First detailed look at Sphere beta |
An invite arrived today for Sphere beta and it encouraged me to check it out and share posted screenshots. I was not given any invites to share, so the only way others can check this out currently is to do what I did: sign up at the site. I don’t know anybody involved here, no connections, I went to the website awhile back, filled out the form and the invite came today.

Sphere is a new blog search engine that seeks to compete against IceRocket, Feedster, Technorati, Yahoo, Google and more. Even though the blog search space is currently crowded most of these engines are pretty weak and contain a lot of spam. In order to get the best results one currently has to search/track multiple blog search engines. There really is no clear cut champion in the blog search space.
Here are my first thoughts on Sphere and where it currently fits:
- design is clean and straighforward with results along the right hand side and two colored tabs on the right “Related blogs” (green) and “New Articles” (orange).
- currently no ads
- default search is by relevance, in English and shows results over the last week.
- Sphere founders, Tony, Steve and Martin indicated that the blog post database actually goes back to 2003, but they are only showing the most recent four months.
- Sphere crawls blog posts, not content contained in RSS feeds. I guess this means they are scraping pages as opposed to listening to reading what is being syndicated. This makes Sphere work more like a search engine spider than an rss aggregator.
- I was happy to see my php-scripts blog which was started back in August already being indexed in Sphere. Webmaster Cookbook was there as well and of course this blog too. Neither my blogspot blog based on Gmail or my very first blog (my personal blog) or my MSN Spaces blog made the cut. Same deal on tabletpcblogs.com.
This leads me to wonder if Sphere has a bias against third party hosted blogs? This might not be a bad thing considering that someone is willing to pay for their own hosting maybe means they will be less likely to spam or splog operation.
- the relevance searches seemed pretty good and relatively spam/splog-free. I found some splogs while doing a vanity search on my name. Sphere seems better than most at keeping this crap out, but there is still some dirt in the doorway (see: broadband-internet-blog.com). They have an optional date search as well.
- a popup mini profile lists stats for each blog containing the following information: avg. posts per week, avg. words per post and avg. links per post (see below).

The stats for MakeYouGoHmm were inaccurate. Here’s the stats based on last month’s numbers (which were lower than normal) vs. Sphere reported stats:
average posts/week:
sphere: 19
Hmm: 25.2 (3.6 avg. posts per day)
average words/post:
sphere: 636
Hmm: 249 (based on 643,231 words in 2,574 posts)
average links/post:
sphere: 1695 (this is definitely wrong)
Hmm: do not currently track this figure, but I know the average post doesn’t have 1,695 links in it. I suppose I could run a quick database query to see how many links were in each post and then arrived at an average figure. I would guess it’s something in the neighborhood of 5 links per post on average. The other blog stats I have in there look closer to reality.
- The popup works but has errors in Internet Explorer and doesn’t work at all in Opera. I didn’t test in Firefox or Safari.
- I didn’t find the news articles or related blogs section very related or relevant. This might be one area that is still heavily under construction (?)
- search results not available in RSS yet (?)
Overall first impression
Sphere didn’t blow me away but it does seem promising. I certainly liked it better than Google’s blog search which was littered with splogs at their beta launch, but has gotten better since. Technorati, which has flash periods of greatness and suckness currently is closer to good than bad. Feedster is ok. IceRocket is trying hard and has some cool stats like their trend deal. Where does Sphere fit into this crowded space? Compared against the rest it’s vanilla.
Sphere doesn’t really have any gotta have new feature that separates them from the rest except that they are scraping pages versus scanning RSS feeds. This may give them an edge on finding quality content that doesn’t current reside in RSS feeds. Also, there are some good quality third party hosted blogs out there, so hopefully their program will include more of these blogs when they get to a public beta launch.
The ideal, dream useful blog search will give me exactly what I’m looking for in whatever format I desire (HTML, RSS, etc) as close to real time as possible with little to no noise (splogs / spam) and provide some useful comparative analysis tools. Finally it should have that special hook that makes it stand out from the rest.
I’m not sure if Sphere has that last part worked out yet.
Related Posts- How to add Sphere It function based on number of words in post
- IceRocket icing blogger and .info splogs
- BlogCompare v0.3.1 changelog
- Feedburner CEO: Click outs in RSS going down
- [site news] Average post length numbers
- How to add a dynamic digg vote this button, show related posts and more using Wordpress




in your face with sphere beta
A new blogging search engine is in town, but unfortunately this was the first time I got my invite… six days after good ol’ TDavid got his over at MakesYouGoHmm. I was definitely curious as to if things were fixed…
Trackback by LUX.ET.UMBRA — November 14, 2005 @ 5:55 pm PST
[…] Techcrunch points out that Time has started using the Sphere It function on its articles. I took a look at the functionality and liked the results it returned. Note: Sphere beta was reviewed at Hmm in November 2005. […]
Pingback by Make You Go Hmm: » How to add Sphere It function based on number of words in post — May 23, 2006 @ 5:29 am PST