Many agile proponents advocate the “Cancel-After-Any-Phase” approach. Work is prioritised by business value and the customer can halt the project after any phase.  You can fix price or scope, but not both.  Most commonly, the price is fixed and scope is cut if necessary.

This approach is a natural fit for most agile processes.  However, other options should also be considered, because:

• Firstly, Alistair Cockburn points out that methodology and contract type are not tightly linked.  Like many people, I initially made the mistake of bundling “agile” together with “flexible contracts” in my mind.   He points out that the two issues are virtually independent.  His comments prompted me reconsider my own views, and to write about fixed price contracts here on this site.
• Secondly, I doubt that Cancel-After-Any-Phase, on its own, can satisfy customers’ concerns about cost overruns…

A Weakness in Cancel-After-Any-Phase?

Consider this scenario:

The project is about 25% complete.  It has become clear that the amount of work required will significantly exceed the initial estimates.  The customer can cancel after any phase, but will they really cancel at this stage just to get a “cheaper” supplier?  Will they carry on, and cut scope to fit the budget?  Can they really cut the scope that far, or do they have to find additional funding?  Is any additional funding possible?

Customers have a legitimate interest in protecting themselves from problems like this.  A good process will seek to meet that interest.

Conclusion

There’s a limit to scope cuts.  Cut too much and customers are bound to be disappointed.  Many customers can’t take that risk. So, it makes sense to look at other contractual options, such as Target Cost or Two-Phase Contracts.  These can be combined with Cancel-After-Any-Phase to offer better risk protection.

Finally, even the best contract in the world is no substitute for trust and collaboration.  That’s what we really need, and our contracts mustn’t stand in the way.

Background Reading

Martin Fowler, on fixing price but not scope.  This is Cancel-After-Any-Phase, but with a pre-arranged maximum expenditure.

By the way, he says you can’t fix both price and scope.  So why do I say you can?  Because Martin’s describing reality, and I’m describing contracts 😉  Traditional contracts are not about reality, they’re about a plan, and about making the supplier responsible for the inevitable difference between the two. That’s OK, as long as you understand the downsides.  See this page for more discussion of Martin Fowler’s comments.

(Update March 2011:

1. Glen Alleman has posted a perspective here, on the inadequacy of cancel-after-any-phase.  While his kitchen remodelling example might be a bit waterfall-ish, I agree with his general premise that,  “How much will this cost” is a legitimate business question.  Personally, I think it is best if both parties can agree that any answer to that question should be just an approximation, but the question itself is a valid one.  It should be asked, and it should be answered.

2. Ron Jeffries is quoted here as saying “The value provided should be so visibly more than the cost as to make detailed cost accounting clearly a waste of time.”  I’d like to point out the difference between detail and accuracy. Yes, we don’t need task-by-task actual costs, but we do need an accurate grand total.  The latter can be obtained easily, from total hours worked by each team member per week (as Ron himself goes on to say).
As for the total value being so clearly more than the cost, I’m not sure that’s necessarily an adequate solution to the cost management problem.  Sure, if the value is approx $50m and the cost is approx $1m, then don’t worry too much about the cost.  But what if the value is approx $3m and the projected cost is $2m?  The gap between those two figures is large enough to make the project worth doing, from a business perspective, but not large enough to let anyone relax about cost tracking.  If you don’t manage cost and scope carefully, overrun could eat the projected benefits.)

Finally

No contract can fully protect you from a supplier who makes promises they can’t keep.  Therefore selecting a supplier must be about finding a competent partner not just a low bidder.