Philippe Kahn coined this law in 1992:

“The productivity of a software developer in a team of N people is diminished by dividing it by the cube root of N.”

I’m now calling this team-size driven coefficient that reduces the productivity of a programmer The Kahn Factor. I derive it by computing the percentage of individual productivity lost because of the team size.

For a team of 1: 0%.

For a team of 2: 20%

For a team of 3: 30%

For a team of 5: 40%.

For a team of 10: 53%.

For a team of 25: 65%.

For a team of 50: 72%.

For a team of 100: 78%.

For a team of 1000: 90%.

According to the law, roughly, you are a third as productive if there’s three of you and half as productive if you work with 10 people.

Steve Yegge once said:

I believe that code weight wrecks projects and companies, that it forces rewrites after a certain size, and that smart teams will do everything in their power to keep their code base from becoming a mountain.

I wonder if smart companies also do everything in their power to keep their software development teams as small as possible, and let them interact through pure service interfaces.

Interesting consequences of this:

• This is not like decreasing marginal productivity, because the output of the first unit is affected by the amount of subsequent units.
• If you bring really productive people at the beginning, bringing normally productive later can really throw a wrench in the productivity of the whole team.
• The curve is really aggressive at the beginning and more gentle later. From an overall productivity approach, once you’re in the plus 50% zone, it makes sense to keep on hiring, assuming that the extra value is higher than the extra cost. The law seems to push you towards the extremes of either a very small, very talented team, or a large, well-organized, average team.

I wonder how true this law is. It sure feels close to truth, though it doesn’t seem to fully capture it either.