Just as understanding the science to music does not make you a good musician or allow you to build a great band, understanding the science to teams does not make you a good team builder.
I look at these things as performance art. Theory goes in the background when you perform all that is left is execution.
It's interesting. I used to work in digital agencies for quite a number of years and the pattern you see is that for a project "the management" composes a team based on who's available ("resourcing" they call it.)
A lot of projects went over budget or were delivered very late and over time I came to notice that the projects that weren't late or over budget were executed by people who'd worked together before on two or more projects. It meant that this team had already gone through a bonding experience and that their lines of communication could be kept short. I referred to this at the time as "the band" since it was like playing in a band. If the chemistry matches between people then communication will happen automatically and magic "just" seem to happen.
In professional services, the economics of team building are much more challenging, given the short horizon of teams, both due to projects duration and due to high attrition rates. The traditional approach is to build teams of equals or of the same (c.f. same colour business suits in old IBM.) It did not work very well, but still it was good enough until it created issues with people not being willing to even participate in such organisations (c.f. casual Fridays.) Even with a complete framework (we have a psychometric classification across different personalities and working styles) this can be applied best when the Business does not go well; bad environmental conditions tend to make team dynamics worse and you cannot therefore see the improvement. Otherwise, as per your experience, teams are built based on availability with the expected random outcomes and increased role for the structural aspects of the team.
> "Just as understanding the science to music does not make you a good musician or allow you to build a great band..."
I'm not a musician, but I disagree with the first part of this statement. My impression has been that understanding the theory can make the difference between a good musician and a great one.
One could apply the same to programming -- understanding the fundamentals of computing is not necessary but it can mark the difference between good and great.
I'm a musician and a programmer, with minimal theory knowledge but extensive work experience in both areas. I can do productive work, but at the same time, can see my own limitations. For instance, I would have a hard time organizing a large scale effort in either area: In music, that would be something like creating a long composition or arrangement involving multiple musicians. In programming, it would be coming up with the architecture for a big project.
Still, in a vein similar to a comment that I made about education yesterday, I wonder if there is such a thing as a "theory" of the formation and administration of work teams, i.e., of a level of reliability similar to music or programming theory.
I look at these things as performance art. Theory goes in the background when you perform all that is left is execution.