![The Big Lebowski (Limited Edition) [Blu-ray Book + Digital Copy]](http://ecx.images-amazon.com/images/I/512F90yEqdL._SL75_.jpg){kind=small}
How Search Will Revolutionize Social Networking
Young Urban Professional
Social networking is on fire. eMarketer predicts that in the US the category will reach 44.3% of Internet users by year's end. According to Google Insights, related searches are up 3,000% over the last four years.
It has a ways to go before it's truly mainstream on a global level.
(More than half of adults in 17 countries don't know what social networking is, according to Synovate.) Still the phenomenon is a sure thing, even though the individual winners and losers will surely shift.
What has me most excited though about social networking is a capability that isn't really in place yet in a powerful way - and that's search.
Much like the early days of the web, social networks have yet to fully
exploit search. Recall that before Google came along 10 years ago web search was
woeful at best and also un-monetized. Eventually that all changed. Even though Facebook's search is weak, already it's one of the fastest growing search engines. That's remarkable.
Search will become a core feature of the social network experience, add in social elements,
usher in easier monetization and in the process revolutionize advertising. Here's a look at some trends to watch...
TRUSTED SEARCH TRUMPS UNTRUSTED SEARCH - Do you trust Google? I do as does most everyone.
Do you trust what's in Google? For me, that depends on what I am
searching for and where it comes from. However, I do trust the 1,000
people I have added to my social network on Facebook.
In fact, it's why I limit my connections there to people I have either
met or corresponded with. I value what they talk about and share
there.
However, there's a gaping hole in the Facebook experience. While I can
search through my friends, find new friends and also groups, I can't
search the content my network creates. In addition, I can't go a
layer deeper to see what my friends' friends are sharing (as I can on Friendfeed). Look for
search to get embedded deeper into the social networking experience and create a split
between trusted and untrusted search. The impact on PR will be major here too.
Microsoft's forthcoming integration of Live Search into Facebook could be the first step toward trusted search. MySpace has already site-wide search and can tweak it to achieve the same. (MySpace and Microsoft are Edelman clients.)
CONTEXTUAL SEARCH ADS GET SOCIAL - Google and MySpace have an advertising agreement going back to 2006. Facebook and Microsoft have a similar arrangement that started last year. So the search engines clearly view the social networks as a monetization venue and vice versa.
Social
network advertising to date, though, has been a mixed bag. Everyone is
innovating. But the draw on social networks is your friends, which
makes it harder to be distracted by ads. Enter search. Watch for
contextual search advertising and programs like Facebook's social ads to mix. New models
will emerge where contextual ads are surfaced based on the content
created and recommended by your friends.
SOCIAL NETWORKS BECOME SEARCH ENGINES - If you went through my browser
history, you would be bored. I spend most of my time in Google's
universe of sites and on The New York Times site. Beyond that, you will
find a bushel of social networks - Facebook, Friendfeed, LinkedIn and
Twitter.
Now, what if I could interact with any or all of my favorite sites all
from a single social network and have my friends add value to that
experience? It's coming. Today, for example, on Facebook I use Six
Apart's BlogIt to Twitter. I also catch up with my favorite sports
teams using Sportsline's Facebook application. These are simplistic though. Notice what's
missing - I can't search the web yet from inside Facebook. However, on
MySpace I can. But this is the beginning.
In the near future the search engines will all create applications or
hooks into soc nets that let you search and annotate the web in
conjunction with your friends, changing the web experience. The image above from the Shifted Librarian shows how she is able to search her local library direct from Facebook. Now imagine that same search application gets social and you can see that a major evolution in how we mine the web with friends is coming soon.
Reader Comments (10)
=)
I agree with your statement that "social networks have yet to fully exploit search." However, I think I'd go even further and say, "Social networks will find that without search and other organizing pricciples, users will be increasingly lost in such environments."
There's a reason eMarketer shows search catching up with email as a primary activity. ( http://tinyurl.com/67oe47 ) There's just so much stuff and we're movng so fast we most often choose search over other organization priciples.
Unfortunately, there are some clear and likely unknown major challenges in applying search in some social environments. Search, (for the most part anyway), depends on text. (Yes, there's a variety of meta data that can be used for relevancy ranking; from link popularity to social graph, etc. etc.) But still, text is at the core. Consider for a moment the Twitterspace of 140 character messages. All manner of odd abbreviations are being used. And on FunWalls or FriendFeed clustered results, multiple concepts are expressed in very short order. One of the core aspects of most search indicies is some sort of inverted index, vector space model and so on.
The very nature of the type of content expressed in many social spaces is likely to offer some serious new challenges to successful findability in these spaces. Some challenges can be easily handled with synonym dictionaries and such. At Twing.com, we're optimized to finability in forums and other discussion social spaces, which itself has special challenges. But sometimes, like just having very little to go on, things can get really dicey in terms of indexing and relevance ranking. Witness the current debates about Google forcing Twitter to nofollow links. ( http://tinyurl.com/597am9 ) Why are they doing this? I have to believe that essentially they haven't figured out a way for this content type to not really screw up their ranking algorithms.
So-called social search cannot alone solve these basic challenges. Clearly, some basic search tools will be applied soon. They'll likely not get the best results; though users may never really know that. The fun part is this means there's still room for those of us who work on specialized vertical search engines.
Scott
When search began it was an arms race to see who could index the most URLs. Google used to have a running tally on their home page. As the web grew, so did the "spammy" sites. The only way to ensure quality results was to assume that all URLs were spam, and require them to "earn" their place within the index.
The social networks came along. Steve you are absolutely correct. Search has to become core to the social networking infrastructure because of the very nature of the social networks themselves. But, that is inherently an issue, since most of the social networks are relative silos, even with OpenSocial and other efforts. But social networks have added an important element to search - Trust.
And, as you rightly state, trust based search was born. The major difference between trust based search and standard search is that with a trust based search system, all the URLs are assumed valid and trusted, and the index consists only of URLs that are important to you.
There are two companies that are doing something interesting along these lines. Friendfeed, which is taking the stance of providing the flow of information to users and allowing them to search through it.
Lijit (where I work) is taking the opposite tact. By allowing the publisher to be the center of his/her universe, the reader is applying intent, and Lijit is providing the publisher's information stream that matches that intent.
Social search, with its arbitrary ranking system will be much like Digg where the results are really popularity not relevancy.
Trust based search is certainly the wave of the future of search.
Thanks for clarifying. What you're saying also makes sense. However, I'm still not at all sure how one gets away from the reality of the linguistics of things. Even once you've scoped a collection, (in this case to some trust circle based on whatever criteria), the initial user query is going to be just that... a text based query of some sort.
I checked out Lijit. Seems like a very cool approach to an important aspect of some problems. And you yourselves seem to have already discovered some scope issues. Such as the fact that you have differing topic networks. Can I assume these are different circles of trust? In which case, does a search for "scuba" go across only certain topical networks or all of them? In any case, there's still issues in the morphological tiers or layers. Specifcally, the small amount of discrete units in the layers of morphology, parts of speech, syntax discourse and so on. For example, I have trouble believing any kind of lexical functional grammar tools will work with really small text segments. Certainly, other approaches will work to varying degrees. I'm just saying getting GOOD relevancy ranking will be very hard.
If you've solved this in satisfactory ways, well... congrats. You all will ideally do very well in that case.
Scott
Yes, when you get down to it, its just a text query, but the major difference between Lijit (or other trsut based search systems) is that the trust itself (in Steve's example - Facebook) applies context and relevance. So, for example, if you do a search on my blog for "billie" you will get posts, pictures and videos of my dog. Do the same search on Google, and the results are all over the place. Thats where trust makes its real difference.
If Steve used a trust based system, then a search for social search on this blog would bring up this post, same search on Google, nada.
Yes, ranking algos and relevancy is always the battle with any search system. We think that having an individual at the center of the network allows us to make better choices around relevancy, because its about degrees of trust, not links, or meta tags.
To answer your question around topical networks, trust based search has to be organized around the individual's trust network. So its defined by the individual versus some algorithm. At Lijit, we have also organized our publishers into topical networks, mostly to make it easy for advertisers.
Regardless, search without trust has become unwieldy.
Don't know what is wrong what is rite but i know that every one has there own point of view and same goes to this one xlpcnm xlpcnm - justin bieber shoes.