In Thinking for Living, Thomas H. Davenport argues that managers of knowledge workers should be paid a premium for assuming the position as manager and letting go of the relatively safety and security of their individual contributor knowledge worker job. His reasoning is simple - managing and organizing knowledge workers is so vital to their productivity - self-organization and empowerment only go so far then management has to step in. However, the first management level job requires the individual to take a huge personal risk - abandon the skill set that made them successful and learn a whole new skill set as a manager. In order to attract the correct candidates in to management goes his thinking, it is necessary to pay a premium. How much of a premium is hard to say but 10%-20% would seem appropriate.

The interesting thing is that compensation professionals tell me that a premium salary is not supported by the market for managers in software engineering. I think that there is one main reason for this and a number of secondary reasons that might indicate a root cause for the problem. The first reason is basic economics and the fact that uncertainty in the nature of the work forces managers to maintain a flexible workforce of contingent labor (contractors). This is typically 10%-50% of the resource pool. The contingent nature of contract labor requires that a premium is paid for the hourly paid contract labor. Good people, confident in their technical skills and their ability to renew and refresh those skills regularly can earn a premium as contractors. Often earning more than middle managers and junior executives. Put another way, a geek can earn more as a contractor than he or she could make suffering through 10 years of climbing the corporate ladder as a manager. Hence, permanent full time individual contributor knowledge worker jobs fetch higher rates too. And as such, there is no premium for management. In fact, the market would suggest that managers should really be paid less!!!

Clearly this is a problem! If knowledge worker productivity is primarily a factor of effective management - and Barry Boehm made a science of this discovery in software engineering - then there is a conflict when the open market will not remunerate managers appropriately. What is causing this conflict?

I feel there may be a number of reasons and feel free to comment and add to my list [Update: Aarrrggghhhh, Murphy's Law - this would have to be a post where that bug in Haloscan kicks in and the comment option is unavailable. If you want to leave a comment use the comment box for the previous post. Thanks.] First, I feel that geeks always look for a technical solution to a problem, before they will pursue a people or process solution to a problem - in other words, tools over operational innovation or sociological or psychological changes within a workforce. If the answer is always to deploy a new software-based tool, then the demand will always be for high-end geeks who can make the best software. Secondly, technical innovation and problem solving is valued over operational innovation and an ability to quantitatively management to a plan. All that tribal individual value is bundled in to an ability to create great product innovation and solve significant problems in computer science rather than process innovation and the soft skills it takes as a manager to motivate a team.

In conclusion, I feel that if we are to deliver on Davenport's vision that good management will come from offering a premium for knowledge workers to make the leap to a new skill set then we must first start to value management skills more highly. In order to value management skills more highly, I believe that we must embrace Barry Boehm's observation from 25 years ago - poor management can increase software costs more than any other factor. So far, we're an industry in denial of this basic truth. Until we face our own brutal reality - that good management works and bad management hurts - then there is little hope for fixing the situation.

Related blog post: Social Networking for a Living, Why Good Managers Still Matter to Agile Development
External Links: Crosstalk article from 1996, Management Impact on Software Cost and Schedule  Technorati tag: Agile, David+Anderson, Software+engineering, Thinking+Living, Thomas+Davenport, Management, Knowledge+Worker