Wednesday, October 01, 2008

Boiling the SOAcean - anti-pattern

This is the SOA strategy that tries to much and sets its aspirations far too high. Often it starts as a little point problem before some one suggests that "we could do this as well" before long the SOA strategy is listing solving world hunger and the middle east peace process as secondary objectives that will come from meeting the now agreed two objectives of the project, interstellar space travel and universal harmony.

There are two basic causes of this problem. Firstly its powerpoint driven aspirational stuff with the people involved knowing they have no real worries about actually having to deliver the stuff they are writing about. Secondly its a belief that there is infinite budget available at least in the long term which is the key part of the this anti-pattern. Planning is for a future in which everything is fixed so you might as well get it down on paper today, this leads people to flights of fancy on trying to conceive of all possible outcomes and how SOA could be used to solve them. The centre to this anti-pattern is the lack of grounding that it has in reality, when combined with aspirational planning this leads to the sort of programme that is doomed to under achieve as it would be impossible to deliver.

The biggest effects are those of perception, firstly people will think that the programme is over stretching itself (they'd be right) and secondly will be disappointed when it fails to get close to its goals. The lack of focus in the aspirational planning often means that expensive investment is made in infrastructure for potential future requirements. These projects are often cancelled after making that infrastructure investment but not delivering the level of benefits expected for the spend.

The key to resolving this problem is grounding it in reality. Push the programme into tight iterations (no more than 3 months) with each iteration delivering quantifiable business benefits this helps to move the focus away from infrastructure and towards the end game and also helps to shift the event horizon of the programme into something more manageable, around a year is ideal. The next bit is to adopt YAGNI as a core principle, infrastructure is procured no more than 1 iteration ahead (i.e. if you are on 1 month cycles then you procure at 2 months) with the first iteration being a selection process.

The final piece is making people get their hands dirty, aspirational planning comes from people who deliver to powerpoint rather than to live, make them responsible for go-lives and it will focus their mind on achievable objectives.

Technorati Tags: ,


Daniel Spangler said...

This is a killer one here. I'm facing this one right now on one of my projects. It's not SOA, but still relevant. My "powerpointers" are the C-level types and they aren't being forced to stick to something by my director. The result is one project that has been in a "kick start" mode for a year now.

Steve Jones said...

Remember to make sure you have a preliminary before the pre-meeting Dilbert says so

Tough luck on the project.