We've been pushing down to 8 hours as the max estimate, but I like Andy's idea of pushing down to 4 hours. The 8 hours came because we don't believe someone can accurately estimate a task of more than one day at a time.
The other thing we've been doing is "factoring" - applying a factor to each 4-8 hour task:
1. If you've done this task before many times, make your best
estimate and add a 10-20% fudge factor.
2. If you've done this task 2 to 6 times before, double your estimate
to allow time for errors that you can't anticipate.
3. If you've done this task only once before, triple your estimate,
allowing time for normal learning curve and the usual errors.
4. If you've never done this task before, multiply your estimate by
four times, to allow for the learning curve, errors, and more errors.I can't take credit for this, but I can't remember who to give the credit to either.
Mike Schroeder DonorWare LLC
