ingridleiria Annual Planning and Headcount
The plan and the headcount plan are usually built by different people, in different tools, a fortnight apart, and they contradict each other from the day they are signed. The plan assumes a sales team that arrives in month two and is productive immediately. The budget assumes requisitions opening in month five. Neither assumption is written down anywhere the other author would see it, so the contradiction is invisible until the third quarter, when the revenue gap is too large to close and the hires who would have closed it have not started.
The cost is a year of quiet under-delivery that nobody can attribute, followed by a hiring freeze announced as a surprise. The freeze is the expensive part: it lands on people who joined three months earlier on the strength of a growth story, and it costs the company its credibility with exactly the group it just paid to recruit. A pause trigger written in advance is a policy. The same decision taken in the moment is a crisis.
When to use this, and when not to
Use it to run a full annual cycle, a quarterly re-plan, a budget build or rebuild, a hiring plan for the year, a mid-year reforecast after a material miss, or the affordability check on a plan somebody else has written.
Do not use it to decide direction. That is strategic-plan-and-action-plan, which sets the destination and the bets, and which must exist first, because the whole reconciliation step depends on knowing which two or three things get funded properly when the numbers do not add up. Do not use it to write the measured objectives for the period; that is okr-planning, which runs after the plan is approved and turns it into things that can be scored. Do not use it to design the meetings that review the plan through the year; that is operating-cadence-design. Do not use it to build the revenue model itself, which is revenue-forecast and financial-model-builder, or to define the roles and run the interviews, which is hiring-scorecard-and-interview-kit. Do not use it to write the board narrative; that is board-and-investor-management, which takes this plan as its input.
The boundary: this skill answers what the plan costs, who does it, when they arrive, and whether the company can afford to find out.
What you need before starting
The bets and the destination. Without them the reconciliation has no basis and the gap gets closed by whoever argues hardest. Missing: stop and get three sentences from whoever decides. This is the input the method turns on.
Current headcount with cost, start date and function. Missing: reconstruct from payroll and note anyone hired in the last quarter separately, since they are the group most often missing from a spreadsheet and most likely to be mid-ramp.
The revenue plan or forecast, with its assumptions visible, in particular what productive sales capacity it assumes and when. Missing: build against a range and say plainly that the headcount sequence cannot be finalised until the revenue assumption is fixed.
Cash position and current monthly burn. Missing: get it before anything else. A plan built without a cash number is an essay.
Time to hire by role and market, and attrition over the last twelve months. Both are usually available and both are usually left out. Missing: use conservative defaults, say what you assumed, and mark them for correction. A plan that omits attrition overstates every year.
Fully loaded cost assumptions and comp bands by level and location. Missing: use salary plus a stated uplift for taxes, benefits, equipment, software and recruiting, commonly a quarter to forty percent above salary depending on the country, and label it an assumption to be replaced.
The approval calendar: board dates, budget sign-off, and when the company is told. Missing: propose the dates and get them confirmed before building starts, because planning fails on timing more often than on content.
The method
Fix the calendar first, working backwards from the approval date, with a named owner per step. A workable annual shape: strategy settled and communicated eight weeks out, functional builds due five weeks out, reconciliation three weeks out, leadership decision two weeks out, board approval one week out, communication to the company in the first week of the period. A quarterly re-plan is the same shape compressed into two weeks, touching only what changed. Without an owner per step, functions submit spreadsheets in different shapes that cannot be added up.
Set the top-down envelope and publish it before the bottom-up build starts. The revenue target, the margin or runway constraint, and the resulting spend envelope: a small number of numbers set by the leadership team, so functions are not planning into a vacuum.
Issue one bottom-up template and require it. Each function states what it will deliver, what it needs, and what it stops doing, in the same columns as everyone else. Return non-conforming submissions the same day rather than reformatting them yourself, which is how a planning owner loses a week.
Reconcile the gap by revisiting the bets, never by cutting proportionally. The two builds will not match, and that gap is the actual work of the cycle. A proportional cut is the default and the worst option, because it underfunds the two or three things that have to work while protecting everything else at a level where nothing is done properly. The rule: fund the bets at the level they need and take the shortfall from what the strategy already named as not a priority. Where that is not enough, the destination has to change, and that is a decision to escalate rather than absorb.
Record every trade-off with its reason, at the moment it is made. In eight months somebody will ask why a team is under-resourced, and the answer should be a decision on record rather than an oversight. It takes ten minutes during the cycle and is unreconstructable afterwards.
Sequence headcount by the month each hire produces output. For every role: the outcome it exists to produce, the requisition month, realistic time to hire, the ramp, fully loaded cost, and what happens to the plan if it is not filled. The month that matters is the start date plus the ramp. A senior sales hire starting in month four with a six-month ramp contributes nothing to that year's number.
Check the three things people skip: span of control, the manager to individual contributor ratio, and backfill. A management plan with three reports each is a plan to build layers. The manager ratio drifts upward silently, one reasonable promotion at a time. Attrition is a real line, and leaving it out means every replacement consumes a slot the plan believed was growth.
Build the cost curve by month, not as an annual total. Twenty hires spread evenly and twenty in the first quarter cost very different amounts in the same year, and the second changes runway materially.
State runway under the plan and under two alternatives: revenue landing materially below plan, and hiring delayed by a quarter. Runway is the number the board asks about first, and it should never be the number you compute in the meeting.
Write the hiring pause trigger before anyone is optimistic. Below a stated revenue level by a stated date, or above a stated burn multiple, hiring pauses except for roles named now, individually. Get it approved with the plan so that invoking it later is administration rather than a debate.
Publish it three ways with the same numbers: a short narrative saying what the company is doing this year and what it is not, which is the only part most people read; the objectives with owners; and each function's own version. Where the three disagree, the narrative wins and the others get corrected.
Fully loaded cost, ramp, and the output month
Three quantities are routinely conflated, and separating them is most of the accuracy in a headcount plan.
Fully loaded cost is salary plus employer taxes, benefits, equipment, software licences and recruiting cost, commonly a quarter to forty percent above salary depending on the country. Using base salary alone understates a twenty-person plan by roughly the cost of five more people.
Time to hire is from requisition open to signed offer, not to first day. Add the notice period, which is two weeks in some markets and three months in others, and which is the most common omission in a plan built by a team that has only hired locally.
Ramp is from first day to full productivity, and it varies enough to be worth stating per role rather than averaging: a support engineer four to six weeks, an enterprise account executive six to nine months, an engineer on a complex codebase three to four months. The output month is requisition month plus time to hire plus notice plus ramp.
Where a workforce or applicant tracking system is connected, pull actual time to hire and attrition from it; where it is not, use the recruiting team's last twelve months of offers; where neither exists, state the assumption and mark it for correction at the first quarterly review.
Worked example
Situation. Northbank Software, 180 people, revenue 42m, planning a year with a target of 58m and an intention to reach 260 people. Cash 31m, monthly burn 1.6m. Two bets: move upmarket to larger accounts, and reduce implementation time so services stop capping growth. Board approval in seven weeks.
Task. An approved plan and headcount sequence, runway shown under a downside, and a hiring pause trigger agreed in advance rather than improvised in the second half.
Action. The calendar was fixed first, backwards from the board date, with an owner per step. Functional builds came in at five weeks out and asked for 94 hires against an envelope that supported 61, a gap of roughly 7.4m annualised.
The first reconciliation was a proportional cut of about 12 percent across every function, and it balanced on paper in an afternoon. It was abandoned two days later when tested against the bets: at 12 percent off, implementation gained four people instead of nine, which left implementation time where it was, which meant the upmarket bet could not be delivered. The cut had protected eight functions equally and defunded the two that mattered.
The second reconciliation went back to the strategy. Implementation was funded at nine, the full ask, and enterprise sales at seven of a requested nine. The shortfall came from two places the strategy had already named as not priorities: the self-serve growth team held at current size rather than adding five, and a planned expansion of partnerships was deferred a year. Both were recorded with the reason and the date, and both leads were told before the plan circulated rather than reading it in the document.
Sequencing then exposed a second problem. Working the output month for each of the seven enterprise hires, using 11 weeks to hire, a 6 week average notice period and a 7 month ramp, showed that four contributed nothing in the year and the other three contributed roughly a third of a year each. Productive capacity was 38 percent below what the revenue plan had assumed. Two honest options went to the leadership team as a choice: pull four requisitions into the first quarter, raising burn early and shortening runway by about six weeks, or reduce the target to 53m. They pulled three forward and reduced the target to 55m, which was the useful outcome of the cycle, and it happened in week four rather than in the third quarter.
Runway was 19 months under plan, 14 months under a downside where revenue landed at 48m, and 21 months where hiring ran a quarter late, which made the point that late hiring is not automatically bad news for cash. The pause trigger was agreed with the plan: if revenue through 30 June is below 24m, or burn exceeds 2.1m in any two consecutive months, all requisitions pause except four named roles listed in the appendix.
Result. Approved with one change, a request to name the critical roles more narrowly, done in a day. Revenue through June came in at 24.8m, above the trigger, so hiring continued. Median implementation time fell from 71 days to 44 by the end of the third quarter. Six roles went unfilled, five in engineering, and because each carried a stated consequence the conversation about what slipped took one meeting rather than three.
A second scenario, where it goes differently
A 130-person professional services business where headcount is revenue rather than cost and the constraint is utilisation.
The method inverts. Hiring is sequenced against booked and probable backlog rather than ambition, and the pacing question is not affordability but bench: hire too early and utilisation and margin fall, hire too late and you turn work away. Utilisation targets by level replace the productivity ramp as the key assumption, and the cost curve becomes a margin curve. The pause trigger becomes a utilisation floor rather than a revenue threshold: if billable utilisation falls below a stated percentage for two consecutive months, hiring pauses except for roles committed against a signed statement of work. The plan also carries a trigger in the other direction, a utilisation ceiling above which hiring accelerates, because in a services business being late is as expensive as being early.
What did not change: the calendar with named owners, reconciliation through the bets rather than proportionally, fully loaded cost, and a trigger written in advance.
Output
The plan on a page, then three tables and a narrative.
| Function | Bottom-up ask (headcount, cost) | Funded (headcount, cost) | Difference | Trade-off made, and why | Consequence if unfilled |
|---|
The headcount sequence, one row per role:
| Role | Function | Outcome it produces | Req opens | Time to hire | Notice | Ramp | Output month | Fully loaded cost | Consequence if unfilled |
|---|
Cost and runway:
| Month | Headcount | Payroll cost | Other opex | Revenue | Net burn | Cash | Runway (months) |
|---|
repeated for the downside case and the delayed-hiring case. Where charted, keep it monochrome, separate the three scenarios by dash pattern and marker rather than colour, put the legend outside the plot area, and reserve one accent for the month cash crosses the threshold.
The pause trigger, as a fenced statement approved with the plan:
HIRING PAUSE TRIGGER
Condition: [revenue below X by date, or burn above Y for N months]
Effect: All open requisitions pause on the date the condition is met.
Exceptions: [named roles, listed individually, not "critical roles"]
Who confirms: [one name] Reviewed at: [cadence]
Restart: [the condition under which hiring resumes]
Failure modes
The plan and the headcount plan built separately. Recognise it when the revenue plan cannot say which hires it assumes. Build the output month for the revenue-generating roles and compare it against the revenue curve; the gap is the finding.
The proportional cut. Recognise it when every function lost a similar percentage. Reopen the reconciliation against the bets and be prepared for the conversation to be harder, because a proportional cut is popular precisely because it avoids choosing.
Start date used as output date. Recognise it because no ramp column exists. Add time to hire, notice and ramp per role, and expect the in-year contribution to fall by a third or more.
No attrition line. Recognise it when planned headcount equals starting headcount plus hires. Add expected leavers by function using last year's actual rate, and show backfills separately from growth hires.
Trade-offs made and not recorded. Recognise it when nobody can say why a team is small. Write the reason in the same session; it cannot be reconstructed later.
A pause trigger written as sentiment. Recognise it when the condition has no number, no date, or no named confirmer, or when the exception list says "critical roles" rather than naming them. Fix both before approval.
Edge cases
No revenue plan yet, or one still moving. Build the sequence against a range and present the cost curve for the low and high cases. Do not wait; the sequencing work is valid across both, and the decision to be made is which case to staff for.
Cash is short enough that the plan is a runway plan. Invert the order: start from the cash-out date, work back to an affordable monthly burn, and let that set the envelope before any function builds. The bets shrink to one or two and the pause trigger becomes the primary artefact.
An acquisition or a large contract lands mid-cycle. Re-run the envelope, the reconciliation and the cost curve only. Do not rebuild everything, and do record what changed, because an unmarked amendment makes year-end variance analysis impossible.
Multiple currencies or locations. Model in the currency each cost is incurred in, convert once at a stated rate, and state the rate. Loaded cost multipliers differ enough by country that a single blended uplift will misprice a distributed plan.
Quality bar
- The calendar is fixed, with a named owner per step, before any building starts.
- Both directions are built, and the gap is closed by revisiting the bets, never by a proportional cut.
- Every trade-off is recorded with its reason and the date it was made.
- Every role has a requisition month, time to hire, notice, ramp, an output month, a fully loaded cost, and a stated consequence if unfilled.
- Attrition and backfill appear as their own lines, separate from growth hires.
- Cost is modelled by month, and runway is shown under plan, under a revenue downside, and under delayed hiring.
- The hiring pause trigger has a number, a date, a named confirmer, and an exception list of named roles.
- The narrative, the objectives and each function's version carry the same numbers.
Adapting this to your context
The numbers here come from venture-funded software companies of a hundred to three hundred people planning against a cash runway. They are defaults, not standards.
- The loaded cost uplift. A quarter to forty percent above salary assumes commercial employment in a high-tax market. Rebuild it from your last four hires; universities and public bodies are usually given a fixed overhead rate instead and should use theirs.
- The ramp figures. Six to nine months for an enterprise seller, three to four for an engineer. Use the time your last five hires in that role took to reach standard, and where nobody measured it, say so on the page.
- The eight-week calendar. It assumes a board approval date. Where approval sits with a partnership, a trustee board or a parent company, work back from their committee date, which is often fixed a year ahead and cannot move for you.
- The runway constraint. A cash-out date suits a company burning money. Substitute the covenant, the grant period end or the reserves policy, and run the same three cases against that.
- What not to change. Close the gap by revisiting the bets rather than cutting proportionally, and sequence every hire by its output month rather than its start date.
Related skills
strategic-plan-and-action-plan sets the destination and the bets that this plan prices, and must come first. okr-planning runs after approval and converts the funded plan into measured objectives. operating-cadence-design places the reconciliation, the approval and the quarterly re-plan into the company's rhythm, and weekly-status-update and meeting-to-decisions are the artefacts that rhythm produces during the cycle. revenue-forecast and financial-model-builder supply the revenue side of the envelope. hiring-scorecard-and-interview-kit turns an approved role into a definition and a process. board-and-investor-management presents the approved plan and the runway cases to the board, and decision-memo carries any single trade-off that has to be escalated rather than absorbed.