kpi-design

Start from your success factors and revenue model, then select one leading metric that predicts each outcome and one lagging metric that confirms it. Cut any metric that does not map to a success factor, initiative, or risk, and keep the set to roughly a dozen headline KPIs.

Leading KPIs predict outcomes and let you act early, such as pipeline value, win rate, or cash runway. Lagging KPIs confirm results after the fact, such as MRR, revenue versus plan, and net margin. Each key outcome should have at least one of each paired together.

Recurring models use MRR, ARR, NRR, and LTV:CAC, while one-time models should use order frequency, repeat-purchase rate, and average order value instead. The revenue model determines which formulas are valid, so it must be confirmed before designing the metric set.

Threshold bands (green, amber, red) are derived from existing targets or forecasts. When no basis exists, thresholds are marked as to-set and the founder must supply the target, since a threshold is a decision rather than a default value.

The KPI is kept but flagged with a data gap naming the source or integration needed, such as time-tracking for utilization. Closing the gap becomes an actionable task rather than silently dropping a valuable metric.

If metrics already exist and only need arrangement into a dashboard, use an executive dashboard builder. For comparing actuals to forecast use a monthly business review, and for root-causing a specific miss use variance diagnosis. KPI design is for defining the metric layer itself.