Couple of good points from the ZDNet article by Toby Wolpe: Jive Software: Where firms are going wrong with social rollouts. First, choose strong, well managed site/team to start with:
Proponents of collaboration tools are sometimes overly ambitious and choose the wrong place in the organization to exploit it initially. They saw the things they wanted and they would want to start in one of those areas they knew the company had a problem. The trouble with choosing one of these problem areas is that by their nature they are often failing precisely because of poor controls, measurement, and management. So even if you improve, no one's going to believe it.
To get started, start in a part of the business that's actually important to you. Start in a place where you measure something today, where you know exactly what the baseline is - because if you don't measure it today, it's not that important to your business. Starting with an area of the business with sound measurement already in place, such as manufacturing or sales, provides a solid foundation for subsequent deployments.
Once you have enough people on the platform, a lot of these more aspirational areas - where information finds you and people feel more connected and you start using it for other parts of the business - those start to happen.
Second, start with very clear purpose:
The second mistake in deploying social software avoids this first error of picking the wrong business grouping but perpetrates another by drawing too widely on participants from across the company and choosing employees with little need to collaborate with one another.
Collaboration without purpose makes no sense. You've got to have a job you're trying to get done or something very purposeful you're trying to do if you want it to stick and get the value.
For the closing - easy to understand video on basic principles of collaboration: