Workforce planning math: headcount modelling, attrition forecasting, and hiring lead times

60-Second Summary

A workforce plan is a math model, not a spreadsheet of wishes. The core equation: Ending Headcount = Opening + Hires − Attrition. Everything else is forecasting those three terms.
Annualised attrition is misleading. Always use monthly attrition × 12 and segment by tenure (most exits cluster in months 3–9 and around month 24).
Hiring lead time = time from approval to first day. Median for engineering 2025–2026: 8–14 weeks. Plan backwards from when you need someone productive, not when you'd like to post the role.
Ramp curve: a senior engineer hits ~50% productivity by week 8, ~85% by week 16, full by week 26. A workforce plan that ignores ramp over-counts capacity by 15–25%.
Recruiter capacity: one experienced tech recruiter can sustainably close ~3–5 hires per month. Plan recruiter headcount from hiring plan, not the other way around.

Most workforce plans are wish-lists with a spreadsheet skin. A real plan is a quantitative model that says: given our current headcount, expected attrition, hiring lead times, and ramp curves, here is the realistic headcount trajectory — and here is what we need to spend, in recruiter capacity and management bandwidth, to hit it.

The core equation

The only formula that matters

Ending headcount (period t) = Opening headcount + Hires − Attrition. Run it monthly, segment by function, and never use annual aggregates for planning.

Every workforce plan reduces to forecasting the three right-hand terms. Opening headcount is known. Hires are a function of demand, lead time, and capacity. Attrition is a function of cohort, role, and conditions. Most planning errors are attrition errors.

Forecasting attrition (the right way)

Why annualised rates lie

An 'annual attrition of 18%' is a backward-looking aggregate. It hides the fact that 60% of those exits came from one BU, that 40% happened in months 3–9 of tenure, and that the bump around the 2-year cliff accounted for most of the remainder. Plan with monthly rates segmented by tenure cohort.

Segmentation that actually predicts

1
By tenure band
0–3mo (early-exit / mishire), 3–12mo (onboarding / fit), 12–24mo (the second-year wall), 24–48mo (the equity-cliff effect), 48mo+ (loyal core). Each has a different exit rate.
2
By function
Sales attrition is 1.5–2x engineering. Customer support is 2–3x. Don't apply company-average attrition to any one team.
3
By level
ICs attrit more than managers. Senior+ leaders attrit less in absolute terms but cost more per exit.
4
By manager
If you have ≥ 5 reports under one manager, attrition under that manager is a stronger predictor than role-level rates. Bad-manager teams attrit 2–4x baseline.
5
By location
Remote roles attrit 10–20% more than office in most 2024–2026 studies. Regional differences are larger than function differences.

The monthly attrition formula

Monthly attrition rate

Monthly attrition = exits in month / average headcount in month. Average headcount = (opening + closing) / 2. Annualise by × 12 only for executive comms — never for planning.

Build a 12-month rolling forecast: take the last 12 months of monthly attrition by segment, apply a simple weighted average (more weight to recent), and project forward. Adjust for known events (post-bonus cliff, leadership change, layoff aftermath).

Hiring lead times by role

Role family	Approval → posted	Posted → offer accepted	Accepted → first day	Total lead time
Junior IC (eng / ops)	1–2 wk	4–6 wk	2–4 wk	7–12 wk
Senior eng / IC	1–2 wk	6–10 wk	4–8 wk (notice + relocation)	11–20 wk
Manager (line)	2–3 wk	8–12 wk	4–8 wk	14–23 wk
Director / executive	3–4 wk	12–20 wk	8–12 wk	23–36 wk
Sales rep	1 wk	3–5 wk	2–4 wk	6–10 wk

Plan backwards from need-by date

If you need a senior engineer productive (not started — productive) by Q3, work backwards: 16 weeks ramp + 16 weeks median lead time = the role must be approved 32 weeks earlier. This is the single most common workforce-planning error.

Ramp curves and effective capacity

Headcount in HRIS ≠ productive headcount. A senior engineer who joined yesterday is on your payroll but delivering ~0% of their eventual output. Workforce plans that ignore ramp systematically over-promise on capacity.

Role	Week 4	Week 8	Week 16	Week 26	Week 52
Junior IC	20%	40%	70%	85%	100%
Senior IC	15%	50%	85%	100%	—
New manager (line)	10%	30%	60%	85%	100%
External executive	5%	20%	50%	75%	95%
Sales (full quota)	10%	30%	60%	100%	—

These are typical curves from published benchmarks (DDI, BCG, RHR) and from internal data at most well-instrumented tech companies. Calibrate to your own ramp data once you have 30+ data points.

Effective FTE formula

Effective FTE this quarter = Σ (each employee's tenure-weighted ramp factor). A team of 10 where 3 joined within 8 weeks delivers ~7.5 effective FTE, not 10.

Recruiter capacity math

Recruiter capacity is the most-underestimated constraint in any hiring plan. The math:

Recruiter throughput

1
Tech recruiter (full-cycle)
3–5 hires/month sustainably. Above this, quality drops and recruiter burns out within 6 months.
2
Sourcer (no closes)
Generates pipeline for 8–12 hires/month — usually paired 1:2 with full-cycle recruiters.
3
Executive recruiter
1–2 placements/quarter. Plan for one external retained search budget per Director+ hire above the recruiter's strength area.
4
Coordinator
Supports ~3–5 recruiters. Often the first hire missing in scaling teams.

The recruiter-to-hire ratio

If your plan needs 60 hires this year and you have one recruiter, the plan will fail. Either reduce the plan or hire a second recruiter 4 months before the hiring crunch (recruiters themselves have a 10–14 week lead time).

Worked example: 80 → 130 in 12 months

A Series B company is at 80 employees on Jan 1, targeting 130 by Dec 31. Look like a 50-person hiring plan? Not quite.

Forecast attrition: 80 × 18% annual = ~14.4 exits. Round to 15.
Real hiring need = 50 net adds + 15 backfills = 65 hires (not 50).
Apply ramp loss: of the 65 hires, the average tenure-weighted ramp at year-end means effective capacity is ~55, not 65. So 'productive headcount' at Dec 31 is closer to 80 + 55 − 15 = 120, not 130.
Recruiter capacity: 65 hires / 12 months / 4 hires per recruiter/month = need ~1.4 recruiters sustained. Round to 2 (one will be ramping for 3 months).
Lead time: most of these need to start by Sep 30 to count as productive. Backwards 16 weeks of recruiting + 2 weeks approval means approvals for the back half need to be done by early June.

The honest plan

A 'realistic' 80 → 130 plan is actually: hire 65, expect 120 productive at year-end, budget 2 recruiters + 1 sourcer, and front-load approvals into Q1/Q2. CFOs who hear 'we'll be at 130' without the asterisks will plan revenue against capacity that doesn't exist.

The minimal workforce-plan spreadsheet

One row per role. Columns: Function, Level, Approval date, Posted date, Need-by date, Recruiter, Status, Expected start date, Ramp-adjusted end-of-year FTE contribution.
Monthly summary: opening HC, hires planned, hires actual, attrition planned, attrition actual, closing HC, variance.
Recruiter capacity check: hires planned per month ÷ 4 = recruiters needed. Highlight any month over capacity.
Attrition forecast: monthly rate × current HC, with seasonality (Jan / post-bonus spikes).
Refresh monthly. The plan that's not refreshed is a wish-list within 60 days.

FAQ

Frequently asked questions

Should we plan for regretted vs unregretted attrition separately?

Yes for executive reporting (top performers leaving is the metric leaders care about). No for capacity planning — a backfill is a backfill regardless of why the role opened.

How conservative should the attrition forecast be?

Use trailing-12-month actual + 1–2 percentage points if conditions are deteriorating (recent layoff, leadership change, equity refresh denied). The cost of under-forecasting attrition is much higher than over-forecasting it.

What's a good 'ramp factor' for purposes of headcount budgeting?

Most CFOs accept a blended ramp adjustment of 15–20% off new hires' annualised cost in their first 6 months — equivalent to assuming roughly 60–70% ramp midpoint. Negotiate this once per plan cycle.