Stop Trusting Estimates Now: See Why
Please stop asking for granular estimates. The are typically useless, and sometimes harmful!
Let’s breakdown the true request and think about what actually is being asked.
“What’s your estimate for this?”
This question requires background to answer:
- How long did it take the last 5 times you did this? Can we use this as an average?
- What’s different this time that will alter your estimate?
- Is there external pressure pushing you towards a certain result?
The big question that is not on this list is “what do you not know that will affect your estimate?” This question is significant because it might be the largest gap in your understanding, and you cannot answer it beforehand by definition.
These questions point at the core of the issue of estimates. They show that the task is either trivial, we have all the answers, or unanswerable. If all you have is trivial problems, you probably shouldn’t be making the software cause something else out there probably already solves your problem. If the questions are unanswerable, then you can’t answer it and you can’t give a good estimate.
So, if you aren’t going to use estimates, how are you going to hit a budget or deadline? We’ll answer that tomorrow…