The distributed teams question is always an interesting one. We did one card,
The Only Agile Tools You'll Ever Need, which ruffled a few feathers. Basically we suggested that moving in the direction of more distributed is moving in the opposite direction from agile values--primarily the value of "high levels of face-to-face communication."
Tim's right, there are many technical solutions that help. But recognize that you are making a concession to a sweet spot that can really work well, and the technical substitutes pale in comparison to just being there. The better answer is "shift your teams so they don't have to be distributed." One team in Dallas, one team in Krakow, one team in Bangalore, etc., instead of three teams each distributed across three geographies.
Agile aside, there are studies that show it takes years to obtain true positive return on investment with a distribute teamd. While the bean counters looking only at hire rates may quibble, it takes a long while and a lot of wasted effort to work through the inevitable issues of distributed teams (which do include quality).
As a remote guy for a year, I realized there were so many things I missed out on, to the point where I felt I wasn't as real a part of the team. It's also a lot harder to obtain consensus and enact lasting, positive changes in such a team.
You can make it work, and Tim provides a few suggestions. (Cameras are essential, and not just face-to-face cameras--you need to be able to see the teams as well in their work areas and in meeting spaces.) But is that really what's best for your company and product?
Regards,
Jeff