Link Details

There are several things you can do. One is to fix the scope at a high level but leave out the details to give yourself as much wiggle room as possible. This is where the notion of card trees in Mingle came from. What you are doing then is setting a budget, and during the project the client spends it on "details" that give them the top level outcomes. The risk is that you can not be creative enough to achieve those outcomes with the given budget or that the client won't agree that the outcomes have been achieved with your cost-limited solutions. (There's no getting away from the client relationship management needed). Another thing you can do is agree a much smaller scoping project where you determine how productive the team would be in that environment (story points per iteration) and roughly agree on what a story point gets you - find some good example stories and their size. Depending on the level of trust with the client, they can they either buy a fixed number of story point from you without knowing in advance what they will be spent on, or you can do some upfront analysis to create an initial story list with story point estimates. The contract can then say you'll deliver those stories or ones of equivalent size with some basic rules on how to swap stories that won't commercially harm you and won't stop the agility of the project. There are lots of variations on this too - get in touch if you're interested.,