Directory · P
New here? Start with the premise →
Priorities
2 min read · 409 words
Priorities are the operator’s ranking of what gets the limited resources first.
The hardware has finite bandwidth — finite time, finite energy, finite attention. Every operation consumes some of these. The total demands always exceed the total capacity. This is structural, not a temporary condition. There is no future state where the operator has bandwidth for everything. Priorities are the ranking that determines which demands are funded and which aren’t.
Most operators have implicit priorities, established by default rather than by design. The squeaky demand gets funded. The familiar demand gets funded. The demand someone else cared about gets funded. The actual most-important demands often go underfunded, because they don’t make noise, or because they require sustained attention rather than reactive response, or because they are about long-term outcomes the system doesn’t feel pressing today.
The cost of unconscious prioritization is large. The operator at the end of a year, looking back, finds that the things they would have said were most important got the least time, while peripheral demands consumed the bandwidth. This isn’t because they didn’t care. It is because the priorities running their schedule were not the priorities they would have chosen. The defaults were running, and the defaults had different priorities than the operator did.
From the chair: priorities have to be made explicit and applied as filters. Not aspirations. Filters. This is what gets the bandwidth this week. The operations that align with the priority get funded. The ones that don’t get declined, deferred, or done at lower investment. The decline is uncomfortable. It is also the entire mechanism by which priorities become real. Priorities that don’t produce decline of competing demands are not priorities. They are wishes.
The hard part: priorities require the operator to disappoint or refuse some demands that have legitimate claims. Not every reasonable request can be funded. The operator who tries to fund every reasonable request ends up funding the loudest ones at the cost of the most important ones. Saying no to the legitimate is the only way to say yes to the priority. This is not selfishness. It is the operating constraint of a system with finite bandwidth.
Set the priorities. Apply them. Notice what gets refused and accept the discomfort of refusing it. The alternative is the year of misallocated bandwidth that the operator can no longer recover.