Tagged with brook’s law

Brook’s law: revisited

Ha. I went and read the Mythical Man Month again. Found the spot where he speaks about the effect of adding more people. Then I went and read comments from you guys and thought.

The result is that, indeed, if I dial the effects really high, the project indeed moves slower than it was before. For a while. The critical effects I found were:

  • Productivity drop. The rest of the team needs to effectively to focus on teaching the newcomers
  • Hiring speed. The speed at which the new people come on board has an interestingly strong effect. When the new people get added gradually, the resulting impact is way smaller
  • Error rate. The rookies need to make a ton of mistakes for the result to show

Actually, this is pretty much exactly what Brook says. And actually my model behaves exactly like Brooke says, too. You see, he talks about a late project. A project that is either close to or having already passed a deadline. And if you add so late in the project that it ends before they learn the ropes and the productivity gains show, he is right. In the long run, though, adding more people will work out.

What a relief! Both me and mr. Brook were right! Not that the latter would come as a surprise, though.

Oh, and I have already some tasty numbers for this weeks episode, so stay tuned and observe SD in action!

Tagged ,

Brook’s law

There really is no other way of saying this but I was wrong. I went into this being fully confident that it’d be simple to show how adding more people to a project can make it take longer (which is what Brook stated in his legendary Mythical Man Month). Well, in my model, it doesn’t. Whatever I do, I end up with a work-to-do graph that shows a minuscule blip above the normal behavior when people are added and then a rapid decline to a much faster project end. Mind you, doubling the team size still does not halve the project duration but still. What I get is something like this:

People, obviously, are added at week 105. By varying the number of people added, the the effect they have on productivity and the way they change the error rate, I can change the shape of the curve but it inevitably crosses the blue line (the scenario without people being added) at about the same point.

Well, whaddayaknow. I will be travelling this weekend and settling in next week so am not sure if I’ll get to that but I’d really like to find out what the hell happened. I’ll read The Mythical Man-month again. I’ll look at the model and play around with it. It might be that I neglected some important point Brook is making like add-more-people-productivity-drops-add-even-more-people feedback loops (although based on current results there is too little of effect to trigger that). It might be that I’m just interpreting the output incorrectly or that there’s a bug in the model. In any case, I’m baffled. Which means I’m learning. Which hopefully meant you are learning as well.

Talk to you next week! Take care and enjoy System Dynamics in Action!

Tagged , ,