Data & Intelligence
KPI Design
We turn your objective into a small set of connected metrics, so each team knows what it owns and how that number feeds the result leadership cares about.
The problem
Every department hits its targets, and the business doesn't move the way those targets promised.
Marketing reports leads, sales reports bookings, product reports active users. With no model linking them, teams can win locally while hurting each other: lead volume climbs, quality slides, and reps waste the quarter.
The usual diagnosis is a wrong metric. More often the metric was never defined: what counts, which date it lands on, who answers for it. And anything used as a target changes behavior. That's Goodhart's law, and it's why a target needs a counterweight.
You may recognize this
- Meetings spend more time disputing the number than deciding what to do.
- Terms like active customer or qualified lead mean different things across teams.
- The board pack lists many KPIs with no order of importance.
How we work
Build the metric tree
We start from how the business makes money, choose a north star that reflects value delivered to customers, and break it into input metrics a team can move.
Pair leading and lagging
Revenue arrives late. We pair it with leading inputs you can manage weekly, test those links against your own history, and flag the ones that remain assumptions.
Write the metric spec
Each KPI gets one page: formula, inclusions and exclusions, date logic, grain, source table, owner and refresh cadence.
Add guardrails and cadence
Every target gets a counter-metric that exposes gaming: lead volume sits beside sales acceptance, delivery speed beside refunds. A monthly review changes metrics that drive the wrong behavior.
What we deliver
- Metric tree diagram
- One-page KPI specs
- Team ownership matrix
- Monthly KPI review agenda
How we measure
- Spec coverage: an agreed definition behind every leadership KPI ends repeat arguments.
- Decisions per review: a KPI meeting that produces no action isn't doing its job.
- Guardrail breaches: shows what a target is costing you elsewhere.
A wholesaler moving to online ordering
Illustrative scenario: Think of a regional wholesaler moving trade customers onto an ordering portal. Sales is paid on revenue, the e-commerce team on sign-ups and service on ticket closures, so nobody owns repeat ordering. A north star of active ordering accounts would give all three a shared target, with service contacts per order as the guardrail.
Common questions
How many KPIs should leadership track?
There's no correct count. A useful test: if an executive can't say what they'd do when a number drops, it's a monitoring metric and belongs on a team view.
Does every company need a north star metric?
It suits subscription and repeat-use businesses. If you win a handful of large contracts a year, one number can mislead, and two or three balanced metrics work better.
Where do target numbers come from?
Your own baseline, seasonality and capacity. With thin history, we use the first period to set a baseline. We're wary of industry benchmarks, since their definitions rarely match yours.
Before we work together, let's think together.
Requests are answered by the founder.