Choose work within a real capacity boundary
What This Concept Is
A priority is a choice about constrained capacity. Separate unavoidable obligations, discretionary investments and uncertainty reserves before comparing projects. Opportunity cost is the valuable alternative displaced by a choice; it is not automatically a monetary loss.
Why It Matters Here
The earlier capacity lesson sized computers. This lesson accounts for engineering attention. A roadmap that assumes all calendar hours are project hours hides on-call, review, support and coordination. Adding a new platform also adds maintenance, migration and support obligations, even if its demonstration looks inexpensive.
Concrete Example
A fictional team has 120 engineering hours available over two weeks after leave and regular meetings. On-call/support is estimated at 32 hours, an incident repair requires 24, and a 16-hour uncertainty reserve leaves 48 discretionary hours. A 36-hour automation experiment fits; a 60-hour extraction does not. Doing both requires 96 hours, exceeding discretionary capacity by 48. Moving the reserve into planned delivery does not eliminate uncertainty.
The automation might save six hours per week, but requires two hours weekly maintenance. With an immediate stable four-hour net weekly saving after completion, its 36-hour investment breaks even after nine operating weeks. This excludes ramp-up, discounting and demand changes. At one hour gross saving and two hours maintenance it never breaks even on time savings alone. A safety benefit may still justify it, but must be argued separately.
Common Confusion / Misconception
Multiplying ordinal scores such as impact 5 and confidence 3 produces a ranking convention, not measured expected value. A high score does not override a security constraint or prove that two teams' estimates are comparable. Time saved is not automatically cash saved or spare deployable capacity.
How To Use It
Name the decision horizon, available capacity and estimate owner. Record a range for support demand, reserve, work and adoption cost. Compare do nothing, a small reversible experiment and the larger investment. State the displaced work and a decision-changing assumption. Consult the people who carry the operational cost before calling an option efficient.
Check Yourself
What happens if support demand rises by 20 hours? Does a partially built platform produce the claimed benefit? Which costs continue after launch? Who can approve moving a reliability obligation?
Mini Drill or Application
Recalculate the example with 52 support hours: discretionary capacity is 28. Propose a smaller experiment with an explicit result, or defer the 36-hour option. Do not quietly borrow hours from another team. Attach this calculation to the original strategy memo and roadmap.
Read This Only If Stuck
Fundamentals of Software Architecture — leadership, negotiation and effective teams. Google SRE: Eliminating Toil motivates measuring repetitive operational work and considering its root causes. Its organizational targets are not universal staffing rules.