We had forty-one live projects at the start of this year. A realistic assessment of what my department can run concurrently, with the people we actually employ, is about twelve.
Nobody had ever decided to overcommit by that margin. The investment committee assesses each case on its merits, one at a time, against a budget. It has never assessed a case against capacity, because no capacity figure existed in a form the committee would recognise, and a business case that stands up on its own is very difficult to refuse on the grounds that we are busy.
What overcommitment produces is not a queue. It is a uniform slowing of everything. Six named specialists appeared in nineteen separate plans, so every one of those plans was built on a person who could give it a fraction of a week. Elapsed times roughly doubled, which meant the benefit assumptions in the cases were decaying while the work was in progress. And because attention was allocated by whoever asked most persistently, the order in which things finished had very little to do with their value.
The part that took me longest to see is that nothing was ever stopped. Pausing a project requires no decision, no meeting and no admission; it happens by itself when the people are needed elsewhere. Cancelling one requires somebody to say out loud that an approved case was wrong, and no organisation makes that easy. So the portfolio only ever grew, and a third of it consisted of work that had been effectively dormant for months while still counting as live.
We now publish a capacity statement in the same units as the plan, by team and by month, and a project is either staffed or it has not started. There is a visible queue with an order in it. Starting something requires naming what finishes or stops. And we report work started against work finished, rather than percentage complete, which is a number that can rise forever without anything being delivered.
In the first quarter we stopped four, paused six deliberately and in writing, and completed nine. Three of the four we stopped turned out to have no sponsor left at all.
Prioritisation is not deciding what matters. Nearly all of it mattered. It is deciding what waits, and somebody has to be permitted to say so.
– Serguey Shinder