Low efficiency and flawed estimating of time

From my experience, creating software or knowledge work stuff is complex and has low process or flow efficiency if not actively managed. This is because we often underestimate complexity from the start — so our time estimations of work to be done is then flawed straight away.

The challenge of time estimating is that you estimate in isolation — yet work is being done alongside other tasks. Indeed, a massive percentage of worktime is often wait time: 85–95% can be waiting for other people, waiting for dependent work or even waiting for you to get back onto a task, which you’ve had to delay or prioritise because of other work.

This means, even if we aced the estimation of a piece of work, our estimates would only be accurately predicting on a low percentage of actual delivery time. If we implement probabilistic forecasting through Monte Carlo simulations however, we can mitigate this flawed thinking and estimations.