Headcount & Capacity Planning

Tap menu for chapters & tools

Chapter 18

Headcount & Capacity Planning

Staff the work you can finish, not the wishlist you can present

22 min read

Headcount conversations fail when they start with “we need more people” and end with a number. Strong leads start with the work: streams of change, service ownership, interrupt load, and the outcomes that matter this half. Capacity planning turns desire into a plan stakeholders can fund or cut with eyes open.

You will not always get the hires you want. Your job is to make trade-offs explicit: what ships if we stay flat, what slips if we add scope, and what burns people if we pretend math is optional.

Capacity is not headcount

Headcount is seats. Capacity is the fraction of time that can do planned product and tech work after meetings, on-call, hiring, and drag. A team of eight with heavy interrupts can have less shipping capacity than a team of five with clean ownership.

  • Planned capacity: roadmap and committed projects.
  • Interrupt capacity: incidents, support, drive-by asks.
  • Investment capacity: debt, platform, hiring, learning.
  • Unavailable capacity: PTO, leave, ramp-up, part-time.
Rule of rough numbers

If you do not track interrupts, assume 20-40% of time is not roadmap. New hires are not full capacity for months. Managers are not full IC capacity. Plan with those truths, not with hope.

A simple capacity model

  1. List the work streams for the next quarter (product, reliability, platform, keep-the-lights-on).
  2. Estimate size in engineer-weeks, not story points theater. Ranges are fine.
  3. Compute available engineer-weeks: people × weeks × focus factor (often 0.6-0.75).
  4. Subtract known ramps, PTO, and on-call weeks.
  5. Compare demand vs supply. Cut or sequence until the plan is honest.
  6. Name the explicit “won’t do” list so scope does not sneak back in chat.

When to ask for headcount

Hire when the constraint is sustained demand against a clear stream, not when one bad quarter of meetings made everyone tired. Bring evidence: lead time, WIP, on-call load, roadmap commitments, and risks of staying flat.

  • Role brief: problem to solve, not a stack shopping list only.
  • Why now: what fails if we wait a quarter.
  • Alternatives considered: scope cut, vendor, platform leverage, rebalance.
  • Ramp plan: who onboards, when the seat becomes net positive.
  • Success signal: what changes in 6 months if the hire works.

Allocation patterns that protect delivery

  • Keep a stability reserve: do not plan 100% of capacity to roadmap.
  • Limit concurrent major bets per team (often one primary, one secondary).
  • Separate stream ownership from temporary swarm work with an end date.
  • Fund platform or debt with a fixed percentage when pain is chronic.
  • Rebalance when topology is wrong before you add more people to a broken interface.

Adding people to a confused ownership map multiplies confusion. Fix the map, then staff the streams.

Conversations with your manager and finance partners

  1. Lead with outcomes and risks, then the capacity math.
  2. Offer scenarios: stay flat / +1 / +2 with different delivery shapes.
  3. Be ready to recommend the cut if headcount is denied.
  4. Never accept infinite scope with finite capacity in silence.
  5. Review the plan monthly; capacity plans rot faster than roadmaps.
Manager ops link

Pair this chapter with hiring scorecards (Ch 11), team topologies (Ch 17), and the priority stack. Headcount without a hiring bar or ownership model is just cost.

Diagrams for this topic

Visual models you can redraw on a whiteboard or in a design review.

Diagram

Demand vs capacity

Demand

  • Roadmap bets
  • Keep-the-lights-on
  • Debt / platform
  • Hiring & onboarding

Supply

  • Engineer-weeks
  • Focus factor
  • PTO / ramp
  • Stability reserve

If demand wins

  • Heroics
  • Hidden cuts
  • Burnout
  • Missed SLOs

Honest plans balance both sides

Diagram

Headcount decision loop
  1. 1Map streams
  2. 2Model capacity
  3. 3Scenario options
  4. 4Hire or cut scope
  5. 5Review monthly
  6. Loop: after the last step, return to the first.

Diagram

Hire vs not yet
SignalHireDo not hire yet
ConstraintSustained stream demandTemporary spike or unclear ownership
EvidenceLead time, WIP, on-call loadVibes and wishlists only
AlternativeTried cut/vendor/platformHave not considered scope cuts

Worked examples

Concrete situations: what goes wrong, what to try instead, what changes.

Headcount without a role brief

Setting. You ask for two engineers because the team feels busy.

Common miss

Present a number and a stack list. No capacity math, no won’t-do list.

Stronger move

Show demand vs supply for the quarter, three scenarios, and role briefs tied to streams.

Outcome. Leadership can fund or refuse with shared facts.

100% roadmap utilization

Setting. Planning fills every engineer-week with feature work.

Common miss

Smile when interrupts arrive and expect heroes to absorb them.

Stronger move

Keep a stability reserve, track interrupt load, and replan when reality shows up.

Outcome. Fewer surprise slips and less silent burnout.

Hiring into a topology mess

Setting. Three teams own the same flow. You request more people for speed.

Common miss

Add headcount to every team and hope handoffs get faster.

Stronger move

Clarify stream ownership and interaction modes first, then staff the stream that owns the outcome.

Outcome. New people accelerate a clear system instead of a meeting network.

Resources specific to this chapter

Go deeper with books, videos, and tools matched to this topic, not a generic dump.

Practice

End-of-chapter drill

For the next 6-8 weeks, list demand in engineer-weeks (roadmap, KTLO, debt, interrupts) and supply (people × weeks × focus factor minus PTO/ramp). Write one scenario if you stay flat and one if you add one hire.

Hint: If you cannot estimate, your planning system is the first gap, not the headcount number.

Capacity plans are moral documents: they say who will be overloaded if you stay silent.