the most decisive. For each indicator, write down:

• the alert threshold, for example DSO above 58 days;
• the named owner, for example the finance director;
• the action planned when the threshold is crossed, for example escalated reminders to the ten largest overdue customers.

These rules change the meeting. Nobody comments on a curve any more. The team notes that a threshold has been crossed and triggers what was agreed in advance.

Step 4: one definition, one source

A leadership meeting that debates gross margin for twenty minutes has a definition problem, not a performance problem. Write a one-page glossary: what each indicator includes, the period, the source, the data owner. Settle the conventions that invite argument: the CHF/EUR rate used, the treatment of freight costs, the scope of subsidiaries. If the dashboard includes individual HR data, remember that the revised Swiss Federal Act on Data Protection (FADP, in force since 1 September 2023) requires use to stay within the stated purposes. Aggregated indicators almost always suffice, and our Data, AI & compliance expertise covers this framework.

Step 5: set the rhythm and the format

Not all indicators move at the same pace. Cash and sales are reviewed weekly, profitability and organisation monthly. One page is enough: current value, comparison with budget and with the latest forecast, twelve-month trend, and a two-line comment written by the owner. The meeting lasts 45 minutes at most, and decisions are recorded with a name and a date.

Step 6: automate what can be, in that order

Once the earlier steps are stable, production of the dashboard can be automated. Take an example with explicit assumptions: two people in finance each spend three days a month assembling figures from the ERP, the CRM and spreadsheets. That is six person-days a month, about 72 a year, close to a third of a full-time position (assuming 220 working days). An automated executive dashboard reduces this to a review step and frees time for analysis. For the most sensitive indicator, cash, a continuous cash-flow forecast is a better answer than a quarterly exercise.

AI can help in two specific places: drafting a first version of the commentary and flagging unusual variances. It does not fix an ambiguous definition. Automating a poorly designed dashboard simply produces, faster, numbers that nobody uses.

A worked example: DSO

Take, for illustration, a Swiss industrial company with CHF 60 million in revenue, assuming for simplicity that all sales are on credit. One day of sales is about CHF 164,000. If DSO moves from 55 to 62 days, roughly CHF 1.15 million of additional cash is tied up.

With monthly reporting, the drift shows up at closing, sometimes six weeks after it began. With a weekly indicator, a threshold at 58 days and an agreed action, the response comes within one to two weeks. The difference does not come from the sophistication of the tool but from the decision prepared in advance.

Three pitfalls

• Stacking rather than choosing. Each department asks for its own indicator, and the dashboard becomes a directory of numbers.
• Mistaking precision for usefulness. A margin to the cent is worth little if the decision rests on a trend.
• Never revisiting. Decisions change as the business changes; a half-yearly review of the indicators is a minimum.

Getting started in 30 days

Week 1: decision workshop. Week 2: select indicators, thresholds and owners. Week 3: glossary and data sources. Week 4: a first one-page version, tested in a meeting and then adjusted. Automation follows, on a model that has already proved itself.

If you want to see where your current dashboard stands and which decisions it fails to inform, our diagnostic can establish that in a few conversations, with no commitment.