[hluska]

I've found a middle ground between micromanaging the team and giving them enough autonomy to work independently. It works best on definable tasks when the remote team's manager is technical enough to do their job on her/his own.

Instead of just a requirements document, give them a requirements document but also define a set of red/green tests to show whether the requirements were met. It's hard to do this with front end, where feel is often very important, but in my experience, it works well provided that the remote team's manager is qualified to make the types of architectural decisions necessary to pull this off.

You can even start tying compensation into passing red/green tests!

[eloff]

That's not likely to be a good idea re the compensation. There are a lot of studies showing knowledge workers perform more poorly at the task with financial incentives.

[hluska]

This is either an unfortunate reality or a limitation of my experience, but I've only had to resort to writing red/green tests when my remote teammates have been at a level where poor performance would be something of an improvement. It's more useful for me when I see tons of commits, yet no forward progress on a measurable goal.