Choose the method for the decision, not for convenience alone
An allocation method answers a bounded managerial question: what share of a defined pool should be assigned to a job, product or service for this decision? A driver is more useful when it has a plausible relationship with resource consumption and can be measured consistently. Easy-to-export revenue is not automatically a good driver for workshop power, setup labour or quality-control work.
| Pool pattern | Candidate evidence | Challenge before use |
|---|---|---|
| Labour-led supervision and support | Productive labour hours or team time | Does time quality vary enough to need separate pools? |
| Equipment use, maintenance and power | Machine hours, cycles or setup events | Does idle capacity belong in active-object rates? |
| Order handling or customer support | Orders, tickets or service events | Does one complex event consume far more than another? |
| Occupancy and facilities | Area, workstation or time occupied | Is the space committed, shared or unused? |
Select and challenge an allocation base
Prerequisites, sequence and checkpoints
- State the decision, cost object, pool total and period. (not complete)
- Separate traceable costs, shared overhead and unused-capacity cost. (not complete)
- Shortlist drivers such as labour hours, machine hours or direct cost. (not complete)
- Allocate the identical pool under each candidate basis and reconcile the total. (not complete)
- Keep the unallocated contribution view and stop if the decision reverses without better evidence. (not complete)
Compare bases without changing the pool
| Driver | Job A allocation | Job B allocation |
|---|---|---|
| Labour hours: 75% / 25% | 9,000 CU | 3,000 CU |
| Machine hours: 25% / 75% | 3,000 CU | 9,000 CU |
| Direct cost: 60% / 40% | 7,200 CU | 4,800 CU |
Reconcile the allocation decision
Reproducible user scenario
Every basis allocates the same 12,000 CU pool; only the object shares change.
| Step | Input or arithmetic | Decision meaning |
|---|---|---|
| Pool check | Job A + Job B = 12,000 CU under every basis | A failed reconciliation invalidates the table |
| Sensitivity | A ranges from 3,000 to 9,000 CU | Driver choice is decision-relevant |
| Unallocated view | Show contribution before the 12,000 CU pool | Prevents allocation from hiding underlying unit economics |
What to do when reasonable methods disagree
Job A receives 9,000 CU under labour hours but only 3,000 CU under machine hours. That six-thousand-unit range is not noise to average away. It tells the reviewer that the chosen pool probably contains activities driven by different things, or that one proposed driver is weak for the decision.
Decision branch
- If one driver clearly matches the pool and records are reliable, use it and document the rationale. (not complete)
- If the pool contains materially different activities, split it into a small number of explainable pools rather than adding arbitrary precision. (not complete)
- If the decision reverses across reasonable bases, keep the unallocated contribution view and gather better driver evidence. (not complete)
- If unused capacity is material, show it separately instead of forcing active jobs to absorb it silently. (not complete)
Limitations, evidence and next action
Use the calculation owner for the next step
- Project Overhead Recovery Planner
Choose a transparent overhead recovery base across jobs
Questions and boundaries
- Which allocation base is correct?
- No universal base exists. Select for the stated decision, explain the connection and test sensitivity.
- Should all overhead be allocated?
- Not automatically. Keep unused capacity and an unallocated view visible where they matter.
Sources and scope
- Cost drivers — OpenStax: Supports cost-pool and driver selection; allocation is not proof of causation.