« A Marketer's Guide to Social Bookmarking & Tagging | Main | Effectively Leveraging Social Networking »

April 22, 2008

Keynote - Social Search: The Human Challengers

Posted by Lisa Barone

It's a keynote. In the afternoon. Aren't these supposed to be in the morning? I'm so confused. And tired. Are you tired? Hopefully our speakers aren't.

This afternoon (not morning) we have Jason Calacanis (Mahalo), Steven Marder (Eurekster) and Jimmy Wales (Wikia Search) speaking. Danny Sullivan and Chris Sherman will be acting as our lovable moderators. Danny Sullivan is giving me the evil eye. Then he told everyone to sit down and shut up. Lovingly, of course.

Steven Marder is up next. Or first, rather.

They're about the distributed play, not about creating a destination. They believe most Web sites have formed some type of community.

Eurekster enables publishes of various shapes and sizes to install branded and highly targeted search widgets and environments that harness the knowledge and behavior of online communities.

Value Proposition:

• Publisher control: Ability to customize feeds and source of content, choose how to monetize and create and manage Buzzcloud
• Reinforces publisher brand and improves user experience
• SEM/SEO Benefits: Syndication and distribution opportunity.

Swicki: Social Search Widget.

The swicki is a customized widget. It surfaces community activity including popular, suggested and recent terms. Drives Grabbing and Sharing. Multi-media search capability and tags. Drives user/community engagement.

Eureksters Social Search:

• Social media meets search
• Algo + Humans
• Blend of Intent
• Publisher-guided, community sharing

Social Search for Marketers: Leverages key characteristics of social media and applies marketers and publishers brand.

Steven shows the evolution of Eurekster, looking at pages from 2004, 2005 and today. They've always believe in Universal Search. They don't think people should have to ask for specific types of content. There is a need for a trusted relationship or expert source.
Potentially increasing search result relevance by applying the social graph to search
How to effectively apply the social graph to search? Socially connected ala Friendster vs profile.

How to create/surface additional high quality content? Need for community and collaboration. How can site owners establish feedback loops.

The social search marketing opportunity:

• End Users: Empowered to collaborate. Empowered to comment on and post results and to implicitly and explicitly affect search result rankings.
• Discovery via enhanced content" Access to another layer of quality content providing for an improved user experience.
• Leverage Community

Jimmy Wales is next to show us a bunch of screenshots of Wikia Search 0.2.

They're doing everything as open source software. They're releasing all of the code they're creating. All of the data they're also releasing under CC. They're doing that because search has been a black box for too long. It's a problem for a lot of social and political reasons.

First, enter in a search term. They provide the results - which aren't really good right now. They're only crawling a small subset of the Web. They're in Alpha mode, trying to do everything publicly and out in the open. They're developing features live. Anyone can add a URL to the displayed results. And anyone can edit that. It's all AJAX and live immediately. Anyone can append short comments to an entry or spotlight that entry. [Yeah, that's not going to get abused or anything. --Susan]

Other search terms may be related to the current one. Anyone can add related search topics the results and toggle between them. After time, results will improve but spam will inevitably emerge.

[You guys are missing about a gazillion screenshots that help to explain all this. Go do some searches on Wikia and see what he's talking about.]

How do you control spam? Everyone gets a profile page. They encourage the community. Everything people do on the site appears in their profile. It makes things really transparent and helps to weed out the bad eggs.

Lots more social features coming in the next month or so.

Rounding out the keynote is Jason Calacanis.

Mahalo is a human powered search engine. They pay people to maintain their search results. The pitch for human search is that it's spam-free and ungame-able. The fact is, it's kind of borked already. Props to Jason for using the word "borked" in a presentation. He shows how Mahalo results can be more relevant than "traditional" results. He also takes a shot at Ask.com. I'll give him that.

He mentions the second click service. They're looking at the behaviors that users do on the second click. They're taking that and surfacing it on their page. It's somewhat controversial. They've spent a couple of hours on the Paris Hotel Page and they go to all the different sites that they link to and they cross reference what those different sites are saying about it.

Mahalo launched a new feature today. It's a microformat. When you're on a Mahalo page, you can do different things with them - like Map them, add them to your address book, etc. It will help people. They also have a social network.

Every link that goes into Mahalo is human-edited. The only way to "game" it, is to create great content. This builds private trust scores.

SEO is a wasted industry. You're wasting your time fighting off ranking problems instead of creating great content. You're just spinning your wheels hoping the Google gods won't kick you out. It's a bad way to live your life. Using a human service is a better way to go about it.

Later on Jason says a lot of the people who hire SEOs don't deserve to be ranked.

Is Jason 1.0 returning? I think I actually just heard someone hiss when he said that. [Wow, Jason, really? I thought he was getting out of the link bait through negativity business.--Susan]

Question and Answer

The personalization stuff has been the holy grail. The downside has always been a scale issue. What you're doing works really well for the popular subjects but how will you scale it?

Jason: If you look at some other projects, the scale issue is something that seems like a real issue. The way Mahalo will scale is they'll spent 50 million to build Mahalo and then it will take 10 million a year to maintain. We'll probably have thousands of fulltime people working on it. The reason there are so many long tail searches is because search is failing. People have to type more words to get what they're looking for.

Steven: It's not that big of an issue. They're relying on publishers to do most of the work.

[Unfortunately, I got really sick during the Q&A portion of this session and had to leave. I apologize.] Just take care of yourself, Lisa. --Susan

Posted at April 22, 2008 5:53 PM
View related entries in: Social Media, liveblog, smxsocial08

Comments

Posted by: Igor The Troll at April 22, 2008 8:35 PM

Posted by: Blog For Money at April 23, 2008 8:59 AM

Posted by: Feydakin at April 23, 2008 9:03 AM

Posted by: Joshua Steimle at April 23, 2008 10:55 AM

Posted by: Matt McGowan at April 23, 2008 1:05 PM

Posted by: billse at April 23, 2008 3:59 PM

Post a comment

Name:

Email Address:

URL:
Remember Me? Yes No

Comments: (you may use HTML tags for style)