His answer was, naturally Yes. What he said was that you should get a complete product backlog from the client and then you would estimate the work for the entire backlog and say to your client that the total points (either function points or story points) in the product backlog would be, say 450 points. Then in your contract, you would write that you will deliver a product with the features, which are 450 points worth in total.
Furthermore, you would add a few more clauses in the contract, which says the following:
- the client will prioritise the requirements so that the most important or most risky features are listed first.
- the client is able to change their requirements without penalty during the project, provided that the client reviews the existing feature list (or backlog) and remove or de-prioritise one or more features (or stories) having the same points as the new requirements, to ensure that the total points of the project remains constant. This is the 'Change for Free' process.
- the client is able to terminate the contract before the agreed end date, if you manage to deliver the product ahead of time or when the client deems the product is sufficiently developed for their needs, and it's not worth their time or money to spend any more. In this case, the client agrees to pay you 50% of the fee that you would have received if the project carries on to its full term. For example, if the project is 9 months long and the cost is $100,000, then if the client decides to terminate the project in 6 months, they need only to pay you 50% of the $33,000 remainder. This is what he calls 'Money for Nothing' principle.
To achieve this, you would naturally develop the product by adding features in the order provided by the client, i.e. the most important ones first, and you will provide demos (or reviews) of the product to the client at regular intervals (i.e. at the end of each sprint).
By developing using SCRUM with the most important/valuable features first, what would tend to happen, according to J. Sutherland, is that the client, in many cases, after seeing the product demos over the course of several sprints they may come to the conclusion that they don't need the 200 or so features (which equated to 450 story points) that they originally spec'ed out, but instead they only needed 80 or so of the most important features that they asked for and which you have already implemented, and that the remaining 120 features are of low priority and not worth the money they would have to spend.
In this case, you are effectively saying, "OK, if you want to terminate the contract, just give us 50% of what you still owe us and we'll go our own separate ways." This will save them 'some' money for not having the spend the full remainder that is due at the end of the contract, PLUS, you get half of what you would have got for not having to do anything.