The Hiring Plan Is the Budget: Headcount Planning for Early-Stage Startups
The first thing I ask a new client for is the budget, and the first thing I do with it is sort the rows by size. The top rows are always people. Everything below them, added together, is usually smaller than one senior engineer. Between seed and Series A, headcount is most of the spend and nearly all of the forecast error, and a budget that treats salaries as one line among forty is forecasting the wrong thing carefully. The budget that actually runs the company is the hiring plan.
The sorted budget looks the same at nearly every company I walk into at this stage. Six or eight rows of people at the top. Somewhere around row nine the amounts drop by an order of magnitude — hosting, software seats, legal, insurance, travel — and below that a long tail of numbers small enough that getting every one of them wrong by half would matter less than one engineer starting a month early.
This is not a criticism of the founders who build these budgets. It is what a budget template does when it is applied to a company whose only real input is people. A template designed for a business with inventory, a lease and a marketing line treats salaries as one row among many. In a software company with twelve employees, that row is the company, and the hours spent forecasting the rest are hours not spent on the only forecast that moves cash: who starts when, at what fully loaded cost, and on what condition.
Where the forecast error actually lives
When I reconcile actuals against plan at an early-stage company, the variance almost never comes from the rows the founder spent time on. Software came in a little high, travel a little low; both are noise. The variance comes from three places, and all three sit inside the hiring plan.
The first is timing. A hire budgeted for March starts in June, because the search took eleven weeks, the candidate had a notice period, and the offer was reworked once. Three months of salary not spent is good news for cash and bad news for the plan, because the work that person was supposed to do has not happened either, and the revenue line that assumed it had is now wrong too. The reverse happens as well: a competitor makes an offer to the same candidate, and the "maybe next quarter" hire lands next week. I have watched a single pulled-forward hire take a month off a company's runway — more than every non-people row in that budget could have moved it combined.
The second is cost. The budget had the base salary. Reality had employer-side social contributions, which run from under ten percent of base in some jurisdictions to more than thirty in others and are not optional. It had health insurance, a laptop, a recruiter's fee at fifteen to twenty-five percent of first-year salary — or the founder's own weeks of interviewing, which cost more and appear nowhere — a seat in every tool the company pays for per head, and the time of the two people who onboarded the newcomer instead of doing their own work. Added up, a person costs between 1.25 and 1.5 times their base salary depending on where they sit, before the option grant, which costs no cash and a permanent slice of the company. A budget built on base salaries is understated by a quarter or more on the one line that matters.
The third is ramp. A salesperson budgeted at full quota from month one. A senior engineer expected to ship in week two. A head of marketing whose spend entered the budget on their start date but whose pipeline did not. Everyone knows people ramp; almost nobody puts the ramp in the plan, so the plan assumes each hire is productive from the day their payroll begins. Sales ramp alone — three to six months for a competent account executive selling to businesses — is the difference between a plan that describes the coming year and one that describes a fictional company.
Put the three together and the picture is simple. A fifteen-person company planning to reach twenty-five in a year has ten start dates, ten loaded costs and ten ramps in its plan, and every one of them is an estimate. If each hire lands a month late, the year-end cash position moves by a month of the incremental payroll; a month early, and it moves the other way by the same amount. No row in the non-people budget can shift cash by that much through ordinary variance.
A hire is a trigger, not a date
The change that fixes most of this is small on the page and large in practice. Every planned hire gets two things instead of one: a start month, and a trigger — a condition that has to be true before the offer goes out. The second account executive is hired when the first has been at eighty percent of quota for two consecutive months, not in the third quarter. The customer success hire is made when active accounts pass a number, or founder time on support exceeds a number of hours a week. The third engineer is hired when a specific roadmap commitment is funded and a customer is waiting for it, not because the plan said "engineer" next to a month.
The trigger turns the hiring plan from a list of intentions into a set of decisions the company has already taken, waiting for evidence, and that changes how the plan behaves under stress. When the month arrives and the condition is not met, the hire waits, and nobody has to have a difficult conversation about "cutting" something, because nothing was ever committed. When the condition is met early, the hire moves forward with the justification already written down. And when a founder wants to hire outside the plan — which happens constantly, and is often right — the question is no longer "is this in the budget" but "what is the trigger, and has it fired."
A start date in a budget is a hope with a month attached. A trigger is a decision the company has already taken, waiting for the evidence to arrive.
Triggers also produce the two scenarios that matter, without any additional modelling. The case where no trigger fires is the floor: committed payroll plus committed non-people cost, the burn the company cannot get below without letting someone go. The case where every trigger fires on schedule is the plan. The distance between the two is the company's real operating range — a more honest version of the scenario analysis in the post on runway as a model, because it is built from decisions the company has pre-agreed rather than from percentages applied to a total.
The order of hires is the strategy
Read a hiring plan in sequence and it tells you more about what the founders actually believe than the deck does. If the first three hires after the round are salespeople, the founders believe the product is ready and the constraint is distribution. If they are engineers, the founders believe it is not ready yet, whatever slide four says. Investors read hiring plans this way in diligence. Founders should read their own the same way first, because the sequence is a statement about the bottleneck, and the deck ought to agree with it.
This is also where hiring ahead of revenue gets its honest answer. It is right when the revenue is not the thing the hire is supposed to prove. A founder who has closed the first fifteen customers personally and cannot physically run more conversations is hiring a salesperson to add capacity to a motion that works. A founder who has closed three and is hiring a salesperson to find out whether the product sells is paying a year of loaded salary for an experiment they should have run themselves in a quarter. "When the founder is turning down qualified conversations" is a trigger. "When we have raised" is not.
What the plan looks like on one page
The plan I build with clients is a single table, and the discipline is keeping it to one page. Each row is a role rather than a person until an offer is signed, and the columns are the ones that drive cash and nothing else.
One row per planned role: title and function; planned start month; the trigger, written as a condition that can be measured; fully loaded monthly cost — base, employer contributions, benefits, equipment, tools and recruiting; ramp, in months to full productivity; the owner of the search; and a status of open, offer out, started or deferred. Three totals at the bottom: committed payroll for people who have started or signed, planned payroll under triggers, and the gap between them. Reviewed monthly alongside the runway model, with each trigger tested against that month's actual numbers, not the plan's.
Rows are roles because a plan written in names changes when a candidate says no, and candidates say no. Costs are monthly and fully loaded because the annual figure hides the start date and the base figure hides a quarter of the cost. A plan with those two properties drops straight into a cash bridge as one line per role, which is where it belongs.
Where hiring plans go wrong
The failure modes are consistent enough to name. The most common is the plan that is really a fundraising narrative: twelve hires, all starting the month after the round closes, none with a trigger, because the plan was written to justify the amount being raised rather than to run the company. Investors see through it, and — worse — founders sometimes come to believe it once the money arrives.
The second is ignoring the slip. Recruiting a senior role takes two to four months from decision to start date, longer in markets with statutory notice periods. A January hire that is decided in January starts in April. The honest cash forecast is the one in which hires start a quarter after the founder thinks they will.
The third is hiring to the calendar rather than to the trigger. It is the third quarter, the plan says a designer, so a designer is hired into a company whose product has not reached the point where a designer changes the outcome. The trigger existed to prevent exactly this and was overruled by the date beside it. A plan is not a promise to hire; it is a promise to hire when a condition is met.
The last is the hire that is really a founder stepping away from something. A head of finance at eight people, so the founder does not have to learn how the numbers work. A head of sales at the first difficult conversation with a customer. Some of these are right. Most, at seed, are expensive ways of avoiding the part of the job the founder needs to understand before they can manage the person who will eventually do it — I have written about when a startup actually needs a CFO, and the answer is later than the founder's discomfort suggests.
The hiring plan sizes the round
The final reason the hiring plan is the budget: it is the document the fundraise is built from, whether or not the founder knows it. "We are raising two million to reach this milestone" is a statement about a hiring plan. The two million is the loaded cost of the people needed to reach the milestone, for the months it takes, plus committed payroll, plus non-people cost, plus a buffer for the slip and for the year in which two triggers fire early. Founders who arrive at the number the other way round — pick the round size the market seems to be paying, then build a plan that spends it — end up with a narrative instead of a plan, and a diligence process that finds this out.
The diligence question is always some version of "walk me through the hires." The founder who can go row by row — this role, this month, this trigger, this cost, this is what it unlocks — is making the strongest case for the round without a single slide. The founder who says "eight engineers and three salespeople" and cannot say what the fourth engineer does that the third could not is telling the investor that the plan was written for them.
The budget I actually build for a seed-stage company is two pages long. The first is the hiring plan, with triggers. The second is the list of committed non-people costs, which changes rarely and can be reforecast in an afternoon. Everything else — the forty rows, the software line to the last euro — is a monthly reforecast from actuals, and nobody should spend a week on it. The founder who can defend every row of the hiring plan with a trigger rather than a date has done the only budgeting that matters at this stage. The rest is rows.