The cost of doing everything at once

When a management team tells us that everything is late, the instinct is to look for the bottleneck. Occasionally there is one. More often there is no single blockage at all, and the delay is produced by the number of things being attempted at the same time by the same small group of people.

This is not a motivation problem and it is rarely a capability problem. It is arithmetic, and it is worth setting out plainly because the arithmetic is genuinely counterintuitive to people who have been rewarded all their careers for taking on more.

The arithmetic

Take a team that can complete four projects in a year. Give it four projects sequentially and the first finishes in three months, delivering value for nine months before the year ends. Give it the same four in parallel and, in the best case where no time is lost to switching, all four finish in December. Same output, same effort, same people, and three quarters of the value has quietly disappeared.

That is the optimistic version, because it assumes switching is free. It is not. Every time somebody moves between two substantial pieces of work they pay a reload cost, and for complex work that cost is measured in hours rather than minutes. Add coordination, add the meetings that exist only because four things are moving at once, and the parallel version does not merely deliver later. It delivers less.

Why it happens anyway

Everybody knows this in the abstract, which raises the more interesting question of why it persists. The answer is political rather than analytical.

Starting a project is cheap and pleasant. It costs one meeting, it creates optimism, and it makes a sponsor feel heard. Stopping a project is expensive and unpleasant. It requires telling somebody their priority is not going to happen, and there is no ceremony for it, no announcement, and no credit. So the natural drift of any organisation is towards more open work than it can carry, and the drift accelerates the more responsive the leadership is.

Note the trap in that last sentence. The behaviour that creates the problem is a virtue. Attentive leaders who take internal requests seriously fill the queue faster than dismissive ones.

The two limits worth setting

The correction is unglamorous and mechanical. It comes down to two numbers that somebody has to own.

• A cap on what is open. Decide how many significant initiatives the organisation runs at once, write the number down, and treat it as fixed. When something new is genuinely more important than something open, the new one enters and the other stops. The point is not the specific number. It is that adding now requires removing.
• A stopping rule per project. Agreed at the start, in writing: the condition under which this is abandoned. Without it, projects do not end, they fade, occupying attention and headcount long after everybody privately knows they are finished.

The stopping rule is the harder of the two to introduce, and the more valuable. A portfolio where nothing is ever cancelled is not a portfolio of successes. It is a portfolio in which failure is undiscussable, which means the same money keeps being spent to avoid a conversation.

Where to start

Count what is open. Not what is in the plan, but every initiative that currently has somebody's attention, including the ones that were never formally approved and the ones that are technically paused but still generate messages on a Tuesday evening.

In our experience the count comes back at somewhere between two and four times what the leadership team believed, and the list itself does most of the persuading. Nobody argues for keeping forty things open once they can see forty things written on one page.

Then order them, take the top few, and give everything else a clear status: stopped, or waiting with a named date at which it will be reconsidered. Waiting is an honest answer. Quietly continuing at ten percent effort is not, and it is what most organisations do instead.