I got into an argument with a manager. The manager had a big vision of DigitalTransformation[tm], wanted to do a big project to implement this dream and required the team to expand rapidly in order to do so. My attitude with projects is to start small and have a workable feature as soon as possible instead. I tried to explain this to said manager but I couldn't get through to the guy. The manager was from an expensive strategic consultancy firm who specializes in powerpoint and reading gartner. This means I couldn't blame the guy too much for his lack of delivery experience but it got me looking for a simple example or calculation to help convince the anybody else that his idea might be a very bad idea.

So let's model the decision a bit.

Let's assume that we're an online business and we need to build a recommender for our website. Let's also assume that you're given two options, the assumptions of which you can change by dragging left and right.

  • a recommender that takes months to make and will cause extra money each day
  • a recommender that takes months to make and will cause extra money each day

With these assumptions it takes about months before both options earn the same amount of revenue (see chart). If you go for the shorter project that means that you have earned yourself months of free time to iterate. This time can be spent on other revenue generating projects.

This argument does not always apply (otherwise tesla would be making faster horses) but for many IT projects this is something that needs to be taken into consideration. Short goals are a good idea because you can benefit from it quicker. Even if there is a risky project you might want to do within your company, consider doing a few short ones first. If you can do a few short projects succesfully first, you may have earned yourself enough "free" time to venture towards something more risky.

Sofar I've assumed that we have perfect knowledge of the future. In reality, we won't know the expected benefit (or even succes) of a project and neither can we pin point the amount of time a project needs. I've also not considered marginal effects of effort. I would argue though that these modelling flaws only add to the point that short term projects are usually better.