It's clear that a mission critical change needs oversight and risk mitigation. But think of the waste of this process from the Lean perspective: Waiting, Handoffs, Checking, Inspecting, Obtaining approvals, Reviewing, Filing, and Rework.
IT groups need shepherds to be able to guide these things through the system of checks and balances. If you look at this scenario, Ed, the programmer, was the orchestrator. The end-to-end process was ad hoc, and not designed purposely. There were actually several processes at play, designed by different departments (IT demand, support, delivery, QA, etc.), with different value systems, and the interaction between them isn't well defined towards the end value of "shipping software changes that work". The division of labour across review, testing, etc., without incentivizing end customer results, and minimizing waste, has devolved into a bunch of school marms that redline all documents or code given to them without actually working to help with the end goal.
The "this test plan isn't good enough" for example is a near and dear to me. Why can't QA dedicate a resource for a brief period to help make it better? Or at the very least, give guidance on the what they want to see for approval? Usually this (and ornery change management boards) wind up being the primary source of senior management overrides - the QA group doesn't improve quality, it just slows down change thus maintaining the current (functional) mediocrity.
IF the process was shepherded by a manager to actively involve QA, Code Review, Change Management, etc. earlier, Ed might be less frustrated, and it would have been done in 3 days. ;)
It's clear that a mission critical change needs oversight and risk mitigation. But think of the waste of this process from the Lean perspective: Waiting, Handoffs, Checking, Inspecting, Obtaining approvals, Reviewing, Filing, and Rework.
IT groups need shepherds to be able to guide these things through the system of checks and balances. If you look at this scenario, Ed, the programmer, was the orchestrator. The end-to-end process was ad hoc, and not designed purposely. There were actually several processes at play, designed by different departments (IT demand, support, delivery, QA, etc.), with different value systems, and the interaction between them isn't well defined towards the end value of "shipping software changes that work". The division of labour across review, testing, etc., without incentivizing end customer results, and minimizing waste, has devolved into a bunch of school marms that redline all documents or code given to them without actually working to help with the end goal.
The "this test plan isn't good enough" for example is a near and dear to me. Why can't QA dedicate a resource for a brief period to help make it better? Or at the very least, give guidance on the what they want to see for approval? Usually this (and ornery change management boards) wind up being the primary source of senior management overrides - the QA group doesn't improve quality, it just slows down change thus maintaining the current (functional) mediocrity.
IF the process was shepherded by a manager to actively involve QA, Code Review, Change Management, etc. earlier, Ed might be less frustrated, and it would have been done in 3 days. ;)