Every team's first scorecard fails the same way: fifteen numbers nobody owns, reviewed monthly, abandoned by week six. A scorecard that works is small, owned, and weekly — 5–8 numbers per team, one name on each, read out loud by that name every single week. This pack gets you there in an afternoon.
How it works
Pick your department tab
Sales, Marketing, Customer Success, Product & Engineering, or Operations & Finance — each with 8 candidate metrics and a one-line case for why each earns a row.
Pick 5–8. Cross out the rest.
A scorecard with fifteen numbers is a report; a scorecard with six is a meeting. If everything is important, nothing is.
Put one name on every number
The owner reads their own number out loud in the weekly review — on target or off. No owner, no metric.
Set your target
Our suggested targets are ~50-person-SaaS starting points. Your baseline beats our suggestion: measure two or three weeks, then set the target just above trend.
Fill one column per week
W1–W13, with a "weeks on target" score that computes automatically and respects direction — higher-is-better and lower-is-better both handled.
Assemble your team scorecard
Copy your picks into the final tab — capped at eight rows on purpose. Need a ninth? Retire one first.
The doctrine baked in
Outcome over activity
If you could hit the number while the business gets worse, it's the wrong number.
Pairs, not solo metrics
Speed pairs with quality, volume pairs with cost — the libraries pre-pair the classic gaming traps.
Weekly or it isn't real
A number you can only read monthly can't steer a weekly meeting. Every metric here reads weekly.
Calibrated for B2B software and recurring-revenue teams. Different industry? The structure — pick 5–8, one owner, direction, weekly grid — works anywhere; swap the metric library for your world's equivalents. Writing the targets themselves? Ten minutes with how to write key results people can't game will save you a quarter of gamed numbers. And when the scorecard is ready to be read out loud, the Weekly Review agenda pack gives the readout its ten timed minutes. Unsure which numbers lead and which merely report? Leading vs. lagging indicators is the ten-minute answer.