The Headcount Plan

Most lines in a model are abstractions. Revenue is a curve, opex is a category, depreciation is a schedule. The headcount plan is the one where the cells have names: row 47 isn’t a number, it’s the person two desks down, the hire you haven’t made yet, the team you’re about to ask to do more with the same six people.

That’s what makes it unlike anything else in the model. It’s the most financially consequential plan you’ll build all year, because in most businesses people are the single largest cost on the P&L. It’s also the one plan where getting the arithmetic right and getting the decision right aren’t the same thing.

So it deserves more care than a line its size would normally get, not less because it’s “soft,” but more, because it’s both.

Why it’s the heaviest line in the model

Start with the money, since that’s why finance is in the room.

For most companies, labor is the largest single cost they carry. The U.S. Bureau of Labor Statistics puts it at up to 70% of total operating costs in labor-intensive industries; in services, payroll routinely runs 30–50% of revenue, higher in the most people-dependent models: agencies, professional services, care, anything where the product is the people. Even in capital-heavy businesses, compensation is rarely far from the top of the stack.

That weight has a practical consequence: headcount is the line with the most leverage. A 5% miss on a line that’s 8% of cost is a rounding error. A 5% miss on the line that’s half your cost base shows up in operating margin, runway, and covenant headroom. When a plan needs to find money, headcount is where the money is, which is why the hardest conversations in budget season are usually conversations about people, whatever else they claim to be about.

Because it’s the biggest line, it’s the one most worth modeling properly: not a salary total dragged across twelve months, but a roster of real positions, each with a start date, a cost well above its salary, and a way of changing over time the other lines don’t have.

Why it’s the most fragile

This is where the headcount plan stops behaving like the rest of the model.

A revenue line can be wrong and you re-forecast it. A cost assumption can run high and you trim it. The headcount plan has a property no other schedule shares: its numbers correspond to decisions about people’s lives. A cell that reads “−3” in Q3 is a layoff. A start date that slips from March to June is an offer rescinded, or a team left to carry the gap. A backfill quietly dropped to make the math work is a promotion that won’t happen.

That’s the fragility, and it cuts both ways:

So the discipline isn’t “be lean” or “be generous.” It’s being clear about what the cells mean. A good headcount plan keeps the human decisions visible as decisions: when a role is held open or a team is asked to stretch, someone chose it on purpose, rather than it falling out of a formula. The arithmetic serves the decision; it doesn’t get to make it.

Headcount is the one schedule where a clean number can hide a hard decision. Every cell is someone’s livelihood: the job is to keep that visible, not to abstract it away until it’s comfortable.

What the plan actually is

Under the weight and the fragility, the mechanics are concrete. A headcount plan is a roster that becomes a cost: two halves joined by time.

The first half is the roster: the positions, filled and open, the organization intends to carry. Who’s here, who’s leaving, who’s being hired, and when. A position is a row with a start date, an end date if it has one, a team, a level, and a status.

The second half is the cost: each position carries far more than the salary on the offer letter: taxes, benefits, bonus, the employer’s share of everything, the desk and the laptop and the software seat. (How much more, and how to build that number honestly, is the fully-loaded cost, its own subject, which the next articles take apart.)

What makes it a model rather than a list is what happens between the two over time:

• People start mid-period, so a hire costs only a fraction of the year it lands in, and start dates slip, the most reliable source of savings nobody planned.
• People leave, on a curve you can’t name in advance but can model in aggregate.
• Pay rises on a cycle: the annual merit wave that lifts the whole base at once.
• The roster counts in more than one unit: heads, FTEs, and hours are three different numbers, and mixing them up is how finance’s total and HR’s never match.

Each of those gets its own article in this section. The opener’s point is just this: the headcount plan is where org design, timing, and cost meet, and it’s hard because all three move at once, and one of them is people.

In Novi, the roster and the P&L are the same model

In a spreadsheet, the headcount plan lives a double life. There’s the roster: a tab of names, dates, and levels someone maintains by hand. And there’s the cost: a second tab of formulas that turns the roster into monthly numbers and feeds the P&L. The two are linked by references that break the moment someone inserts a row, and reconciled by whoever has the patience to check that tab one still matches tab two. They drift. The headcount reconciliation is a recurring chore precisely because the spreadsheet keeps the people and the money in separate places and dares you to keep them in sync.

In Novi, the roster and the cost are one structure, not two tabs. A position is an element with a start date, a level, and a team; its fully-loaded cost is derived from that element by formula, like any other calculated line. Change the start date and the cost moves to the right month. Add a hire and it flows into the P&L with no reference to repoint. Slice the same plan by team, location, or scenario (a hiring freeze beside a ramp plan) without rebuilding the roster three times. The count and the cost can’t disagree, because they come from the same place.

That matters most for the part the spreadsheet drops: keeping the human decisions visible. When the roster is the model, a held-open role is a position you can see, with a status and an owner, not a number missing from a salary total. The plan stays honest about what its cells mean, which, for the one plan where the cells have names, is the whole job.