Social/wiki search engine provider Eurekster has been down for the past two days and users are wondering if the company will return to provide the site search functionality that many have come to depend on. Things look rough for the company.
Eurekster provides a service called a Swicki, a search function that lets site owners identify what other sites in their community they would like to have included in their Swicki search results. We've used Swicki here at ReadWriteWeb for some time, as have many other blogs and online communities. The company offered revenue sharing from search ads. Eurekster says that more than 100,000 sites have created Swickis and traffic to Eurekster used to be strong.
Both Eurekster and Swicki.com are down and users report to us that's been the case for the past two days. Traffic to the site has plummeted since the start of the year and key executives have departed. According to their LinkedIn profiles, VP of Engineering Blair Cassidy and VP of Product Development Tac Leung left the company in March and April respectively - but both are still listed on the company's management page. As we publish this, neither emails nor phone messages have been returned by the company. Channel Marketing Manager at Eurekster Alex Holmes sent out one Twitter message two days ago reading "temporary service outage on http://www.eurekster.com . Swickis will be back shortly."
Eurekster was founded in October of 2003 and raised more than $6 million in venture capital, most recently a $5.5 million round in March of 2007.
It's always interesting to ask why a startup may not have succeeded, though to be fair Eurekster could come back at a moment's notice. The company's method of determining relevance in search was always a little unclear. Google Site Search is easy to install and its results are fairly predictable. Many users wanted search results in chronological order, but that's not the first place Swicki searches took you. Finally, the Swicki usually lived in a blog sidebar. That's a hard place to build a business.
Those are our theories about why Swicki has struggled. What are yours? We hope to hear from the company soon, it's never a pretty sight to see a startup sputter out.
Comments
Subscribe to comments for this post OR Subscribe to comments for all Read/WriteWeb posts
I like this quote;
"Finally, the Swicki usually lived in a blog sidebar. That's a hard place to build a business."
=-)
Posted by: James | May 22, 2008 1:21 PM
I have used this cool device since 2007. i like it alot and hope our friends at eurekster hold on to a good thing.
Keith
Posted by: Keith Kennedy | May 22, 2008 1:22 PM
I heard that they are going out of business.
Posted by: jenkins | May 22, 2008 2:39 PM
If it is true, then it is sad to see a New Zealand linked company go out of business. I've never used Eurekster before.
Here is my honest opinion. There are lots of different applications using the same algorithm as its core technology & foundation. So, my point here is that when you hear someone says, hey , look we've got a cool app, here that can do this and that, but in fact, the algorithm is not new at all, it might be a new domain of use for the algorithm, but the idea of pitching for funding to a VC based solely on the new use of the algorithm, to me, makes VCs naive. I don't know what technology/algorithm that Eurekster use as core engine, but I wouldn't be surprised if they're using LSA (Latent Semantic Analysis), which everybody today use LSA in their site search engine (both search engine vendors & developers). LSA is not new, however it's always reported in the literature of it use in a new domain. For example, LSA is solved by an algorithm called SVD (Singular Value Decomposition), which was developed in the late 1960s has been used in collaborative filtering & recommendation engine recently as is reported in the following paper:
Using singular value decomposition (SVD) approximation for collaborative filtering
SVD is old, however its application in recommendation engine is new and if I am a VC, I would ask questions about the algorithm or technology involved, because there is not one way to do recommendation or site search but many. Suppose that a vendor X with a recommender system product that uses SVD for its core technology. A new emerging vendor Y will realizes the shortfall of SVD and adopted a similar algorithm but more robust than SVD such as NNMF (non-negative matrix factorisation) as reported in the following title:
Learning from Incomplete Ratings Using Non-negative Matrix
NNMF is also used for text search engine (site search, etc...) which is again superior in performance compared to SVD-based LSA algorithm.
See, it wouldn't take long for potential competitor of a business to figured out a better way (may be some new algorithm) that do the same thing but do it better and when this happens, the incumbent in that specific market (for example, the site search domain) will find it hard to compete with this emerging competitor.
I believe that Eurekster is based on a technology that was built in the late 1990s by the founders. It was reported that they sold the technology, then bought it back some years later from the vendor who bought it from them in the first place. So, in the last 10 years, there had been lots of new things (algorithms) emerged. So, to not adapt to the latest and newest algorithms, IMO is a commercial suicide in the long run. I am pretty much convinced that the core Eurekster technology was the one from the original company they sold in the late 1990s then bought it back again (the technology) in the early 2000s.
There had been so much happening in the domain of online advertisement since then for Eurekster not to adapt to the latest algorithms. For example, new researches in online advertisement had been made available recently such as in the Data mining journal special issue on Audience intelligence for advertising.
IMO, the key is to adapt the technology to be competitive, because there is no guarantee that what one has developed and claim to be the world first, will stay true forever.
Finally, for VCs out there that need an independent evaluation of the technology of a startup company which is bidding for funds, perhaps you might find the consultancy services of Dr. Isabelle Guyon's very useful. Dr. Guyon's work (peer review publications in various computing journals) is outstanding in the field of pattern recognition, machine learning, search engine, general artificial intelligence and more... She is recognized by her peers as an outstanding researcher. It is better to pay consulting fees to get Dr. Guyon's opinion & evaluation of a bidder's (startups) core technology, before plunging millions of dollars to some very good business plan on paper and found out a few years later that the startup company's lost out to a new competitor because the startup company's technology wasn't unique at all. Business plan always looks better on paper, but the core technology is quite difficult for VCs to evaluate as they might find it too technical. VCs do rely only on the words of the bidder.
Posted by: Falafulu Fisi | May 22, 2008 10:26 PM
http://www.techcrunch.com/2006/10/24/googles-choice-of-words/
Posted by: Google's A Bitch | May 23, 2008 4:50 AM
Too bad if they won't come back up. I have used this on the sananselmo.com community site and the daily feedback of search words provided a simple and practical way to know of visitor's interest's and what they aren't finding on the site.
Hope that they can overcome some adversity.
Posted by: Bill Jacobson | May 23, 2008 11:31 AM
I have used Swikis for as long as they have existed. Now all my sites look like crap. Without my Swiki, there's a big white space above my header graphic!
By the time I finish changing some 200 plus web pages to eliminate the white space, I don't know that I'll ever re-add my Swiki. That's a shame, too. It was worth over $200,000 if Eurekster had ever implented their trading scheme.
I hate to see it go. It was a fun little search tool to have on my sites and it got more use than my Google search ever did.
By the way, it was nice to find your post.
Posted by: Tim Woodard | May 23, 2008 11:40 AM
Finally some news! I have nearly 150 swickis mostly stand alone and I spent A LOT of time building them. We were both earning from Google but Eurekster dropped 3/4 of them from search and Google awhile back. I also have 100's of domains directed to...nothing. I am pissed off.
Posted by: texidriver | May 23, 2008 1:14 PM
I bet Google payed off Eurekster to close down perhaps leaving them with the option of opening up under a different name and targeting a different audience. When Eurekster opened up this "last" time, I turned up a blog entry from around 2004 where this guy was complaining about how Eurekster or whatever the company was called then shut the door with no explanation leaving him holding a bag of empty swickis. Since this is the second time they have done this perhaps it is just a Kiwi way of doing business. I think the core Eurekster team has learned a lot from all the earnest swickiers out there that thought this search approach was worth sinking some time into for a bit of pleasure and profit. They will take that knowledge just like they did before and open up under another cool trickster name.
Looking at the chart above its obvious that Google thought that most of the swickis were spam. Eurekster basically said this in their blog. A lot of those spammy swickis were bringing in decent money. When you combine that with the incredible ease of making swickis, swicki farms and swicki universes, Google had something to worry about.
Posted by: Kim Remmington | May 23, 2008 7:53 PM
Just FYI, I linked to your article from Swicki Tag Search Down - Eurekster Dead?, but the trackback doesn't show up. Perhaps I'm in your spam filter?
Posted by: bernie | May 24, 2008 1:54 AM
I use it, i like it, i miss it, Where is it?
Posted by: Keith Kennedy | May 24, 2008 2:39 AM
http://eurekster-sux.compro-serve.com/
Posted by: Crian | May 24, 2008 6:53 PM
Looks like this company disappeared. Glad you posted this story because my swicki stopped working. They should have a better answer than a random Twitter message that isn't accurate. Being offline for over a week isn't a temporary outage.
Posted by: Ed Gillet | May 26, 2008 7:20 PM
I can't believe they just let it die like this without a peep or warning. Seems like they had a decent amount of traffic that was at least worth a fire sale. Any longer and all of their customers will be gone.
Posted by: dumbfounder | May 27, 2008 8:16 AM
Well one of the big reasons why they have dropped is probably because Google made an change in their search algo and what to keep in their index at the end of Nov 2007. And basically all swicki "subdomains" dropped out of Google index. From 2 million indexed swickis to currently 300.000.
Trust me I know this since I monitored my swickis on daily base and was trying to build up a nice portfolio of swickis on "many" different topics that I came up with while watching TV and reading magazines.
I started creating swickis in September 2006, and basically continued doing them until January 2008.
During 2007 i managed to climb from a few cents per day from Adsense to a couple of days in December when I finally broke 100$/ day barrier.
And then first week in December 2007, they they made the decision to drop adsense revenue for any swicki that was not in their index.
And guess what I only had 3 in there. Never occurred to me to ask them to add my swiciks as the first ones popped up there automatically.
So my +2000 swickis suddenly dropped from a nice +100$/day to not even 1$/day...... and they were really showing a nice pace towards even 200$/day.
1 swicki maybe gave only 0,03$, a few gave 1$ per day, but you add upp +100 swickis that have 10-30 visitors a day then is suddenly picks up.
I made 4300$ from adsense revenue from my swickis during 2007, and was already looking for a prosperous 2008.
Talk about wasted time from my side, to promote and submit my swickis and giving "eurekster" linkpower over the +1 year that I was doing daily management of my swickis and creating new ones.
So for now I have not touched or tried to promote them over 6 months. It's going to be interesting to see what happens.
If they would allow the Adsense or other revenue back on for any created swicki, then I might be interested to promote them again as it would be worthwhile.
Go to http:// shoutouts .swicki .com - and you will see that they are "working on it...
2006 and 2007 I thought this was a nice way to make some extra $$$ in the years to come. Now....... Nope.
Luckly I have other income sources and no debts....
But that is life.. It sucks.
Now is time to concentrate on new ventures... Time for Poker News in Asia..
// TomYam
Posted by: TomYam | May 28, 2008 7:45 AM
Looks like the service is back up, see their post here.
Posted by: Jack | June 2, 2008 1:09 AM