After having used mybloglog and 103bees for quite a while to dig into blog analytics, I'm starting to look at how Google Analytics can be used for similar purposes.
The Google tool (formerly Urchin) is very numbers-heavy in its approach, which means that you hardly ever get a nice page full of words to chew on to inspire you. That said, there are a few gems I have been able to dig out that are worth further study.
One metric of success for a blog might be the number of pages that a visitor views when they go to look at it. If you have a blog where people look at a single page in disgust and move on quickly, then you're not really going in the direction you want to go for content. So I decided to run some reports and see if I could find the source of unusually sticky content - where are people coming from when they land on something that keeps their interest?
The report I picked up is under "Marketing optimization", "visitor segment performance", "referring source". The standard report shows the average pages/visit for the top hits. I turned that on its head and looked for the unusually productive referrers, and this is what I got for this blog. (sample size about 1 week, and the data is combined for superpatron + vacuum). To give you a sense of scale, Google hits result in about 1.4 pages/visit. Here's some good ones:
Steven Abram's Steven's Lighthouse had one referrer who started at his The Library 2.0 'Bandwagon' post, clicked through to my Superpatron blog main page, and looked at 5 more pages. (Note to self: Superpatron needs to have a new "popular posts" left hand column, which isn't correct now.)
Dan Klyn's Wildly Appropriate blog yielded one person who looked at 4 more pages.
Someone who hit my Flickr page went on to look at 6 more.
The data set is too small so far - I need to go back further in time - but I think the approach is right: find people who are sending you traffic who seem to be actually interested in the traffic, and give some feedback in their direction.
(for my analytics category)