I've been interested in Collarity for a few months now - ever since I read about Collarity in Read/WriteWeb.  Attempts to meet for an interview at AlwaysON NYC fell though as did our first and second attempts at setting up a phone interview.

But we finally met yesterday and I spoke to Levi Cohen and Rob Rustad of Collarity for an hour - we had a very spirited exchange where I asked several questions and  looked at their demo and a few sites that are running versions of Collarity right now.

What I came away with - Collarity has some very interesting capabilities for Web Analytics - something which they're not fully marketing yet - in fact, Collarity is painted more as a Search Engine, but that's only part of what it does - and the other capabilities of Collarity are more interesting to me.  Let me illustrate what Collarity can do.

Visitors to a website running Collarity are automatically categorized into "communities" based on keyword associations and the pages they visit on the site, and a site may have up to several hundred "communities". 

For a Web Analyst - a "community" is another word for "customer segment" and Collarity is actually - a customer segmentation engine - perhaps the best one in existence today (again, it's not really being profiled that way right now).

When you log into the Collarity backend and look at your site - you'll see all the "communities"/"segments" on the far left and when you select a community, you'll get the pages they found most interesting and a tag cloud of the keywords that segment/community found interesting!

Why would you care that Collarity can segment your audience for you?  Well, for one thing, you could look at the audience you think your getting VS. the audience your actually attracting to your site.  You can also find out what pages of your site, what content, is interesting to each audience based on their behavior.

For example - say you have a page on your Corporate Portal that is geared towards the CEO Audience and you have a white paper you want them to download.  How do you know the people who came to the CEO page and downloaded the CEO Study were really the audience you wanted?  You don't know, perhaps can't know for sure today.   In fact, we'd solve that issue by doing a reverse IP lookup of the IP Address of the visitor and put them into a customer segment that way.   But Collarity is much more powerful than that - as it can tell you what the attributes of the actual audience that came to the CEO page are.

What if you wanted a CEO Study page to be visited by CEO's but Health Care professionals came there instead?  Would that page be as successful - even if the CEO Study white paper was downloaded frequently?  No, it wouldn't.  With Collarity you can know the audience of the page and you can change the keywords on the page, along with the copy until it matches the audience you want to see your page - and you can do it in real time!

How's that for Customer Segmentation.  And I did not get to the Search Engine part - which will be my next post.