Even More on SOA as a Corporate Responsibility

I took some well deserved flak on my original post about SOA as a Corporate Responsibility.  Based on the responses, my message clearly didn't get through. For example, Lars Olufsen commented that the "tactical focus" that he thought I was proposing was off base. Here is an excerpt from the his comment:

"Shaping your entire organization to support [SOA] takes longer to reach ROI and it WILL take a period of zero added value and quite a bit added cost..."

Here's the thing to remember: The average CIO changes jobs every 3 years.  And, when you get a new CIO, one of the things the CIO will do is take an inventory of all the initiatives going on, to decide what to do with each one.

If you haven't shown quantifiable business results with your SOA initiative, the CIO really has only three options:

1. Cancel the SOA infrastructure planning and initiative altogether.
2. Spend valuable (and still tenuous) political capital convincing the CFO that this existing RIO—less SOA initiative—is the right horse to place a bet on.
3. Put someone else in charge of the SOA initiative who the new CIO believes can restructure the project to make SOA openly successful in the short term (proving the CIO's leadership made a difference).

If I were a betting man, I'd bet that 90% of the CIOs today would take the 3rd option. Why? Because they know they need to do service-oriented architecture, but they also know their butt is on the line if they don't shake things up and "fix" the "problems" created by the last CIO.  Starry eyed idealism about long term initiatives that cost a lot and don't show ROI for years will get you, at best, to your next CIO but not any further.  You may not want to hear this, and it may not be "fair", but that's reality.

SOA What? Let's look at two different paths to SOA. The first path to SOA success:

• Take 10 years to "finish" your SOA project (by whatever measure you consider it "finished" -- which may be "it's now a way-of-life for every project");
• Show clear business results after 5 years; and
• Spend $2M per year for the first 5 years.

The second path to SOA success:

• Take 15 years to "finish" your SOA initiative;
• Show business results after 1 year; and
• Spend $2M per year, for the first 10 years.

Clearly the first path to SOA is more "optimal" because it takes 33% less time and costs 50% less.  But, every CIO I know would choose the second path in a heartbeat.  Why?  Risk mitigation - both budgetary and political.

With the first path you need to invest $10M before you know whether you've succeeded or not.  Basically, you have to make a $10M bet. Oh, and don't forget, the CIO will have been fired long before the company sees that benefit (CIOs are well aware of the average tenure of their job).  If anyone reaps the benefit, it will be the next CIO, not them.

With the second path, the CIO only needs to make a $2M bet (5x smaller), and results will be shown within the CIO's tenure.  Heck, good results might even extend the CIO's tenure!  Spending money on something that is proven is a no-brainer for a CIO.  It's making expensive long term bets that put you in quicksand.  This is why CIOs would choose the second path in almost all cases.

Your SOA initiative is no doubt long term and strategic.  But to succeed with it in the real world, you need to execute tactically and show results at every step, while continuing to move the puck (yes, I'm Canadian) in the right direction for the long term SOA strategy.