Cost of Delay
The economic loss per unit of time that a decision or delivery is postponed, which turns prioritisation from opinion into arithmetic.
Prioritisation arguments stall because they compare value without comparing urgency, and those are different. A feature worth £2 million delivered any time this year and one worth £500,000 that must land before a regulatory deadline are not ranked by value.
Cost of delay expresses what each week of postponement costs: lost revenue, a compliance penalty, a competitor's advantage compounding, continued operational spend that a change would remove. Dividing it by duration gives a ranking that maximises value delivered per unit of scarce capacity, and it frequently inverts the intuitive order.
The application to architecture work is where it is most useful, because platform, modernisation and reliability work is chronically outranked by features. Expressing them properly changes the conversation: the migration that removes £40,000 a month of licence cost has a quantifiable cost of delay, and so does the reliability work that prevents an outage class costing £180,000 per incident at a historical rate of one a quarter.
The honest caveat is that the inputs are estimates and can be constructed to support a preferred answer. The discipline that keeps it useful is stating assumptions explicitly and comparing relative magnitudes rather than pretending to precision — the aim is to establish that one item is an order of magnitude more urgent than another, which is usually clear and usually sufficient.