Estimation biases

Posted by Feisal Adur on March 5, 2015

One of the most toxic things in a project is estimations. Yet it's the most cast around notion that at any point during a projects lifespan, anyone is capable of estimating the amount of work left.

Obviously time lines get thrown around and projections can be made. But then again, this isn't the statistical "bias of an estimator" equation we are discussing. But when accurate human computation is applied to estimating a task and the amount of oberservable time that is left until the objective is met. One thing is usually left out. The frame of mind the estimator is in.

From my personal experience I tend to think hard about what steps are involved in solving a given task. Usually am not that far off mark. And yet it still ends being completely wrong.

The crucial mistake here is that I haven't taken into considerations any of the other projects, sidetracks and interruptions that take place in this exact same time frame. The estimation bias, is when you make projections and only consider it an isolated "thing". It never is.

The estimation bias comes naturally to all of us. None of us can predict the future, so the framework we use to estimate is entirely based on the current situation. This isolated vacuum in time we find ourselves in. And we try to give ourselves some wiggle room and still end up blind sided.

Estimations are important and we can't plan anything without them. So instead what I've been doing as of late is. Together with the stakeholders, reassess my estimations as often as possible. Sometimes more than once a day. Why? Well the perception of what the project is needs to adapt a new reality. Because sidetracks, meetings and interruptions are now part the project as well.

Worse is when estimation biases are used to measure progress. Progress is one thing and meeting deadlines is completely different beast. But that's for a another post.