Sunday, 31 July 2016

More on estimates - From upcoming book Agile 101

When we talk about estimating using Agile techniques, some things are difficult to accept for those of us who have been planning and calculating timing and deadlines for a few years.

We can find several techniques for Agile planning: Planning Poker, clothing sizes for task classification (S, M, L, XL, etc.) or Team Estimation Game, but, let’s face it, they’re difficult to integrate in a traditional development environment, and even harder to integrate for the clients who purchase our products (or at least, for the clients’ I’ve been able to work with).

We’re not used to seeing a team “playing cards” for a session of Planning Poker (at least, I am not used to it). I do, however, agree with the philosophy behind those techniques. I believe that they’re becoming more and more popular partly because it is necessary to dispel some myths regarding estimations. Let’s try and do just that:

We do estimates wrong

Can we do them right? To estimate is to predict how long a task will take or what materials we will need for a certain job. When we do an estimate with high uncertainty, as for example when we use the data included in a public-tender document, it’s not so much an estimation as a bet.

We need estimations—that’s obvious. We must try to predict what will happen, so we can make decisions. But unless we’ve gone through that specific task many times before, with the same team and under the same circumstances, we need to be aware that our estimates will very likely be wrong.

The more effort we put into an estimation, the closer we’ll be to reality

Estimations have a cost and that cost is high. When we don’t know every specific detail of every bit of the work we’ll be taking on, does it make sense to spend hours and hours to guess about tasks that we do not know very well? When you finish your next project, do this exercise for me. Review the time spent on every task. You’ll see that you estimated 20 hours on tasks that took you 100, and you’ll find that —with luck— you spent 20 hours on tasks that you thought would take you 100. You’ll see that there were some tasks that no-one marked as done. They simply weren’t necessary. However, hours are spent on tasks that were added later and that no-one considered necessary in the initial estimate.

We’ll estimate better with this or that technique

There are many estimation techniques but very few will be as useful as comparing the time spent in similar projects, the decomposition of the work in smaller tasks, or the conclusions of experts (it would be better if we have several of these). The fact that we’re using complex estimation techniques could give us a false sense of security about our estimate, and prevent us from striving for higher reliability. That could actually cause us higher costs (see above) where we invest time that would have been better spent doing something else (like reducing uncertainty).

You can find texts like this and many other about how to manage agile projects in my book Agile 101: Practical Project Management (available on Amazon).

Translation by Begoña Martínez. You can also find her on her LinkedIn profile. Proofreading by David Nesbitt.

No comments:

Post a Comment