Scott Ambler has posted a follow-up on his previous article about fixed price software projects at

As a quick summary, he gives five recommendations:

  1. Give a ranged estimate (+/- a certain percentage or amount)

  2. Do some upfront agile modeling (using people who will actually do the work)

  3. If the customer still insists on a “precise” estimate, pad the number as much as possible to account for the risks

  4. Fix the price, flex the scope (anything added requires removing something of equal effort)

  5. Stage the funding, based on real deliverables (working software)

More to come on this topic…