The Operator Output Calculation

A hiring-math framework for 5–200 person businesses.

The decision this is for

An operator running a 5–200 person business has roughly four to seven hours per week to spend on the operating system itself — meetings, planning, scoring, priority review. The remaining thirty hours are split between customer work, hiring, and escalations.

Every hire, retainer, and tool the operator approves is implicitly a bet that the next unit of back-office spend produces more output per dollar than the last one. Most operators make that bet on instinct. The instinct is usually wrong, because the operator is the most expensive resource in the company and that cost does not appear on any income statement.

This calculation puts the operator's attention onto the same balance sheet as the spend, so the bet can be made with a real number.

The five inputs

• Output Volume (V). Completed units the function produces per period — articles published, qualified leads, candidates screened, tickets resolved. Pick the smallest unit that maps cleanly to a business outcome.
• Cycle Time (T). Elapsed days from request to ship. A blog post that takes six elapsed days has T = 6, not the two work-hours.
• Quality Variance (Q). Standard deviation of unit quality against the function's quality bar. Most operators omit this. It is the input that distinguishes a leveraged function from one that just looks busy.
• Operator Attention Cost (A). Hours per week the function consumes from the operator personally — including the background processing the function imposes on working memory. A function that takes 45 minutes of meetings but lives in your head all week has higher A than the meeting time alone.
• Loaded Spend (S). Total dollar cost, loaded. For a hire: base × 1.4. For a tool: subscription. For a retainer: monthly fee plus operator-side handoff cost.

The ratio

L = (V / T × (1 / Q_norm)) / (S + A_dollarized)

A_dollarized = operator hours × loaded operator hourly rate
               (default: $300/hr for a founder of a 5-200 person business)

Higher L is better. Two functions with very different absolute volumes can be compared. A content function shipping 40 articles a month with high A is directly comparable to one shipping 10 with low A — the ratio tells you which is leveraged.

Worked example: content production

The third configuration is roughly 29× more leveraged than the first, even though the first looks free. Free is the most common pricing illusion in operator software.

Same shape applies to sales pipeline and recruiting — the leveraged configuration is always the one that minimizes A, not the one that minimizes S.

The most common failure modes

Undercounting A. Operators are trained by an industry that values manual founder work — the founder writing the post, making the cold call, doing the screening — as a sign of dedication. Manual founder work has signal value. It does not have leverage. The hour costs $300. If the same post can be produced for a $20 API call against a $199/month subscription with one minute of audit, the founder's hour was the most expensive content in the building.

Undercounting Q. A function shipping at 2× the rate at half the quality is producing the same throughput against the business outcome. Quality variance is what separates a busy function from a leveraged one.

Treating S as a wall. "I cannot afford a $250k CMO" is correct. It is also a category error if the actual decision is between $250k of CMO and $199/month of AI CMO with a one-hour-per-week audit. The two are not the same shape; the calculation should not pretend they are.

The Tuesday morning version

Full calculation is for monthly review. Tuesday version is one question:

"How many hours of my own attention did this function consume in the last week, and is that consumption producing more business outcome than the same hours spent somewhere else would have?"

If the answer is no, the function is not leveraged at the current configuration. Recalculate against an alternative. The recalculation usually points the same direction the worked examples did: the leveraged configuration minimizes A, accepts a higher S in exchange, and produces more output per dollar of total cost than the founder's intuition predicts.

This paper is not an argument for adopting AI executives. It is an argument for measuring leverage with a calculation that includes the operator's attention as a real cost. Operators who run that calculation and find their existing configuration leveraged should keep it. Operators who run it and find what most operators find — that A is dominating S, that L is low, and that the founder's Tuesday is paying the bill — have the data they need to act.

— Everett Steele, founder, Meridian Ventures · April 28 2026