It’s anybody’s guess
Estimates, not guesses, are key to controlling development
Comment If you were to ask me how long it takes me to get to work each morning, I would tell you, “It usually takes about an hour, door-to-door.” I’m pretty sure most of you could also give a reasonably accurate description of your journey to work, too.
It’s worth bearing in mind the use of the word ‘usually’ in my answer. It does usually take me about an hour to get from my home to my office but occasionally, too occasionally for my liking; it can take a little longer. Not withstanding mishaps, however, my experience tells me the journey will be about an hour and I can plan to leave home about an hour before I need to be in the office and expect to be there on time.
This is because I know to a high degree of certainty the value to assign to most of the variables involved in the journey. I have experience of travelling to my hometown station; riding on the train; and getting from the destination station to the office. In other words, when you ask me how long it will take me to get to work I am estimating under conditions of certainty.
There are three conditions under which we can attempt to estimate anything and these are certainty, risk and uncertainty. We’ve already discussed certainty and under these conditions in software development, we would be familiar with using the technologies and we would have experience of the domain too. Estimating under these conditions is quite simple and relatively easy.
However, unlike with my train journey, it is very unusual for software developers to be asked to estimate from a position of certainty. To gain a competitive advantage, organisations need to be delivering new and, hopefully, unique functionality. "New and unique" is, by definition, something we have not done before and have no experience of. It is, therefore, unknown.
When we’re trying to estimate with unknown variables, we are estimating under conditions of risk. Under these conditions, we are certain about some things but not about others. For example, we might be planning on redeveloping our existing application in a new language. Obviously, in this case, we’d be familiar with the domain but not with the technology, so that is where the risk would arise. Estimating under these conditions is complex and difficult.
The very last condition we might be asked to estimate under is uncertainty and this is where we are unfamiliar with both the domain and the technology. It is very tempting to leap into this type of project just for the learning experience or the sheer thrill of doing something novel but, ultimately, it is doomed to failure. These are chaotic conditions to say the least and any estimating you did here would only be guesswork.
A frequent mistake made by developers and project managers alike is to imagine that they are estimating from a position of certainty when they are really estimating under risk or uncertainty. It is a human failing to believe you know more than you actually do or to over-extrapolate from your experiences. Guessing is easier and requires less effort than making a proper estimate. So, what should we do when we’re asked to produce an estimate and we are operating under risk or uncertain conditions? Well, it may sound obvious, but if we can bring the risky or uncertain variables into the zone of certainty, we can estimate with confidence.
Sponsored: Network DDoS protection