During Web 1.0 and before that, companies had to go out of their way to collect feedback from their customers and when they did, it was in the form of customers communicating directly with the company. Web 2.0 changes all of that. Now customer feedback is an integral component of most social computing sites. And, the feedback is public. This provides great benefits to companies by getting more direct and regular feedback but it also changes the nature of the feedback process in that customers can also use the feedback from other customers directly as well. An article by Vicky Burger on Web 2.0 transformation addresses this topic.
However, things are not quite the way you may think they are in the world of social computing. Very few people are aware of a recent study which shows that less than one percent of visitors to social computing sites contribute in any way including providing feedback whether comments or ratings. So, the challenge here is that companies are now able to get additional feedback directly from customers but they may run in to two problems. The first is that unless the site gets a huge number of hits, the amount of feedback the site may collect will likely be very small (remember less than 1% contribute). The second, and perhaps even more difficult to deal with, problem concerns the likelihood of the sample being unrepresentative in some way. I haven't seen any research on characteristics of users who tend to contribute to sites versus those who don't and without that type of information it is difficult to determine whether the feedback a company gets in this way is representative or biased in some way.
Companies that are serious about making strategic decisions on the bases of direct customer feedback coming from Web 2.0 sites would be wise to adopt the best practice from market research of directly testing the non-response bias. The way that's done is to solicit input on the same questions from people who didn't contribute the information unsolicited and then comparing the results to test for a systematic bias. Without a specific focus on this aspect of Web 2.0 user feedback, companies could find themselves honing the designs of their sites based on the feedback of the small group of people who contribute feedback and these changes may not be optimal for the entire population of users who the company is trying to satisfy.