Disclaimer: I have been studying social search from an academic perspective for about a year. I’m sure that I’m not aware of all the “social search” solutions already available on the web and their possible implications as intimately as I probably should be.
However, in reflecting on the strengths and weaknesses of social search, I don’t think it’s going to “topple” Google anytime soon.
There are four important points to consider here: (1) why Google is good at what it does; (2) the drawbacks to Google search; (3) social search as a potential solution; and (4) the limitations of social search. I do believe that there is a place for socially-augmented search, but it will be as a supplement to, not a replacement for, Google.
Why Google is good at what it does
There is a huge advantage to Google search as it currently exists. Google scans and indexes vast amounts of information, catalogs it in databases, and provides quick and efficient mechanisms for retrieval. Your average Google search takes milliseconds to complete, which leaves plenty of time for you, the searcher, to review the result listings, synthesize the information, and possibly refine and iterate on your search. Think about it: Without search engines, all of our time would be spent searching for and retrieving information from a much smaller corpora (based on our very real human processing limitations). We would be left with little time to contemplate our search results. Google, as a “tool”, is a scaffold for the cognitive process of search. For social search to topple Google, it would need to provide a greater scaffold than Google already does.
Additionally, people are hooked on Google. (This point should not be taken lightly!) People are relying more and more on the web for their daily information needs. In support of this, the PEW Internet & American Life Project  recently found that broadband and dial-up users prefer using the Internet over friends and family for help in solving problems. Of course, this isn’t the case for all information needs; community networks are still preferred in certain circumstances (e.g., choosing a new bank, dealing with a spiritual crisis). However, when asked why these individuals prefer web search to asking friends and family, they report two main reasons. First, the web can provide them with information more quickly. Second, the content they receive may be more accurate (presuming that it comes from a larger pool of potentially unbiased and/or more authoritative sources). These are strong motivations for users to continue using the web (and Google) to satisfy their information needs. The PEW study’s conclusion is that social networks are still important sources of information, but that the Internet supplements and facilitates communication and information seeking. I expect the same with social search—that its power lies in its ability to supplement and augment our current search practices.
The drawbacks to Google search
However, there are limitations with our current search tools. The most notable drawback is that Google can’t always satisfy our information needs. This may occur for several reasons, perhaps because some inquiries aren’t well suited for a Google approach (Google deals best with facts, historical references, etc.) or because human search terms don’t match Google’s index terms (the human-system “vocabulary problem” ). Searchers who are unable to properly describe their search problem or whose query is not neatly addressed in the first page of hits may experience failures. It’s unclear at this point how frequently search failures occur. (Errors in accounting for this may always exist since people with difficult queries may be disinclined from even trying Google). On the other hand, Ed Chi & I have found that 87% of failed search reports were from informational queries, where the question was open-ended or the information need poorly defined. These searches are not likely to be addressed by simple facts present on the first page of search results, and would be expected to produce (occasional) failures with current search solutions.
Furthermore, “real world” searches extend across physical and digital media, include friends, family, and coworkers for support, and may even span several sessions, days, or weeks . This has been observed in libraries and other physical settings [13, 15] and more recently in web search studies [5, 11]. A major drawback with current search tools is that they rarely capture (and address) the context and complexity of these types of searches. Simply put, Google is pretty dumb. It does what it does really well—indexing and retrieving data from large databases. But it does not know what room we’re in or what our surroundings are like; it does not know our recent activities; and it does not know who our friends are. I’m not advocating that we’d ever expect (or want) a search engine to become our personal assistant in this manner. Instead, I see Google’s drawbacks as an opportunity for social search.
Social search as a potential solution
The benefits of social search (on the web) are mostly theoretical at this point, although there is mounting evidence that social interactions provide critical support during search and problem solving. For example, people can provide solutions (advice, guidance, assistance) [1, 3, 8], pointers to databases or other people [3, 6], validation and legitimation of ideas [3, 5], and can serve as memory aids .
Plus, social interactions are useful for more than conveying knowledge or supplying certain facts. People can help with problem reformulation  and brainstorming . “Guided participation”  is a process in which people co-construct knowledge in concert with peers in their community . Social psychologists have recently found that social discussions facilitate subsequent cognitive performance . These findings suggest that social interactions could naturally aid users in web search tasks, although we cannot predict how these interactions will play out in practice.
Several “flavors” of social search solutions already exist on the web, each with their own strengths and weaknesses. Some services are collective harvesters, presenting content based on popularity or emerging topics (Digg, OneRiot), or re-ranking search results based on user annotations (WikiaSearch, Google’s SearchWiki). These solutions, although available instantaneously, may lack relevance or validity since they are pooled from anonymous users across the web. Others try to connect searchers to experts (Yahoo! Answers, MetaFiler, Aardvark), providing personalized answers but at a serious cost (a time delay). Searching through social networks is another solution that has the potential to be both personalized and expedient. For example, a search in FriendFeed first returns content that your friends have shared with the system. Delver presents data shared by your social graph (elsewhere) intermixed with their search result listings. Undoubtedly Facebook and Google (who arguably have social graph data through Gmail contacts) are also considering solutions in this direction. While these solutions must require a lot of computing power (among other social smarts), I believe that a personalized social graph solution, if integrated with Google, will be the most successful in addressing real search problems.
Limitations of social search
One concern with these existing social search tools (and others) is that they aren’t quite the solution to the search problem yet. Do these services address real user needs? Is this the type of social support that will be most useful during difficult queries or failed search attempts? As this is yet to be determined, it’s worthwhile reviewing the current limitations of social search.
An obvious limitation is that information from social resources may not be as complete or thorough as information stored in large search engine databases. This is an issue of scale and scope. A social search only solution will lack either the scale or the scope that a Google search can provide. This criticism is not entirely a fair, however, if social inputs are used primarily to augment regular result listings, instead of replace them.
A narrower scale and scope may actually be advantageous by helping searchers in one of two ways. First, since we naturally seek information from other people directly, receiving suggestions of experts or friends to help with a certain problem may be quite useful. Second, socially-filtered data may be more trustworthy, especially if the filtered information comes from people you know.
On the other hand, a searcher may not have knowledgeable or available contacts for a given query. In these circumstances, passively receiving social recommendations may be a good solution. My current work is looking at the tradeoffs between asking friends directly for help and receiving somewhat anonymous replies from large social networks (a la social recommendations) during search and sensemaking tasks. These strategies provide complimentary support for searchers: While the latter may present a diverse set of views on a topic (establishing breadth), the former may help develop depth on individual topics.
Finally, social inputs won’t be the right solution for every user query. Certain questions are likely to be improved by social search (experience-, opinion-based, and tacit information), but others may be hindered. Open-ended, exploratory queries may benefit from social inputs, simple navigational ones may not . To avoid cluttering the interface or cognitively overloading the user, an optimal solution would deliver social solutions appropriate to the target circumstance. Will this type of inference and personalization be possible with search tools?
On the whole, social search is still in its infancy. Even Google, which has been around for 10 years, continually tweaks its search algorithms. It’s naive to think that we will have truly viable, integrated social solutions to the search problem over night.
Unfortunately, current social search tools are fairly crude, presenting solutions to (more simple) algorithmic problems instead of to user problems. This is not to say that the work of social search engines is a easy, but rather that they may be working on the wrong questions. Very few current social tools integrate with Google (with a few exceptions, whose value has yet to be demonstrated: CloudLet and Qitera). Since it’s unlikely that users will rush to adopt new search tools at this point, an integrated, combined approach (Google + social) seems more practical.
Beyond this, we need to know more about how social inputs, interactions, and passive social recommendations can address user needs during search, to help focus targeted solutions to real problems. I believe that Google search will continue to be a part of this solution; I also believe that social search has the potential to supplement Google in powerful ways.
 Bandura, A. Social Cognitive Theory. In R. Vasta (Ed.), Annals of child development. Vol. 6. Six theories of child development (1989), 1-60.
 Churchill, E. Of candied herbs and happy babies: Seeking and searching on your own terms. Interactions 15, 6 (2008), 46-49.
 Cross, R., Rice, R.E., and Parker, A. Information seeking in social context: Structural influences and receipt of information benefits. IEEE Transactions on Systems, Man and Cybernetics — Part C 31, 4 (2001), 438-448.
 Edelson, D., Pea, R., & Gomez, L. Constructivism in the Collaboratory. In B. G. Wilson (Ed.), Constructivist learning environments: Case studies in instructional design. Englewood Cliffs, NJ: Educational Technology Publications (1996), 151-164.
 Evans, B.M. and Chi, E.H. Towards a model of understanding social search. Proc. CSCW’08, ACM Press (2008), 485-494.
 Fox, E.A., Hix, D., Nowell, L.T., Brueni, D.J., Wake, D.C., Heath, L.S., and Rao, D. Users, user interfaces, and objects–envison, a digital library. JASIS 44, 8 (1993), 480–491.
 Furnas, G.W., Landauer, T.K., Gomez, L.M., and Dumais, S.T. The Vocabulary-Problem in Human-System Communication. Comm. of the ACM 30, 11 (1987), 964-971.
 Hatch, T., and Gardner, H. Finding cognition in the classroom: An expanded view of human intelligence. In G. Salomon (Ed.), Distributed cognitions: Psychological and educational considerations. NY: Cambridge University Press (1993), 164-187.
 Karasavvidis, I. Distributed Cognition and Educational Practice. Journal of Iterative Learning Research 13, 1/2 (2002), 11-29.
 Kuhlthau, C.C. Inside the search process: Information seeking from the user’s perspective. JASIS 42, 5 (1991), 361–371.
 Morris, M.R. A survey of collaborative web search practices. Proc. CHI’08, ACM Press (2008), 1657-1660.
 Rogoff, B. Apprenticeship in thinking: Cognitive development in social context. New York: Oxford University Press (1990).
 Twidale, M.B., Nichols, D.M., and Paice, C.D. Browsing is a collaborative process. Information Processing and Management 33, 6 (1997), 761-783.
 Wells, A.T. and Rainie, L. The Internet as social ally. First Monday 13, 11 (2008), retrieved from: http://firstmonday.org/htbin/cgiwrap/bin/ojs/index.php/fm/article/view/2198/2051.
 Wilson, T. On user studies and information needs. Journal of Documentation 37, 1 (1981), 3-15.
 Ybarra, O., Burnstein, E., Winkielman, P., Keller, M., Manis, M., Chan, E., and Rodrigues, J. Mental exercising through simple socializing: Social interaction promotes general cognitive functioning. Personality and Social Psychology Bulletin 34, 2 (2008), 248-259.