Last weeks I spent huge amount of time negotiating with the renovation company about the details of our upcoming bathroom renovation. You know, all this bargaining, market research and choosing between a multitude of option takes time. To make things even more complicated our house manager happened to be very peculiar about what he can allow to be changed - it triggered yet another round of negotiations and renegotiations even though we signed the contract already.
The biggest problem is that there is a decent element of uncertainty. Some technical decision can be made only after they start works and break the walls. Therefore we have to nail down plan B and plan C well in advance not to be offered a bill with the "additional costs" after the fact.
Could we go the agile path to the possible extent (i.e. taking into account that hardware are harder to refactor, than software)? Technically we could formulate our goal loosely crush the walls, buy and install stuff, when we are sure we need it. I guess we wouldn't have to waster time on plan B and plan C - there is anyway very low probability that they will be needed.
But.. it would probably cost us a fortune. Not that renovators were evil. I don't think they would multiply the prices by 10, because they need to care about their reputation to a certain extent. However, they would probably double their non-transparent prices - it is difficult to argue when your bathroom is in the middle of renovation and we are anyway unlikely to be their customers again within several years at least.
Single-shot software subcontracting, especially to the offshore people you never met, often looks like the above renovation story. It is very difficult to go agile, unfix costs and scope, when working with the people for the first time. They meet you once in a lifetime, they need to earn their living and everybody knows that in SW world formulating precise and valid requirements is even more difficult, than with the house renovations.
Going Agile without the intention for long term relationship is possible. It just can be very expensive. Single-use trust is uneconomical or at least is rarely economical enough to resist the temptation of "small" overcharging.
You might find the following articles worth reading:
Are you an offshore development company or were you ever a client of such company? How difficult it was for you to apply agile methods for the first time?
Comments
Agile House Renovation
June 6, 2008 by Ilja Preuß (not verified), 23 weeks 6 days ago
Comment id: 1572
You might be interested in this interview with Alistair Cockburn - his experience is significantly different, or so it seems.
Interesting story indeed. I
June 6, 2008 by Artem, 23 weeks 6 days ago
Comment id: 1575
Post new comment