Four at a time

The rule was set in 2004, when the practice was three people and four projects was an ambition rather than a ceiling. It has survived every subsequent expansion, including two that made it briefly ridiculous.

The reasoning is simple. A director can hold four projects in their head. At five, they begin to manage rather than design, and the work becomes the average of what the team produces rather than the best of it. There is no version of scale that solves this; it is a limit of attention, not of headcount.

The commercial cost is obvious. We turn away roughly six commissions a year, some of them substantial. We have twice delayed a site acquisition by more than a year because the pipeline was full. A larger competitor would call this leaving money on the table, and they would be right.

What we get in exchange is that every project has a director who has visited the site more than thirty times, and who can tell you without looking why a particular column moved 400 millimetres in 2021. That knowledge does not survive a portfolio of twelve.

Forty-one buildings in twenty-three years is not a fast rate. It is the rate at which we can do the work properly, and we have never found a way to make it faster that we were willing to live with.