Before I became a consultant all I cared about was becoming a highly skilled programmer. Now I believe that what my clients need is not a great hacker, coder, architect... or whatever. I am more and more convinced every day that there is something of greater value.
Everywhere I go I discover practices where I used to roll my eyes in despair. I saw the software industry with pink glasses and laughed or cried at them depending on my mood. I was so convinced everything could be done better.
Now I believe that what my clients desperately need is finding a balance between good engineering practices and desperate project execution. Although a great design can make a project cheap to maintain thought many years, usually it is more important to produce quick fast and cheap, just to see if the project can succeed. Before that, it does not really matters that much if the design is cheap to maintain, after that, it might be too late to improve things.
They need people who get involved, who do some clandestine improvements into the project without their manager approval/consent/knowledge... because they are never given time for some tasks we all know are important. Not all good things can be done, some of them must come out of freewill, and some of them must be discussed in order to educate colleagues, managers, clients and ourselves.
Now my big question is. What exactly are the skills and practices aside from great coding that can provide real value to the economical success of software projects? (and not the software architecture alone)