Company: SphereLaunched: Private Beta Launch this Weekend
Status: Funded by Doug MacKenzie, Kevin Compton, Phil Black, Will Hearst, David Mahoney, Vince Vannelli and Mike Winton
Previous Post: October 8, 2005
Location: Palo Alto
I met with Tony Conrad, one of the founders of Sphere, today at our office (ok, its a house, but it’s also an office - just ask the IRS). Sphere is a new blog search engine that quite frankly blows everything, and I mean everything, I’ve seen out of the water in terms of relevance (and, by the way, design…Adaptive Path was involved).
The timing for this meeting was perfect - right on the heels of Yahoo’s new blog search engine, and Robert Scobles already-classic post “The race to time-based and blog search“.
Tony, along with co-founders Steve Nieker and Martin Remy, and advisors Matt Mullenweg, Mary Hodder and Toni Schnieder, have created Sphere in a very short period of time and for “less than $200k”.
I saw the live site and was allowed to bang on it as much as I liked. I did, and came away very impressed.
Relevance in blog search is very difficult. Google-type PageRank analysis, which looks at incoming links to a piece of content, simply doesn’t work because new content doesn’t have much in the way of links. Until now, no one has come up with a way to properly sort blog posts by relevance, and the general default way of showing results is “reverse-chrono”, which simply puts the newest stuff at the top.
Sphere appears to have solved the problem, or at least taken big steps in the right direction. Their approach involves three key algorithms - an analysis of links into and out of a blog, an analysis of metadata around a post (links, post frequency, length of posts, etc.), and something Tony calls their “secret sauce”, which is content semantic analysis to filter out spam and to understand what a blog post is talking about.
Result sets show only two posts per blog on the first page, so no one blog can dominate a category. The result set has auto-generated profiles of blogs that include recent links in and out of the blog, average posts per week, average words per post and a link to a full page profile that can be edited by the blog author.
To the right of the main result set are blog (as opposed to post) links that are relevant to the query, and something they call the “Daily Sphere”, which is links to relevant non-blog news stories. I imagine they may add additional content in the side area as well.
The results page integrates both ajax and flash features in an intelligent way (it’s not there just for show).
While Sphere has been indexing blogs since January of 2003, their index only shows results for the previous four months. They will lengthen this window as they scale up operations.
Sphere has tens of thousands of beta requests and is opening up to 100 lucky people this weekend. Sign up for the beta here.
See all
Post
“less than 200k”
bull. shit.
Just tell me they output searches in RSS.
Chris, yes, they will.
John - actually it’s quite possible to do this today. A very small team can create highly cool web apps in a very short time.
yes, michael, indeed it is, and that is a good thing, which many people on and off of sand hill road need to grok.
but i don’t buy the “$200k” figure for sphere.
they all ate cardboard and stone soup between 2003 and now? nah.
and what — they’re ready to tackle the blog indexing problem with a couple old 486’s? no. there’s some iron involved.
it is nice and fashionable to play “small ball” with development these days, but specious hype leads to unrealistic expectations.
if you’re going to report things like that, a little fact checking wouldn’t hurt, y’know…
John, I’m not able to fact check such a statement, although it is a quote from one of the founders. I also agree that to launch will require a significant ramp up in server strength, but they haven’t launched yet. Finally, I suspect that the founders aren’t paying themselves. So, I believe the number is correct.
I’ve been using the site over the weekend and I have found a few holes in Sphere. It is not picking up real-time events very quickly. “Splogs” is a big topic today on Memeorandum but a query on Sphere doesn’t show any relevant blogs
On the 31st October I received this from Sphere:-
Due to a breach in account security, it is necessary for us to reissue a number of password, including yours. Please use the account information in this email and ignore any previous information. Sorry for the inconvenience.
Is this an emerging problem that we are facing with some startups not having good infrastructure and security behind the new wave of AJAX ‘Web2′ applications now.
Sphere team members think they will solve the Blog relevance problem by concentrating on relationships between Blogs, pushing spam-blogs way down the page, matching Blog content with relevant web articles from mainstream media, analysis of metadata, content semantic analysis -All These seem to me like more of the same old macro-content techniques. I believe that in order to really push forward a solution to the relevancy of Blogs they’d better search the solution with MICROCONTENT techniques. I’ll help them if they’ll ask…
77270r1hO850 7Ipd2fjqAU9 TnNF75ztsDiYP
As of Tuesday it is open to the public. Check it out
E2
http://domains.optionextreme.net/
E1
http://domains.optionextreme.com/
Sphere have worked hard to improve their blog software. I personally think the software is pretty good.
Very beautiful, learning!!!