The expensive costs outside of estimates

on daily, tail, costs, and estimating

Are you missing the expensive hidden costs in your story estimates?

I have been reading ‘No Estimates’ by Vasco Duarte recently and came across a concept that has been in my head for a while, but I have been unable to put into words. Originally thought of by J.B. Rainsberger called accidental complication, it is complication that creeps into our work due to organizational structures.

Vasco goes on to share an email from Troy Magennis. That says that people estimate based on in a perfect world, as we know that is not the case in the real world. Work gets blocked for multiple reasons (like holiday, or they are working on something else). He also says that we should be thinking about the entire system not just the perfect development time.

This is important if we start to count “man-hours” on a story. Where “man-hours” means the time spent on a ticket by all humans, if we have 8 people working on it for an hour that is 8 “man-hours”. If we have one person working for 8 hours that is also 8 “man-hours”.

For quite a few tickets the “man-hours” involved in getting a ticket ready to develop can sometimes be a multiple of the amount of actual work required to fix the problem in code. The same can be said of the release process. However, these are not included in the estimates.

This becomes especially important if you work out the cost of a story point. For one project we found out that the average cost of a story point was around £1000. This blew mind and the teams mind. We thought “why does a spelling mistake fix cost can so much money?”. What we didn’t consider is the work either side of the few of minutes it takes to make that code change and commit it.

It seems that systems have a minimum cost of interaction however big the change is and sometimes the system cost either side of the code change can dwarf the size of the actual change. These are often thought of as constant and unchanging but with thought the team can often be improved but first it needs to be raised to the team.

I would recommend that teams get told how much a story point roughly costs since after hearing that our team changed how it approached small items and the conversations with stakeholders. As we realized that talking about the small tickets was costing a significant amount, we cut sort analysis on simple tickets. It also meant that these costs become known areas at retrospectives as areas to improve upon.