The Timing of Requirements

Monitoring the customer's wish list (or the backlog of requirements) is an important part of a project manager's job. Agile methods have taught us to defer decisions and to keep inventory low, meaning that you should work only on a few requirements with highest priority, and keep other feature requests at bay when people are still uncertain about their necessity. However, you must take care not to ignore the other part of the story: don't wait too long!

I'm a book maniac. My Amazon wish list is like a second home to me. I buy so many books that I'm surprised Amazon hasn't yet named one of their buildings after me. This week my shopping spree has reached a new climax, with 19 new books now on their way to The Netherlands. (I apologize for any problems my order might cause in regular transatlantic air traffic.)

The reason I placed such a large order is that, for a couple of other books on my wish list I simply waited too long! Some books I wanted appear to have been sold out by now. This means I will have to buy them used (which I hate) or I must try to find them on other markets, which is likely to cost me more. And for other books on my wish list Amazon has increased the price significantly, either because the books are becoming rare, or else because because Amazon is evil-minded and they know –from my earlier shopping behavior– that I'm going to buy them anyway.

In software development projects we can experience similar situations. I have seen the licenses of 3rd party components suddenly becoming much more expensive. This usually happens two days before we decide to buy them. Other times the implementation of features cost us significantly more time than they would have cost us originally, because of new interdependencies with other parts of the projects. In these cases a different order of implementation could have saved us a lot of time.

It is often hard to determine the proper timing for anything you do. I not only buy my books too late. I usually buy them too soon. It's a pity I don't have a fireplace, because there are a couple of books that I would gladly throw in. (I would even want to piss on them first, but that would make them harder to burn.) Unfortunately, based on raving reviews –apparently written by raving lunatics– I had already bought the sequels to these soon-to-be-burned-and-possibly-pissed-on books. A clear case of bad timing on my part.

Monitoring a wish list (whether it is a requirements backlog or a wish list on Amazon) is like monitoring a stock market. Acting too soon is costly, but you must take care not to underestimate the cost of acting too late.

Tell me, when did you act too late?