The processes, practices, and methods used for any UX Improvement project are grounded in User Centered Design (UCD) philosophy.
That said, there isn't a one-size-fits all method. Each project requires its own special attention, but in general, research and testing go a long way in rendering a successful design. That tenant is one explanation for the two month design cycles (which I describe in "The Approach"). Designers need time to get into the guts of a design and go beyond just moving pixels around.
The general idea is to break off a chunk of the product and work through it over a period of time, ending with "implementation ready" deliverables.
To be more specific, I'll work with the community to spot areas of the product that need UX help and will put together a design project to tackle the problem. A project might include re-designing an existing tool, working on a fresh design for a new tool, or as in the case of the first project, Improving the CLE 1.
Once a project is outlined, it will theoretically live on ad infinitum (or until design is no longer needed). Actual work on any project will take place over a series of scheduled two-month periods. These two-month periods (or design cycles) will consist of a flurry of design activity ranging from research and analysis to interaction design to usability testing to everything in between.
To keep things simple, the two month cycles will be more of a guide than a rule. Some projects might take less time while other might take more. But in general, all efforts will be made to keep things moving a long at a consistent clip.
Once a design cycle starts, production will consist of an 80/20 mix. 80% of the effort will go toward the current cycle and 20% will go toward the previous cycle.
The net result of the 80% effort will be a set of implementation ready deliverables – or a UX kit, if you will. The kit will include virtually every screen needed to define the design. Think of it as a blueprint containing screenshots of what each part of the UI should look like. The kit might also include sitemaps, diagrams, and other paint-by-numbers artifacts that will help any developer who's interested in picking up a design an running with it.
The 20% effort will be used to help those brave souls who actually pick up a UX kit and decide to implement it. While they'll only be given 20% of a designer's time during a cycle, they'll get 150% attention during that time!
Design support will include: calibrating the design to solve incompatibilities discovered during implementation, production of graphics and *other elements, one-on-one working sessions, mentoring, advice, testing, etc. Whatever you need, just say the word.
For more information on this approach, check out the UX FAQ, where I try to address any concerns you might have.