Re^2: Interesting insights from Software Estimation: Demystifying the Black Art

[dpavlin]

I had similar process when finding out my estimation factor. I also notices that my (positive) estimates are three times too short.

However, when I need to estimate time for group of people, even if I know them really well, and can estimate time for each of them individually, I would probably take a optimistic guess, multiply it by three and than take next time unit.

Yes, convert hours to days, days to weeks, weeks to months... Months would go to quoters or years, but you would already know that :-)

This technique proved very successful in estimating time for small team of 7 people with various skill levels. It also leaves enough time for testing, documentation and deployment :-)