The myth of software time estimations
Sam Johnson

I share the sentiments, but disagree with conclusions. Thanks for the write up, and I’m sorry you got terminated for writing a blog article, that sucks!

It is true that few developers like estimations and accurately estimating software development is probably the next hardest thing after “naming things” and “cache invalidation”, but estimations are crucial for business to operate properly. At least in typical startup environments. You are referring to estimates being taken for investors, but it’s even more important for the internal company processes, like cross team projects. While I agree that thinking of “priorities” is better, estimations are still necessary when it comes to team collaborations.

Tech debt is a threat with inaccurate estimations. There also always exists some tech debt because no business operates in complete certainty. I would argue that it is okay to incur tech debt when you have to, but you have to take time to repay it soon thereafter. Budget that time into the next iteration/quoter/project. If the product and business teams are competent, they will understand that the cost of mainting tech debt over time is high.

Estimations won’t ever be 100% accurate, and they don’t need to be. Some stories will take more, others you will be done with sooner than you expected. It’s fine as long as on average it evens out.