Bob Lewis
Columnist

Handling a difficult combinatorial – namely, growth

analysis
Mar 21, 20072 mins

Dear Bob ...Thanks for doing the math (in "The mathematics of organizational dysfunction," Keep the Joint Running," 1/22/2007).The combinatorials of a growing organization (n(n-1)/2) are negatively affecting my team of systems engineers who install and operate our proprietary systems. In the past year our development teams have increased headcount fourfold yet our engineering team hasn't grown at a commensurate

Dear Bob …

Thanks for doing the math (in “The mathematics of organizational dysfunction,” Keep the Joint Running,” 1/22/2007).

The combinatorials of a growing organization (n(n-1)/2) are negatively affecting my team of systems engineers who install and operate our proprietary systems. In the past year our development teams have increased headcount fourfold yet our engineering team hasn’t grown at a commensurate rate.

What a surprise it is that all those new software developers would be spitting out more code so we now have a big increase in the volume of new software from these dev teams. Compounding the growth in bits is the growth in conversations that need to happen to deliver the bits to our clients which is taking evem more time from an already constrained team.

To mitigate or reverse the effect of n(n-1)/2 should I be reducing n by trying to cut the numbers of interactions between development and engineering? And if that’s a good approach how can I do that and somehow still get the communication we need to understand our inputs?

Yours in permutuity,

– Math guy

Dear Math guy …

The way it generally works is this: Small organizations operate through relationships (which scale according to the n(n-1)/2 formula). Larger ones operate through well-defined processes (which scale according to the formula ax+b).

The short version: It sounds to me like you need to implement a well-defined change control process that pipelines the changes developers spit out. In general this approach also ties into the institution of a release management discipline, so that changes are batched up into releases that are put into production on a schedule, rather than being put into production one at a time.

I’m flying a bit blind here, so if this advice doesn’t fit your situation let me know and give a bit more detail to work with.

Polynomially yours,

– Bob