Personal KPIs: measuring the operator

Personal KPIs are a small set of numbers — five or fewer — that tell you whether you, the operator, are functioning, in the same way a dashboard tells you whether the business is. The useful ones measure controllable inputs (deep-work hours, decisions closed, outreach sent), not moods or outcomes, and they get reviewed weekly, not obsessed over daily. Most founders measure their business to one decimal place and measure themselves not at all.

Why measure yourself at all?

Because you are the most expensive and least monitored asset in the company. A five-person firm will track pipeline value, response times and utilisation, while the founder — whose hours drive all of it — runs on vibes and a vague sense of being busy. When the constraint on growth is founder attention, which in small service firms it usually is, an unmeasured founder is an unmeasured bottleneck.

Measurement also corrects the specific self-deception operators are prone to: confusing activity with progress. A full calendar and a hot inbox feel like output. Only a number — how many hours went to work that compounds, how many decisions actually closed — can tell you whether the feeling is earned. This is the founder-level version of the argument in the UK B2B outbound playbook: a pipeline you do not measure is a pipeline you cannot fix, and the same is true of a week.

What makes a KPI personal rather than performative?

Three tests. It must be an input you control — "deep-work hours" qualifies, "revenue" does not, because revenue lags your behaviour by months and punishes you for things that are not behaviour. It must be countable without ceremony — if measuring it takes more than a minute a day, it will die within a fortnight. And it must change a decision — if no reading of the number would alter what you do next week, it is decoration.

This is where most personal-metrics attempts fail: people import vanity metrics from productivity culture. Streaks, app scores, screen-time percentages. These measure compliance with the tool, not function of the operator.

Which KPIs do I actually track?

Mine, as one worked example rather than a prescription:

• Deep-work hours per week. Uninterrupted blocks on work that compounds. My floor is a number I defend in the calendar, not an aspiration.
• Decisions closed per week, plus the age of the oldest open decision. Stale decisions are the quietest form of drag I know, which is why they pair with a 30-minute research cap.
• Outbound actions sent. Proposals, follow-ups, cold approaches — the inputs to future revenue, counted weekly.
• Days shut down on time. A proxy for sustainability; when it drops for three straight weeks, something structural is wrong.
• One subjective score for energy, 1–5, noted daily in five seconds. Not a KPI in the strict sense, but the correlation between it and the other four is where the insight lives.

Everything sits inside the personal operating system; the KPIs are its gauge cluster, not a separate hobby.

How do you run the measurement without it becoming a job?

The mechanism is deliberately small. When the working day ends, then you spend sixty seconds logging the day's counts — a text file or a spreadsheet row, nothing more elaborate. When the week ends, then the weekly review compares the numbers to their floors, and any KPI below floor gets one named cause and one adjustment for next week — an adjustment to the system, not a promise to try harder. When a KPI has been comfortably above floor for a quarter, then it either gets retired or its floor rises; a metric nobody can miss teaches nothing. When a metric starts driving silly behaviour — logging shallow work as deep to protect the number — then the metric gets rewritten, because a gamed KPI is worse than none.

That last rule matters. Goodhart's law — when a measure becomes a target, it stops measuring — applies to individuals just as forcefully as to sales teams. The defence is reviewing the numbers weekly, in context, rather than chasing them daily.

What do the numbers change in practice?

They convert arguments with yourself into observations. "I have no time for deep work" becomes "I averaged four hours a week last month, and the leak is Tuesday afternoons". "I'm indecisive" becomes "my oldest open decision is 26 days old and it's about a £40 tool". Named problems get fixed; ambient guilt does not.

They also tell you when the problem is not effort at all. When deep-work hours are fine and output still feels thin, the issue is usually interruption structure — the subject of deep work when clients want you shallow — rather than anything a metric can push on. The KPI's job is to rule things out as much as in.

Measure the operator the way you would measure the business: a few honest inputs, reviewed on a cadence, connected to decisions. Everything else is scenery.

---

Keep the thinking coming: The Format — a fortnightly letter on business and personal systems. Subscribe by email.