If ever there was a strong meme in search, it’s the impact of social: Everyone is talking about how Facebook and Twitter are threatening Google for what I’ve called the “oxygen” of the web: distribution of attention.
A little background. Google rose to prominence as the absolute winner in the Internet’s distribution game. The de facto interface for knowledge navigation, Google brought signal to the noise of Web 1.0: Sure, nearly everything worth publishing was now on the web, but how on earth could you find that ONE thing that mattered to your query, NOW?
A hundred billion plus dollar business ensued: we all now use Google to find that which we want to find on the web. In particular, Google is great at delivering authority on the web for those things that had already been published and ranked: In a way, Google has become the reference librarian of the web.
But…just searching a reference library is one thing. What about finding things people are talking about right now? And wouldn’t it be great if you could cross index that reference library with your social graph, so that people you trusted helped you go from query to decision?
Twitter and Facebook promise that next step in search. Let’s tease this out a bit.
We have different modes when we search. Sometimes we are looking for that perfect reference point – an article on how to train a dog, for example, or a how to guide on building a treehouse. But then we hit a critical inflection – we want to validate our reference material with a real live human connection. And Google can’t really do that. In short, we want to cross reference what we’ve learned with the experience of someone we trust.
Before the rise and ubiquity of social networks, the ability to do this was pretty serendipitous – sometimes in our reference search we found humans with whom we could connect and then learn (this happened to me in 1995 as I was searching online for my birth mother, but that’s another story).
But it’s happening more and more online now, thanks to our ability to use Twitter and Facebook to query our social graph. Through status updates or tweets, we can ask real people that which before we asked Google. And, by reading through the lifestreams of our network, we can discover that which we might never has asked, but nevertheless find interesting.
It’s late and I’m working way too many hours to do this line of thinking justice. But I will simply state it this way: Facebook and Twitter, you need to get better at mixing traditional web search with what you’ve already got. And Google/Microsoft, well, vicey versa. You need to get better at mixing social into your traditional web search.
Whoever does it best, wins.
Update: A new study on the interplay of search and social media can be found here.
14 thoughts on “Search Does That. Social Does This. Give Me A Reese’s Cup Please”
You’re so right. The ability to scope a search to your social graph seems like the most obvious and, alas, unmet user need. That’s why we built it into TuneIn from the beginning.
Compare this Twitter search: http://search.twitter.com/search?q=advertising
with this TuneIn search: http://tunein.com/johnbattelle?q=advertising
The TuneIn results are drawn from tweets and articles written by people you’re following.
That’s a good example, Adam. There will be many flavors of social search which will be right for different situations. Having a broad “trends” type search is good for learning what’s going on right now (see OneRiot’s PulseRank algorithm). But opinions on the Palm Pre might come best from your circle of friends (enter TuneIn).
It’s also important to note that the “social” and the “search” are not independent of each other. Just because we can ask a friend for help about something that formerly we would have plugged into Google, does not mean that we will only turn to social networks for search now. The last paragraph of the post puts it best: that we need to mix the traditional search with new social media. I have discussed some of these issues before here: http://brynnevans.com/blog/2009/01/30/why-social-search-wont-topple-google-anytime-soon/
It’s taken Google a long time, in Internet years, to get search good. Unfortunately, most people on the web are idiots. Very few of us have a social network of experts. Most of us have some combination of family and friends, people we could query by using a telephone and/or email (or social media, I suppose, but the rambling, “I’m not really sure I want to build a tree house but the one you built is cool, tell me more” kinds of conversations are not well suited to 140 character conversations). I’d just as soon keep social media — at least as it is right now — as far away as the ranked expertise of the web as possible.
I’m not entirely sure it’s the social aspect people are looking for when it comes to online decision making. Ultimately people are looking for an opinion they can trust and this comes from a recognised source rather than a human opinion. For example if someone was given two set of instructions for building a tree house – one from a random blog they’d stumbled across and one from a national newspaper website which one would they choose?
It’s all about establishing those trusted sites right now, and at the moment, the online user trusts Google more than anyone else to do this.
To both Nora and Craig’s points: true. But there is such a wide spectrum of information needs in search and even a hidden set of interests/knowledge from among your social network — that it seems silly not to try to merge them.
Yes, there will always be obscure searches where no one in your social network will know. “I heard of this “kuksa” cup from Lapland. Where can I buy one?”
But trust can be established even among strangers. I might actually trust a random blog over a national newspaper depending on the topic. Blogs have character and identities (of the real people behind them) — this I often trust more than a newspaper (or professional journalism) since I believe many journalists are just doing a “job” which is often persuaded by editors’ and stakeholders’ opinions.
It really depends. But since there is vast people knowledge and tacit knowledge available in our social networks (and the conversations are often more substantive than “I like building treehouses too”), I believe search will benefit from trying to exploit these networks in some way.
Some data about this specific topic by Nielsen
Social Media: The Next Great Gateway for Content Discovery?
But trust can be established even among strangers. I might actually trust a random blog over a national newspaper depending on the topic.
In our work, we found that when looking for experts in an industrial setting, people prefer those they know (or can connect to through mutual friends) over ‘identified’ experts. A lot of this is related to the fact that social connections increase the likelihood of response.
In relation to Brynn’s point about strangers, clearly there is a need for more research on credibility and trust of unknown sources. For example, some might actually trust newspapers over blogs. We’ve tried to get at some of this using the concept of ‘signals’.
These publications are available from my website as well http://domino.research.ibm.com/comm/research_people.nsf/pages/sadat.index.html
Working for a company that offers real-time social media search at the Enterprise level, I really have to wonder how effective Twitter and Facebook will be in meeting users’ requirements.
I think the experiences will be less than satisfying. Consumer-ready real-time social media search will be much more complicated than simply showing people who said what in the last five minutes.
For argument’s sake, let us believe that I am motivated to build a treehouse. Standard houses just don’t cut it for me and I am compelled to abide in the canopy of my evergreen forest.
I have downloaded a pdf file from a DoItYourself blog, that was on the first page of Google SERPS, so I trust this site. Why do I trust this site? Because it was on the first page of a Google search.
Now if I have any grey hairs, I would do this next step before attempting to erect the house. I would find a forum, a blog, or chatroom, and search for any experiential knowledge of building a tree house. Why not ask my own cybersocial group? Because I want a network of DIY minded people, not my buddy Jim who is “Hating rush hour traffic, Arrgggh!”
It is fair to assume however that if my cybersocial group is large enough, that I am likely to find at least one person with a working knowledge of treehouses. But why waist that time waiting for a response, when that personalized knowledge has already been archived?
So how do we find it? Well this is what I just did in the past five minutes.
I searched for ‘diy forum’.
Copied the second forum in the results which was http://forum.doityourself.com
Did a new search for
It returned 60 pages of wonderful, interactive, experiential knowledge. And this was only 1 site using only 1 term. You could also search ‘diy chatrom’ or ‘contruction forum’ or try any of the other 9 sites that came back from ‘diy forum’.
But I think most people do not know how to do this. So the answer would be some type of SE that solely searches through forums, chatrooms, etc. Either that or make ‘Internet Search’ as fundamental as Algebra.
it is just totally hackish for one to equate Google’s search engine with Twitter’s drivel or pictures of little suzy on Facebook. Idiotic.
I don’t think anyone is equating the two. One is like non-fiction, and the other is like an auto-biography. Both have their merits, it’s just a matter of melding the two.
I like social media but I will always go to Google when I want an answer. It’s great that I have connections on Facebook and Twitter that can help me with a real-time answer to a question I have but when the question is about a complex topic then I’ll want to get an informative article from an authority site pulled back by Google.
It would be great to see how the development of real-time and traditional search will merge in the future.
I would do this next step before attempting to erect the house. I would find a forum, a blog, or chatroom, and search for any experiential knowledge of building a tree house. Why not ask my own cybersocial group?