I’ve used a mantra many times before: If you’re the best person to do a job then you shouldn’t be doing it.
Often I find that a team has specialised knowledge locked up in too few people’s heads—sometimes that’s critical information that only one person knows. This is a problem for the organisation because it’s placing a lot of reliance on that one person. If they leave, or they have to work on something else for a while, then anything that needs that specialised knowledge can’t get done.
The answer, of course, is to share. But hand-over documents and meetings tend not to work. They are too theoretical, and reading about something on paper is vastly different from doing it for real. Usually when you do it for real a whole host of issues and assumptions suddenly present themselves, which would never be in a document or presentation.
So sharing that knowledge has to be built firstly on real work. Real issues that need the specialised knowledge have to be found—ideally it will already be part of on-going work. Then there needs to be a systematised way of ensuring the sharing takes place. Hence the mantra. A less experienced person must be required to do the work and then draw on the expertise of the specialist. Ultimately the skills get spread around.
An obvious example of this is pair programming, where the emphasis is constantly on sharing knowledge and expertise. Also, when I know a specialist is going to leave my organisation then I will try to make sure the specialised work they usually do is picked up by someone else… who should work with them. I imagine it would be useful if the specialist comes in for their last week wearing boxing gloves, specifically to stop them actually using the keyboard, although you’ll be pleased to know I’ve never actually implemented that rule.
Sharing this way can be difficult, so the culture needs to promote that kind of outlook. Ultimately it means everyone can learn and improve. Even experts will often tell you they get better at their subject by mentoring others.