Will The Social Graph Change Search?
Ah, lunch was delicious and now The Strokes is playing in the session rooms. SMX is totally hardcore.
This time around, Chris Sherman is going to give us a quick presentation and then open things up to question and answer with panelists Aditya Agarwal (Facebook) and Sean Lyndersay (Microsoft. Danny Sullivan is acting as moderator for the Q&A. In case you didn’t hear, me and Danny are totally BFFs. He even twittered it yesterday. You’re jealous, aren’t you? I know.
[There's some crazy romantic mood lighting going on over here. I can barely see Chris Sherman up on the podium.]
Chris says there’s no good industry standard definition for social search. The simple definition is that it’s a collection of Internet wayfinding tools informed by human judgment. Informed can mean many things, including egregiously uninformed. Hee.
The social graph is a picture of the search behavior of a group of people, making unseen connections between them. It’s most effective when the group is made up of trusted people.
We’ve always had social search. The very first guide to the Web was created by Tim Berners-Lee in 1990. Yahoo was original created by a team of human editors. Meta tags were created in 1996 to help content owners influence search engines and were a massive failure!
Algorithmic search is social. Fundamentally, search engines reflect human bias (programmer choices). Also, search engines observe human behavior – click paths, popular URLs, etc – and use that information to modify their algorithm. New personalization efforts are also used to refine search for everyone. The search engines are watching what we do.
So, if social search has always been around, why the buzz now? Algorithmic search has plateaued. Innovation is much harder than it used to be. Humans are still better at some things than computers. A major factor is that many, if not most, of the players in social search are leveraging the work of volunteers.
Types of social search
Shared bookmarks and Web pages: The most basic and least useful type of social search. Still used for a personal tool. Sites like delicious, shadows, my Web, etc.
Tag engines (blogs, RSS): Also called "taggregators". They’re primarily search blogs and RSS feeds. Sites like Technorati, Bloglines, Ask Blog Search and (parenthesis: Blogpulse).
Collaborative directories: Created by teams of volunteers. Sites like the Open Directory Project, Prefound, StumbleUpon, Mahalo, and Wikisearch & Wikipedia.
Personalized verticals: It used to be difficult or labor intensive to create a specialized search engine or directory. This is no loner the case. Tools like Google Custom Search Engine, Eurekster Swickis and Rollyo have changed this.
Social Q&A sites: These have been around forever. They used to be called Aska’s, Usenet, BBS’s, etc. Now they’re sites like Yahoo Answers, AllExperts, Answerbag, etc.
Collaborative Harvesters: Users "nominate" interesting content, others "vote" for it. Sites like Digg, Newsvine, Reddit, Sphinn, etc.
The Tyranny of Democracy
Herd mentality dominates collaborative harvesters. People want to support their friends. Digg has a stealthy "bury" squad that it refuses to acknowledge. Catchy headlines often trounce substantive content.
Then we have hybrids that are part search, but definitely social. Sites like Facebook, Craigslist, Judy’s Book, Insider Pages, LinkedIn, etc.
- Scale and Scope: The Web is expanding at an astonishing rate, particularly now that we have video. Video is becoming a key component of content that people consume on the Web.
- Tagging Issues: Ambiguity of language. Lack of a controlled vocabulary. Human laziness. And of course, the idiots. People are out there tagging stuff that has no relation to the terms they’re using.
- Spammers: Black hat tactics are starting to really emerge.
What will ultimately work?
- A combination of algorithmic and people-mediated search (eg Ask’s new Edison algorithm).
- Trust networks
- Increased personalization and user control over result filtering
- Social search will probably work best for non-text content (photos, music, video, widgets, etc)
Social search is disruptive. It’s already impacting algorithmic results and will have an even more significant effect in the future. But social search also offers unique opportunities.
Social search will become more popular and important over time. People are less predictable than algorithms, so expect both potential and problems. If you aren’t using these tools, take the time to explore them.
Question & Answer
What particular aspect of social media gives it an edge?
Aditya: One of the major reasons why social media offers something over and above search is that the type and volume of content being generated is always changing. The other reason is because it allows us to actually create a really accurate mapping of the social graph. Any data that allows us to do that has the chance to impact the degree of relevance.
Sean: We need to understand the different edges of the social graph. There’s the implicit social graph (family, friends) and there’s an outer ring. There’s also a level where I may not be implicit friends with someone, but based on our behavior we have very similar tastes. Recommendations play a huge part in social media in general. Its people talking to other people and helping them make informed decisions. That tension between it being an informed decision and an uninformed decision is where the search engines have a lot of work to do.
A lot of search marketers have tried to engage in the social arena and have faced backlash. How do you get around that?
Sean: Authenticity is the key. If you look at the recommendation engines like Digg, the stuff that attracts people attention is all over the map. I don’t have any tips. It’s more of a case of there are a lot of things you shouldn’t do – creating a false blog, being overly salesy. What you SHOULD do is make it really easy for people to recommend your content.
Aditya: Try and create content that is interesting and engaging. Present information to the user that is useful, then users will want to interact with it. Don’t try to game the system.
In the pure search space we’ve had another outbreak of privacy concerns. What I find most interesting about social media sites is that users seem to be the opposite. What steps do you take to protect user privacy?
Aditya: Facebook is a communication system for everyone. You can build relationships and trust through your network. We’ve been successful because we give users the ability to control how and what they share. Privacy is hugely important to us. There’s a fine line between how you present those privacy options to users. You want to make sure it’s easy enough to use. It can’t just be semantics.
Sean: Some of the problems you see with privacy often happen because information is being used in ways users didn’t expect to happen. They didn’t know their MySpace information would end up in the SERP. We can’t assume that just because someone signed up for one service that they are willing to go to the next step. You have to take into account the user’s intent when they gave you certain information.
For search to be valuable it needs to be objective. Doesn’t the information being put into social search run the risk of being biased?
Sean: He says it needs to be objective in all cases. The bias itself also needs to be objective (uh, what?). Being as objective as possible in search results is key. If you do have a bias, then you have to disclose that.
Between LinkedIn, Facebook, Sphinn, TripAdvisors, etc, how is someone supposed to manage all of their accounts as a marketer?
The panelists are stumped so my BFF, Tamar Weinberg, steps in, She says you have to get good at something, one or two sites, don’t spread yourself too thin. Then you become an expert in one area.
Sean and Aditya both agree with Tamar. Yey, BFF!
What are you starting to see about who is drawn to a particular group?
Aditya: In terms of whom you end up identifying with more, real world relationships don’t change. You’ll still trust recommendations from your friends. A lot of what you identify with is based on the graph that you have.
Sean: Both are equally valuable. When it comes to recommendations, one of the big problems is the extreme effect. People write reviews for things that they really hate or really love. There’s no middle ground. It’s more of an observation then an answer.
Google said social networks were not performing well. What’s going on? Have expectations been unrealistic?
Sean: One of the reasons search advertising is lucrative is because when people are doing a search they have some purchasing intent, so advertising there is straightforward. Advertising in social networks is more brand-related. You shouldn’t expect that advertising in search and on Facebook are the same.
Aditya: The way of thinking isn’t the same. The data that you have to work with is very different. There are different semantics.
It’s important to be genuine. How can brand advertisers get value from social media sites without looking like spammers?
Aditya: He talks about the Geico Cavemen ads on TV. Geico is not an individual. The idea here is that they created this Web site that users could engage with and advertised for it on Facebook. You’re creating something that is worth users’ time.
Sean: You have to engage with the user and make them feel like the product or the brand is helping them achieve their goals.