I agree that you can add tasks up to arrive at the effort (man-day) but how about the time? Most projects will have tasks that run in parallel (e.g. front-end and back-end). Can you explain your formula to get the calendar days estimate?
Think about the initial estimate as a "high level design."
It doesn't have all of the details. You still need to map out the dependencies figure out what you can parallelize, build a (gasp) Gantt chart, etc. Each additional investment in planning increases the resolution and accuracy of your estimation. You just have to make sure that each additional investment is worth the effort (it often isn't!)
In the spreadsheet, I also share how to compute a conversion factor from net development effort to calendar time. This is a "low hanging fruit" of known knowns that are too often ignored. Yes, people have meetings. Yes, there's an on call rotation. Yes, people take time-off. All of this is known, so it's insanity not to account for it.
I agree that you can add tasks up to arrive at the effort (man-day) but how about the time? Most projects will have tasks that run in parallel (e.g. front-end and back-end). Can you explain your formula to get the calendar days estimate?
Think about the initial estimate as a "high level design."
It doesn't have all of the details. You still need to map out the dependencies figure out what you can parallelize, build a (gasp) Gantt chart, etc. Each additional investment in planning increases the resolution and accuracy of your estimation. You just have to make sure that each additional investment is worth the effort (it often isn't!)
In the spreadsheet, I also share how to compute a conversion factor from net development effort to calendar time. This is a "low hanging fruit" of known knowns that are too often ignored. Yes, people have meetings. Yes, there's an on call rotation. Yes, people take time-off. All of this is known, so it's insanity not to account for it.