What it calculates

A KPI dashboard presents a small set of leadership metrics against their targets, with a traffic-light status on each, so a management meeting starts from a shared view of what is on and off track.

How it is calculated

Each KPI carries a current value, a target, a direction of good - higher is better for revenue, lower for churn - and thresholds that set the status bands. Status is derived from the gap to target relative to those thresholds rather than judged case by case, which keeps the colours consistent between people and between months. A trend against prior periods sits alongside the status, because a red metric improving fast and a green metric deteriorating are very different situations.

How to read the result

Keep the set small. A dashboard with forty metrics has no signal, because everything is amber somewhere and attention has nowhere to land. Prefer leading indicators over lagging ones: pipeline coverage tells you about next quarter, whereas revenue tells you about last. Every metric should have an owner who can actually influence it - a KPI nobody owns generates discussion and no decision.

Worked example

Revenue at 4,200,000 against a 4,500,000 target is 93% of goal - amber under a band that puts red below 90% and green at or above 100%. Churn at 2.8% against a 3.0% target is green, because lower is better. Pipeline coverage at 2.4x against a 3.0x target is red and, being a leading indicator, matters more than the revenue miss it predicts.

Common questions

How many KPIs should a dashboard have?

Roughly five to nine at the top level. Beyond that the eye stops prioritising and the dashboard becomes a report. Additional metrics belong in supporting views that owners use, not in the leadership summary.

What is the difference between a leading and a lagging indicator?

A lagging indicator reports an outcome that is already fixed, like closed revenue. A leading indicator moves before the outcome and can still be influenced, like pipeline coverage, trial starts or support backlog. A dashboard of only lagging indicators is a history lesson.

Should thresholds be the same for every metric?

No. A 5% miss on revenue and a 5% miss on system uptime are not comparable events. Set bands per metric based on what a meaningful deviation looks like for that measure, and review them when the business changes rather than leaving last year's tolerances in place.