So I stirred up a bit of interest with the ROI blog. It's always fun to do that by making a broad statement, but I don't think I fully made my point. Then again, I wrote the whole thing from 6am to 6:10am on the day we did our company ski outing :).
Let me clarify several items:
1. External communities are much easier to measure than internal communities. My blog was specifically in relation to internal employee communities. We have done and seen lots of ROI analyses for external communities, and it usually comes down to percent of offloaded support calls, increases in customer satisfaction, new leads generated, increase in purchases and increase in customer insights. I think these are very clear and measurement can do a good job of articulating the benefits.
2. I very much expect that we will be doing ROI analyses for our customers. It's a fact of the industry and we're no exception. My point was merely that it's very difficult to do it for internal rollouts without a lot of fuzzy areas. Even the best analysts will have difficulty measuring it. That said, there are some solid ways to get started: time employees spend researching an answer, the number of redundant questions, total output before and after, process time, employee satisfaction, etc.
3. Most of the ROI analyses I've seen for the team collaboration space have been half-baked, unless it was for a specific company. Once you get into the details of measuring employee efficiencies, it can get pretty sticky and it's almost always unique to the company, department and even employee.
In short, I don't at all doubt ROI analysis will continue to be a factor in all of our lives. What I have yet to see, however, is a compelling analysis that defines the benefits of Enterprise 2.0 on a broader scale.