We've been playing with Sphere for the past week and are able to start sharing it with our clients today. Sphere is a new blog search and discovery tool. (Hey, don't we already have Technorati, IceRocket, Google Blog Search, Feedster, etc.?) Sure, but this issue is far from solved. First, as we monitor for our clients and help them track conversations, it's clear that no one tool is comprehensive enough to rely on. Secondly, we waste a good deal of time on spam blogs & the leading services have yet to really work through that problem.
So, Sphere. Sphere's basic premise is that everyday folks want to get right at the good stuff. They don't need a zillion features that would interest power bloggers — they want to find good posts about topics they care about. In other words, give me fewer results that are more relevant to my search. Give me less to wade through. Filter for me, don't just find. The focus is on simple search & introduction to other blogs and other media discussions related to these search terms.
Sphere builds its index three ways: link structure (which T'rati prioritizes), analysis of meta data and semantic analysis of the post content. Sphere then lets you review its results and use them as a jumping off point in a number of ways. First, you can use its blog search results filtered by relevance (the default) or recency:
You can also use sliders to customize the date range you want to search for something — around a certain event, say.
Once you've identified a relevant link, you can easily find out a bit more about the blog itself by hitting the profile button.
(More is coming on these blog profiles. They'll add descriptor text to this initial bit of information.)
Then you can move off into discovering other relevant blogs for certain search terms by using the Featured Blogs. Sphere CEO Tony Conrad said that Sphere can generate dynamic featured blog lists for about 15,000 terms today.
Here are the featured blogs you get when you search for 'podcasting'. You can also suggest other blogs to include, and I expect the results to improve as time goes on and more users help groom the system.
You can also see how the conversation is being carried out in other media, using Sphere's Related Media link.
Here you can see that a search for 'blogging' gives us Webshot photos using that tag, Yahoo news stories that are related, as well as books and podcast listings.
The cool thing is: Sphere is creating links between relevant content where none previously existed.
The most whiz bang example of this is using the Sphere bookmarklet — Sphere It!. You install the bookmarklet in your browser toolbar, and click it when you want to see what relevant blogs are saying about any Web content that you are reading. Let's say we're reading yesterday's story about Apple's song pricing. Sphere It returns the following results:
Again, Sphere has created these connections through metadata and semantic analysis, so I can follow blog discussion about the topic, even if the blog in question hasn't actually linked to the article I want to use as the basis of my search. Kind of like having a memeorandum engine in your browser, aimed at whatever topic is important to you at the time.
There is a great deal of discussion out there today about Sphere. Check out the tech.memeorandum conversation.
Mike Arrington has a thorough review and a podcast with Tony Conrad & Toni Schneider about the launch.
Om has a rundown the initial investment for Sphere.
Sphere is one of the first companies involved in Adapative Path's New Ventures program, and Ryan Freitas has an extensive essay detailing the design and development process for the service. The whiz kids at Mule Design also lent their handiwork.
Technorati Tags: adaptive+path, mule design, new+ventures, sphere, toni+schneider, tony+conrad