Reference

Capacity and Constraint

In Brief — Step 10 tells you to build a capacity view. It does not tell you how to calculate one, how to find the thing that is actually limiting delivery, or what to do when demand exceeds what the organization can absorb. This page does. It covers the capacity vocabulary, the arithmetic that turns headcount into a credible commitment, a twenty-two step operating sequence, constraint diagnosis, and flow control. It closes with a working Excel model you can use on Monday.

Why This Page Exists

The Introduction names a boundary of this framework plainly: it assumes you can see capacity, not that you command it. That is still true. Nothing here gives an EPMO authority over resource managers who do not report to it.

But there is a difference between not commanding capacity and not being able to describe it. Most governance functions fail the second test long before the first one matters. They report headcount, call it capacity, and are surprised when eight approved projects requiring the same three architects do not deliver. A capacity number nobody can reproduce is not evidence, and it will not survive contact with a functional leader who disagrees with it.

This page is the arithmetic and the operating sequence behind Step 10’s capacity view. It is method-neutral. Apply it at portfolio, program, or team level using whichever unit of capacity fits that level.

The core decision rule. Do not release work solely because the work is valuable. Release it when priority is explicit, required capacity is credible, dependencies are ready, and the system can absorb the work without damaging higher-priority outcomes.

The Capacity Vocabulary

Most capacity arguments are vocabulary arguments. One person means contracted hours, another means hours actually available for project work, and a third means what the team has historically delivered. Those are three different numbers and the gap between them is usually a factor of two or more. Name them separately before you plan with any of them.

Term	Definition	Management use
Theoretical capacity	Maximum possible time before any deduction	Reference only. Never a delivery commitment.
Gross capacity	Contracted or scheduled working time	Starting supply baseline
Usable capacity	Gross less normal availability and non-project commitments	Realistically allocable time
Effective capacity	Usable adjusted for proficiency, focus, ramp-up, dependencies and demonstrated performance	Credible delivery forecast
Reserved capacity	Intentionally uncommitted capacity for variability and recovery	Protects predictability
Demonstrated capacity	Historically achieved output or throughput	Calibrates planning assumptions
Constraint capacity	Capacity of the resource or process governing system throughput	Primary improvement focus
Surge capacity	Temporary exceptional capacity	Short-term response, not structural supply

The Calculations

Three formulas carry the whole model. They are stated here so the number can be reproduced on paper, argued with, and corrected — which is the only reason a capacity figure ever earns authority in a room.

Usable capacity

Gross capacity less planned absence, operational commitments, administrative obligations, and other committed work. Every deduction needs an owner and a source. A deduction nobody will defend is a deduction someone will delete.

Effective capacity

Usable capacity multiplied by a proficiency factor, a focus factor, and a delivery-effectiveness factor. Use ranges and calibrate against what the team has actually delivered. Precision here is false precision.

Committable capacity

Effective capacity less the reserve. This is the only number you may commit against, unless an authorized exception names the work being displaced or the exposure being accepted.

Worked example. A specialist has 160 gross hours in a month. After 16 hours of leave, 48 hours of operations and 16 hours of administration, usable capacity is 80 hours — half the headline number. At a 90 percent proficiency factor, an 85 percent focus factor and a 95 percent delivery-effectiveness factor, effective capacity is about 58 hours. With a 10 percent reserve, committable capacity is about 52 hours. The organization that plans this person at 160 hours is not optimistic. It is wrong by a factor of three.

Calculation	Method	Control
Gross capacity	Scheduled workdays × standard hours, adjusted for start and end dates	Use calendar-specific working days
Usable capacity	Gross less leave, operations, administration, training and committed non-change work	Every deduction has an owner and a source
Effective capacity	Usable × proficiency × focus × delivery-effectiveness	Use ranges; recalibrate with actuals
Reserve	Effective × reserve percentage	Reserve reflects volatility, not arbitrary padding
Committable capacity	Effective less reserve	The primary supply figure used for authorization
Demand-capacity gap	Committable capacity less approved demand	A negative result requires a decision. Do not hide it.
Demand-to-capacity ratio	Approved demand divided by committable capacity	Above 1.0 indicates overload
Commitment reliability	Completed as committed divided by commitments due	Define whether date, scope and quality all count
Forecast variance	Actual required capacity less forecast	Analyse bias by work type and skill

What gets deducted. Planned leave and holidays. Operational responsibilities and production support. Administrative and leadership duties. Recurring meetings and governance. Training and onboarding. Defect correction and rework. Known enterprise initiatives. Vacancies and recruiting lead time. Vendor mobilization and knowledge transfer. Context switching and split assignments. Business review, approval and acceptance effort. And the non-people constraints — environments, funding, procurement, regulatory windows — that stop work just as effectively as a missing person.

The Twenty-Two Step Sequence

Run the sequence once to establish the system. After that it is not linear: when something changes, return to the earliest step the change affected. The six phases below are the executive view; the full table is the practitioner view.

Phase	Steps	Principal decision or output
1. Define	1–2	Outcome, benefit, boundaries, planning horizon and controls
2. Shape demand	3–5	A complete, qualified, force-ranked demand inventory
3. Quantify	6–10	Time-phased demand, effective supply and reserve
4. Resolve gaps	11–13	Gap analysis, constraint evidence and scenarios
5. Commit	14–17	Trade-off decisions, sequencing, readiness and controlled release
6. Execute and learn	18–22	WIP control, performance, reforecast, constraint reassessment and iteration

Step	What it does	Accountable owner	Exit criterion
1. Define intended outcomes	Define output, outcome, benefit, baseline, target, timing, owner, and the consequence of delay.	Business sponsor	Leadership agrees on measurable success
2. Set planning boundaries	Set organizational scope, resource populations, work categories, units, horizons, owners, and cadence.	EPMO / capacity process owner	Units, owners, cadence and detail are defined
3. Create the demand inventory	Capture projects, products, programs, operations, mandates, defects, maintenance, improvement, and likely future work.	Portfolio owner	Known demand has an owner and a category
4. Qualify demand	Apply minimum-information and readiness criteria. Return incomplete items rather than holding them.	Demand owner	Minimum information and readiness thresholds met
5. Prioritize demand	Score value and urgency, then force a rank order or capacity-backed tiers.	Portfolio governance	Leaders accept the trade-offs
6. Translate demand into capabilities	Decompose work into time-phased roles, skills, proficiency, approvals, and non-people needs.	Program / solution lead	Demand can be compared with supply
7. Establish gross capacity	Calculate scheduled availability from employees, contractors, vendors, calendars, vacancies, and hiring plans.	Functional resource owner	Supply population and availability are complete
8. Calculate usable capacity	Deduct absence, operations, administration, support, training, and existing commitments.	Functional resource owner	Operational and availability deductions applied
9. Calculate effective capacity	Adjust for proficiency, ramp-up, focus, rework, dependencies, stability, and demonstrated performance.	Program and functional leaders	Adjustments are transparent and calibrated
10. Reserve capacity for variability	Set reserves from volatility and historical unplanned demand. Define who may consume the reserve.	Portfolio governance	Reserve is approved and visible
11. Compare demand with capacity	Compare demand and committable supply by period, skill, team, initiative, and category.	EPMO / program management	Surpluses and shortfalls are visible
12. Identify the system constraint	Use queues, age, blocked time, rework, and lost capacity to identify the throughput limiter.	Program manager with functional leaders	Constraint is specific and measurable
13. Develop scenarios	Model date, scope, stop and defer, reallocation, sourcing, hiring, automation, and risk options.	EPMO / program manager	Options cover scope, date, cost, capacity and risk
14. Make trade-off decisions	Choose what proceeds, waits, stops, changes, receives investment, or accepts risk.	Authorized governance body	Authorized demand fits capacity, or has a named exception
15. Sequence approved work	Sequence by priority, dependencies, constraint availability, readiness, and cost of delay.	Program manager	No unresolved critical collision
16. Validate delivery readiness	Apply the work-readiness gate and resolve failed criteria before release.	Delivery authority	The system can absorb the work without harming priorities
17. Authorize and release work	Issue formal authorization naming scope, timing, capacity, owner, and WIP slot.	Delivery authority	Entry policy met
18. Control work in progress	Enforce WIP and expedite policies. Swarm blockers. Finish before starting.	Team and program leaders	WIP limits and expedite policy observed
19. Monitor actual performance	Track flow, capacity variance, quality, commitment reliability, and outcomes.	Delivery and EPMO analysts	Leading and lagging measures current
20. Reforecast demand and capacity	Incorporate actuals and every approved change into one integrated forecast.	Program / portfolio manager	Assumptions and impacts reconciled
21. Reassess the constraint	Confirm whether the constraint remains, moved, or was misidentified.	Program manager	Constraint movement is explicit
22. Iterate from the change point	Route the process back to the earliest step the trigger affected.	Capacity process owner	The right upstream step is selected

Return-Point Logic

The value of a sequence is knowing where to re-enter it. Running the whole thing again every time something moves is how capacity planning becomes an annual ritual that nobody believes.

Trigger	Return to	Required response
Strategy or benefit changes	Step 1	Reconfirm intended outcomes and timing
New demand enters	Step 3	Inventory, qualify, prioritize and size before allocating
Priority changes	Step 5	Re-rank, and show what got displaced
Scope or staffing changes	Step 6 or 7	Recalculate time-phased demand or supply
Demand exceeds capacity	Step 13	Develop explicit trade-off scenarios
Readiness fails	Step 16	Hold the release and remediate entry conditions
WIP exceeds policy	Step 18	Stop starting. Finish work. Protect the constraint.
Actuals differ from forecast	Step 20	Reforecast, then route to the affected upstream step
The constraint moves	Step 12	Diagnose, and subordinate the system to the new constraint

Finding the Constraint

Improving work that is not the constraint increases activity without increasing completed outcomes. That is the single most useful idea in this section, and it is the one most governance forums act against every week by pressing the teams that are visibly busy rather than the one that is quietly governing throughput.

Constraint statement template. The current constraint is [resource, decision, process, dependency or policy] because it limits [specific outcome or throughput measure]. We will manage it by [actions]. We will know it has moved when [evidence].

	Focus	Program application
1	Identify	Determine the resource, policy, decision, dependency or process currently governing delivery
2	Exploit	Use existing constraint capacity properly before adding cost. Remove avoidable interruptions, rework, low-value meetings and poor inputs.
3	Subordinate	Align upstream and downstream activity to support the constraint. Regulate release so the queue does not grow without purpose.
4	Elevate	Add capacity only when better use of current capacity is insufficient: funding, staffing, automation, policy change or scope reduction.
5	Repeat	Reassess once performance improves. The constraint has probably moved.

The weekly constraint review asks seven questions. Is this still the primary constraint? How much work is waiting in front of it, and how old is the oldest item? Is it ever idle because inputs are late or incomplete? How much of its capacity is lost to rework, interruption, or lower-priority work? Is it processing the highest-value work available? What must other teams do differently to support it? And what evidence would show that it has moved?

Manage Flow, Not Percent Complete

Percent-complete is a subjective number that can rise for weeks while nothing finishes. Flow measures cannot be reported green while queues grow, because they are counts of observable things.

Measure	Definition	Management question
Throughput	Completed items per week or month	Is the system finishing more work?
Cycle time	Elapsed time from start to finish	How quickly does active work move?
Lead time	Elapsed time from request to completion	How long does the requester actually wait?
Work in progress	Started but unfinished work	Are we diluting focus and growing queues?
Work-item age	Elapsed active time on each open item	What is becoming stuck before it is officially late?
Blocked time	Time an item cannot progress	Where are impediments consuming delivery time?
Queue size	Items waiting at the constraint	Is demand exceeding effective capacity?
Rework rate	Items returned for correction	Are poor inputs consuming the constraint?
Decision aging	Number and age of unresolved decisions	Is governance accelerating or delaying delivery?

Change the status language and the conversation changes. Less useful: “the design workstream is yellow and 75 percent complete.” More useful: “the team completed six of eight planned designs. Two are waiting on architecture decisions, the oldest for nine business days. That decision queue is now the constraint and it threatens the October integration milestone. The architecture director will decide both by Friday.” The second version contains a constraint, an age, a consequence, an owner and a date. The first contains a colour.

Limit Work in Progress

Starting work creates the appearance of momentum. Finishing work creates value. Concurrent work multiplies context switching, queues and handoff delay, and it makes every forecast less reliable at exactly the moment leadership starts asking for certainty.

The operating policies are short. Set explicit limits on active initiatives, work packages or deliverables. When the limit is reached, help finish something, resolve a blocker, improve an input, or support the constraint — do not start. Exceeding a WIP limit requires sponsor approval. No team starts lower-priority work while a higher-priority item is blocked and recoverable. Expedited work requires explicit authority, a documented reason, and the name of the work it displaces. Every item that has not moved within the agreed ageing threshold gets flagged.

The Work Readiness Gate

This is the control that connects capacity to Step 8. Authorization asks whether the conditions for responsible execution exist. This gate asks the same question of the delivery system rather than the project.

Criterion	Pass evidence	If it fails
Outcome	Measurable outcome with a named owner	Return to Step 1 or 4
Priority	Approved rank or tier	Return to Step 5
Capacity	Time-phased committable skills available	Return to Steps 7–14
Dependencies	Critical predecessors committed and accepted	Hold the release; remediate at Step 15
Acceptance	Definition of done and business acceptance criteria	Clarify before authorization
Constraint protection	The constraint has capacity and quality inputs	Re-sequence or improve the inputs
WIP headroom	Release does not breach policy	Wait, or finish active work first
Funding and environment	Required non-people resources available	Escalate or defer

When Demand Exceeds Capacity

There are eight moves and no others. Naming them as a closed set is what turns an argument into a decision, because the room stops debating whether a gap exists and starts choosing which trade-off it is willing to own.

Choice	Use when	Primary trade-off
Extend dates	The outcome stays valuable but capacity cannot be added	Time
Reduce scope	The deadline matters more than full content	Scope and benefit
Stop or defer work	A higher-priority outcome needs the scarce capacity	Portfolio value mix
Reassign capacity	The skills exist elsewhere and movement is feasible	Other commitments
Add sourcing	External capability can ramp in time	Cost and coordination
Develop capability	The need is durable and the lead time acceptable	Near-term speed
Improve constraint productivity	Losses, rework or interruption are material	Process change
Accept risk	The exposure is understood and authorized	Reliability and risk

Publish the trade-off. When an urgent request is accepted, name the work it displaces. If leaders will not defer, reduce, pause or stop anything, the organization does not have priorities. It has an overloaded queue with a ranking written on it.

Decision Rights and Cadence

Decision	Recommended owner	Evidence required
Approve and rank enterprise demand	Portfolio governance / executive sponsor	Outcome, value, mandate, cost of delay, capacity demand
Allocate or reserve shared capacity	Portfolio authority with functional leaders	Capability supply, commitments, scenarios
Release program work	Program sponsor and delivery authority	Readiness, WIP headroom, dependencies, effective capacity
Exceed a WIP limit or expedite	A named governance authority	Reason, duration, displaced work, outcome impact
Change scope, date or investment	Business sponsor / steering committee	Scenario comparison and a recommendation
Accept capacity risk	An authorized executive	Exposure, consequence, mitigation, review date

Cadence	Focus	Mandatory output
Weekly team review	Actual flow, WIP, ageing, blockers, near-term capacity	Finish, unblock and release decisions
Biweekly program review	Cross-team skills, dependencies, shared constraints, forecast	Re-sequencing and escalation
Monthly portfolio review	Demand-capacity gaps, allocation, scenarios	Start, stop, defer, source or reallocate
Quarterly strategic reset	Outcome assumptions, capability direction, investment mix	Allocation ranges and capability actions

A decision that is not made is still a decision — to accept the cost and the risk of waiting. Track the age of open decisions the same way you track the age of open work.

Ninety Days

Period	Objective	Key actions
Days 1–30	Establish visibility	Name the process owner and governance authority. Publish definitions, categories, horizons and units. Build the complete demand inventory. Establish the gross and usable capacity baseline. Identify critical skills and likely constraints.
Days 31–60	Create decisions	Qualify and prioritize demand. Calculate effective and committable capacity. Set reserves and WIP policy. Produce the first demand-capacity heat map. Make the first explicit scenario decisions.
Days 61–90	Control release and learn	Launch readiness gating and controlled release. Start the three cadences. Track actual flow and capacity variance. Recalibrate factors and reserves. Reassess the constraint and institutionalize the return-point logic.

The Workbook

Everything above is implemented in an editable Excel model. It is not a blank template — the calculations, the gap analysis and the dashboard are built. Replace the sample rows with your own demand and resources.

Worksheet	Purpose
Instructions	Use sequence, colour conventions and workflow
Assumptions	Planning horizon, standard hours, reserve and planning factors
Demand Inventory	Capture and qualify all demand
Resource Capacity	Gross, usable, effective, reserve and committable supply
Skill Matrix	Map people to skills and proficiency
Demand Profile	Time-phase capacity demand by initiative and skill
Gap Analysis	Compare demand and supply by period and skill
Scenarios	Compare trade-off options side by side
Constraint Register	Track constraint evidence, actions and movement
Decision Log	Record required and completed decisions
Dashboard	Summarize the key capacity indicators

Download the Workbook (.xlsx) →

How This Relates to the Twelve Steps

This page is not a thirteenth step. It is the depth behind three of the existing ones.

→ Step 6, Prioritize: the scoring model tells you what matters most. It cannot tell you what is executable. The demand-to-capacity ratio by skill and period is what stops a correctly ranked portfolio from being an incorrectly staffed one.

→ Step 8, Authorize: the eight authorization conditions include confirmed capacity. The work readiness gate above is how that condition gets tested rather than asserted.

→ Step 10, Portfolio Governance: Step 10 names the dependency map, the capacity view and the saturation view as the three portfolio-level instruments. This page is how the capacity view is built and kept honest.

Where this stops. None of this creates authority. In a matrixed organization the functional managers still own the supply, and nothing here obliges anyone to honour an allocation. What it buys you is a number that can be reproduced, a constraint named with evidence, and a trade-off published before the meeting rather than discovered after it. That is usually enough, and it takes about two cycles to earn.

This material synthesizes established public practice — PMI’s Governance of Portfolios, Programs, and Projects and The Standard for Program Management, the Theory of Constraints Institute’s five focusing steps, and flow metrics as documented by Atlassian and Scrum.org — into one operating sequence. Tailor it to your governance requirements, delivery method, regulatory context and decision authorities.