Stop tracking everything. Pick six metrics — throughput, cycle time, on-time rate, revenue per person, utilization and client health — and review them once a week, not once a day.
Open most team dashboards and you get buried fast. Tasks created, tasks completed, hours logged, messages sent, a dozen colorful charts — a wall of numbers with no priority. Most people glance at the busiest one, feel good or bad for a second, and close the tab. That's a wasted habit.
The problem isn't that you're measuring. It's that you're measuring indiscriminately. A small team doesn't need an analytics department — it needs a handful of signals that reliably answer one question: is the work flowing the way it should? Below are the six I build into every operating system, and exactly how to read each one.
1 · Throughput
Throughput is the number of meaningful units of work your team finishes in a period — shipped features, closed tickets, delivered projects. It's deliberately blunt. You're not measuring effort, you're measuring output that crossed the finish line.
Track it weekly and look at the trend, not the absolute number. A throughput that wobbles wildly week to week usually means your intake is chaotic — work is arriving in unpredictable lumps rather than a steady flow.
If you can't say this week's throughput number from memory, it isn't visible enough. Put it on the first screen of your workspace.
2 · Cycle time
Cycle time is how long a unit of work takes from "started" to "done." It's the single best early-warning signal you have. Throughput can look healthy while cycle time quietly creeps up — and a rising cycle time always precedes missed deadlines.
Measure it from the moment real work begins, not from when the request first landed. Mixing queue time and work time together gives you a number nobody can act on.
3 · On-time rate
Of everything you committed to this period, what share actually landed when you said it would? On-time rate is the metric clients feel most directly, and the one teams are most tempted to fudge.
Keep it honest by recording the commitment date once — when you make the promise — and never editing it afterward. The gap between intent and reality is exactly the thing worth seeing.
“You don't need more data. You need the same six numbers, reviewed at the same time, every week.”
4 · Revenue per person
For any team running lean, revenue per person is the clearest measure of whether your systems are actually creating leverage. Headcount is easy to grow; revenue per head is the number that tells you the growth is healthy.
You don't need to share it with the whole team, but the founder should watch it monthly. When it climbs as you add systems rather than people, your operating system is doing its job.
5 · Utilization
Utilization is the share of available capacity that's committed to real work. The trap is treating 100% as the goal — a team running flat-out has no slack to absorb surprises, and surprises are guaranteed.
Aim for a band, not a ceiling. Somewhere around 70–85% leaves room for the unplanned without letting capacity go to waste.
6 · Client health
The softest metric, and often the most predictive. A simple green / amber / red status per client — updated by whoever owns the relationship — surfaces churn risk weeks before it shows up in revenue.
Keep the criteria explicit so "amber" means the same thing to everyone. Ambiguous status fields are worse than no status at all.
The 10-minute weekly review
Metrics only matter if you look at them on a rhythm. Block ten minutes every Monday, open one dashboard, and walk the six numbers in order. Ask a single question of each: is this better, worse, or flat versus last week — and do I know why?
That's the entire discipline. Six signals, one screen, ten minutes. Do it for a month and you'll catch problems while they're still small — which is the only time they're cheap to fix.