From the Ivory Tower to the Muddy Trench

I’m not paying for that! February 25, 2010

I’m working on preparing a TED like session for an internal architecture summit coming up in about a month.  At first I considered the rather bland topic, “A Survey of Models”.  Clearly architects use models every day, whether they are analyst models, patterns or architectural, but this topic is unlikely to spark the kind of imagination that I hope for in this kind of event.

Instead I have settled on the topic “I’m not paying for that!”.  In this session, I plan to look at those aspects of a project which are required for it’s success, but to which your business doesn’t feel are necessary.  Things like – project management, architecture and documentation and the dreaded cost of maintenance and stabilization.

In each one of these cases the issue is not necessarily the idea of these things.  Who doesn’t want their project to be managed so that it finishes on time?  Why wouldn’t you want the service plan which ensures this expensive solution continues to run well into the future?  The central issue is that we, as a profession, have shown no these investments to have no material impact on the execution of the solution.  Nearly 80% of all software projects are late and over budget!  Is this because 80% have no project management?  I think not.  I think it is that because of poor planning and the ineffectiveness to plan and execute a project plan that these projects are too expensive.

Have we shown that spending $30/GB leads to more reliable solutions than the $100/TB that business customers are putting on their PCs at home?  We have proved this to ourselves and our own IT brethren, but it is up to us seasoned professionals – the architects that know how to speak the language of business and technology – to make the case the business.  Speaking their language, explaining and proving that the $30 a GB disk is worth the premium.  It is also incumbent on us architects to listen to our customers and communicate what is valuable to them back to our engineering teams.  Maybe that SharePoint site that hosts pictures from the company picnic doesn’t need to have RAID 1 / SAN attached / disaster recovery.  Maybe we can afford to offer a less expensive alternative with a less stringent SLA.

And so I ask you – what is it that your customer won’t pay for?  And what have you done to justify the cost or negotiate a different solution?