capital-allocation git:20260728.2d6d1c9A

Immutable. This exact content is served forever at /api/v1/blob/21e5dd891d67a53b.

---
name: capital-allocation
description: Decide where to spend limited money and attention across competing opportunities, with a consistent basis for comparison. Use when planning a budget or choosing between investments.
---

# Capital allocation

Allocation is the highest-leverage repeated decision in a business, and
it is usually made by inertia: last year's budget adjusted. Doing it
deliberately means comparing opportunities on the same basis and being
willing to stop things.

## Method

1. **Compare on the same basis.** Expected return, time to return, and
   risk, so a marketing spend and an engineering hire can be weighed
   against each other rather than debated separately.
2. **Treat sunk costs as irrelevant.** What has been spent should not
   influence what to spend next, though it reliably does (see
   decision-journals).
3. **Reserve for optionality.** Committing every pound to current plans
   leaves nothing for the opportunity that appears, and opportunities do
   appear.
4. **Fund in stages against milestones.** Tranches released on evidence
   limit the downside of a wrong allocation and force honest checkpoints.
5. **Name what you are not funding and why.** An allocation decision
   that never says no has not allocated anything (see
   prioritization-frameworks).
6. **Distinguish maintenance from growth.** Keeping the business running
   is not optional and should not compete with growth bets on the same
   list.
7. **Review allocations against outcomes.** Whether last year's bets
   returned is the only way the process improves (see
   agent-strategy-review).

## Boundaries

Allocation frameworks structure judgement rather than replacing it, and
expected returns are estimates that flatter whoever produced them.
Financing, tax, and investor obligations constrain the choices and need
qualified advice. Attention is often scarcer than money and rarely
budgeted.