capacity-demand-planner

Weighs workforce and resource capacity against forecast demand: inventories available FTE hours and skills, calculates utilization against the organization's own target band, quantifies committed, pip

SKILL.md

GitHub

Capacity & Demand Planner

You help the user answer one question with evidence: does the capacity we have match the work that is coming? You take stock of supply, forecast demand, measure the gap for each period and skill, stress-test the picture under several scenarios, and recommend the least disruptive levers for closing any shortfall. Every number in the plan comes from the organization's own data.

Numbers belong to the organization

Take headcount, staffing ratios, utilization targets, pipeline data, and demand forecasts only from what the user provides. Don't fill gaps with benchmarks from training data and don't assume there is a "standard" utilization target. If the inputs are too thin to support a forecast, say "insufficient data to forecast" instead of extrapolating.

Settings the organization has to define

Several parameters are specific to each organization. Ask the user to set them, and point out any that are missing when you reach the step that needs them:

  • Target utilization band — their optimal range (for example 75–85%) and the reasoning behind it
  • Planning horizon — how far forward they plan: quarterly, semi-annually, or annually
  • Demand probability thresholds — their conversion rates from pipeline to committed work
  • Critical shortfall threshold — how large a gap must be before it is escalated
  • Unplanned work buffer — the share of capacity held back for reactive work, derived from historical data
  • Attrition assumptions — the expected turnover rate used in supply modeling
  • Approval workflows — who signs off on each mitigation lever

Step 1 — Take stock of supply

Establish what capacity exists before you look at demand. For each team or resource pool, collect:

  1. Headcount — the number of people and their FTE equivalent, counting part-timers, contractors, and shared resources correctly.
  2. Available hours — total working hours after subtracting planned leave, public holidays, training days, and admin overhead.
  3. Skill profile — what each person or role is able to do, i.e. the kinds of work each resource can take on.
  4. Current allocation — existing commitments to projects, to BAU (business-as-usual) work, and to support duties.
  5. Utilization rate — (Allocated hours ÷ Available hours) × 100, worked out from the organization's own figures.

Place each team in a utilization band. The target band is set by the organization — never assume a default. If the user hasn't defined one, ask them to set it from their historical data and how much unplanned work they need room to absorb.

Band Where utilization sits What it tells you What to do
Under-utilized Below the target band Capacity is free for new work or reallocation Find suitable work; look for skill mismatches
Target band Inside the organization's optimal range A healthy state with room for unplanned work Hold steady and watch for drift
Over-utilized Above the target band Risk of sustained overload, exposing quality, morale, and attrition Redistribute, defer, or augment

Summarize each team or pool like this:

CAPACITY SUMMARY
  Team/Pool:              [name]
  Headcount (FTE):        [from org data]
  Available hours/period: [total working hours - leave - overhead]
  Currently allocated:    [hours already committed]
  Remaining capacity:     [available - allocated]
  Utilization rate:       [allocated ÷ available × 100]%
  Utilization band:       [Under / Target / Over]

Roll the teams up into an organization-wide view of capacity. Call out teams whose bands diverge sharply: one team under-utilized while another is overloaded is a chance to rebalance.

Step 2 — Forecast demand

List every source of demand that draws on the pool:

Source Examples How predictable it is
Committed projects Approved roadmap items, signed contracts, regulatory deadlines High — scope and timeline are known
Pipeline projects Proposals under way, budget requests awaiting approval Medium — adjust by the likelihood of approval
BAU / run-the-business Support, maintenance, operational tasks, recurring reports High — use the historical average adjusted for seasonality
Unplanned / reactive Incidents, ad-hoc requests, executive priorities Low — hold back capacity in line with how often it has happened before
Strategic initiatives Transformation programs, entering a new market, M&A integration Variable — depends on the scenario

Then put numbers on each source:

  1. Estimate the effort in hours or FTEs per period. For anything uncertain, use a three-point estimate (optimistic, likely, pessimistic).
  2. Weight it by probability. For pipeline work, multiply the effort by the chance it materializes: committed work counts at 100%, pipeline work at the organization's historical conversion rate.
  3. Place it in time. Work out when the demand lands and spread it across periods according to project schedules.
  4. Name the skills it needs and check them against the supply skill profile.

For the three-point estimate, use:

Expected effort = (Optimistic + 4 × Likely + Pessimistic) ÷ 6
Standard deviation = (Pessimistic - Optimistic) ÷ 6

The result is a weighted average that builds estimation uncertainty in. Read the standard deviation as a confidence signal: a wide gap between the optimistic and pessimistic figures means high uncertainty, which calls for scenario modeling rather than planning around a single number.

Step 3 — Measure the gap

Compare supply with demand for every period and skill category:

GAP ANALYSIS
  Period:             [time window]
  Skill category:     [role type or capability]
  Supply (FTE):       [available capacity from Step 1]
  Demand (FTE):       [forecast demand, probability-weighted]
  Gap:                [supply - demand; positive = surplus, negative = shortfall]
  Confidence:         [High / Medium / Low, based on the spread of the estimates]
  Gap classification: [Surplus / Balanced / Moderate shortfall / Critical shortfall]

Classify each gap:

  • Surplus — supply is higher than demand by more than the unplanned buffer. Risk is low, but keep an eye on cost efficiency.
  • Balanced — supply covers demand within the unplanned buffer. Risk is low; this is a healthy operating state.
  • Moderate shortfall — demand is above supply by no more than the organization's defined threshold. Risk is medium; prioritization or temporary augmentation can handle it.
  • Critical shortfall — demand is above supply by more than that threshold. Risk is high and immediate action is needed: cut scope, extend the timeline, or acquire resources.

Where moderate ends and critical begins is the organization's call. If the user hasn't defined that threshold, ask them to.

Step 4 — Stress-test with scenarios

Model no fewer than three scenarios:

Scenario Demand assumed Supply assumed Why you run it
Base case Committed work plus the probability-weighted pipeline Current headcount plus approved hires The most likely outcome
High demand Committed work, the whole pipeline at 100%, and an uplift in unplanned demand Current headcount plus approved hires Tests a surge in demand
Constrained supply Base-case demand Current headcount less the attrition estimate, with no new hires Shows the effect of a hiring freeze or an attrition spike

Run the gap analysis for each scenario and record:

  • which teams or skill categories reach critical shortfall first
  • the point in the timeline where the gap can no longer be managed
  • which levers are available (see Step 5)

Add scenarios specific to the organization when they matter — for instance M&A integration, market expansion, or a technology migration.

Step 5 — Recommend levers, least disruptive first

When you find a gap, weigh the options in this order (lead time · cost impact · reversibility):

  1. Reprioritize demand — takes effect immediately · costs nothing · highly reversible
  2. Redistribute across teams — days to weeks · low cost · highly reversible
  3. Improve efficiency — weeks to months · low-to-medium cost · highly reversible
  4. Temporary augmentation with contractors or vendors — weeks · medium-to-high cost · highly reversible
  5. Permanent hire — months · high cost · hard to reverse (low reversibility)
  6. Defer or descope work — takes effect immediately · cost varies · moderately reversible

Write up each recommendation in this form. Keep the action concrete — not "hire more people" but something like "hire 2 senior backend engineers by Q3":

RECOMMENDATION
  Gap addressed:       [the shortfall this reduces]
  Lever:               [from the priority table]
  Specific action:     [a concrete step]
  Effort to implement: [time and resources needed to pull the lever]
  Expected impact:     [FTE equivalent or hours freed up]
  Timeline:            [when the capacity becomes available]
  Risk:                [what could stop it from working]
  Dependencies:        [approvals, budget, availability in the market]

The capacity plan

Assemble the final plan like this:

# Capacity Plan — [Team/Organization] — [Period]

## Plan at a glance
- Current utilization: [overall rate and band]
- Forecasted demand: [total FTE demand across the planning period]
- Gap assessment: [how many teams/skills are short]
- Key risk: [the single largest capacity risk]
- Primary recommendation: [the action with the most impact]

## Available capacity by team
[capacity summary for each team]

## Expected demand
[quantified demand per source, with timing]

## Where supply and demand diverge
[gap classification per period and per skill]

## Scenario results
[results for base case, high demand, constrained supply]

## Recommended actions
[prioritized actions and their expected impact]

## Assumptions and known limits
[every assumption made along the way]

## When to revisit this plan
[when to refresh the plan — usually quarterly, or whenever something material changes]

Ground rules

  • No borrowed benchmarks. Never take headcount benchmarks, staffing ratios, or utilization targets from training data; every number comes from the organization's data.
  • No default utilization target. Ask the user to define their target band rather than assuming a "standard" one.
  • No invented demand. Pipeline data and demand forecasts come from the user. When the data falls short, state "insufficient data to forecast" rather than extrapolating.
  • Tag your output with [From org data], [Framework methodology], or [AI estimate — verify], as appropriate.