Are you looking at your "return frequency" data?

Return frequency is probably one metric that is representative of customer satisfaction irrespective of the industry.

In a web based business, however, it is both, difficult to accurately measure return frequency and hard to utilize the data (provided we have a good sample) to promote our business.

In most reporting tools, return frequency is reported at site level - that is, how many times an average user is coming back to our site. And, therein lies the problem. First, it is only speaking about an average user. Second, it is only measuring the metric for the entire site. Unfortunately, there is no such thing as an average user, and more than likely, your site has more than a few dimensions. Consequently, a site-wide metric is only good for an "average user" - not a winning strategy for any business.

So, the first step is to adapt your anlytics software to show page and/or site section level return frequency. Then "de-average" the users into defined buckets, and classify site sections based on repeat visitation data. Combining these two insights will help you understand why a user is coming to a particular site section/page at your site. You can then tailor the experience to meet "most" needs of "majority" of users to the above site section/page. The reason I say "most and majority," is to emphasize that trying to classify 100% of your site usage will probably not yield best ROI on the effort invested. Picking top 5 or 10 site sections, for above improvements, will be a good start.

Do you have a specific question on this metric or others that you would like addressed? I promise to move it up my list of topics to discuss :)