Roughly 30 percent of web searches are for pages that a searcher has actually seen before. That's a pretty wild statistic. It tells us that our existing bookmarking and "history search" mechanisms aren't quite doing the job.
Let's call this the "Find Again" problem. What does social search mean, in the context of the Find Again problem?
Seeing a Shared Page, via Facebook
Let's say that a Facebook friend shares a page with me. The page moves through my information feed, briefly, barely registering. As I see the shared page, I think, "Ah, that's interesting, Steve is saying that this is a great article by John Battelle on how Microsoft's search share just eked past Yahoo's -- I should read that when I get a chance." As time moves on, Steve's recommendation drifts into the past.
If I wanted to search for that recommended page, how would I do that? Would I go to a general-purpose reference search engine, and search for "John Battelle," or any of the other terms I saw in Steve's recommendation? This wouldn't work very well.
It would be hard to find the exact page Steve was talking about. The web has become a machine that takes information and forces it to rapidly metastasize through the body of the Internet. Articles are chopped-up and repurposed, so I know that I'll find all sorts of redundant mentions of the original article, but I'll have trouble finding the specific page my friend recommended.
This is the "Find Social" problem, and it comes up in Facebook, Twitter, LinkedIn, and in any other service that creates a stream of data that's pushed to me by my friends, where the stream can contain recommendations for, and mentions of, web pages.
Circles of "Seen-ness"
The Find Again and Find Social problems are actually just variants of the same idea. A person has looked at a page, and the fact that they've looked at it -- have deployed valuable attention assets against it -- is what matters. I want to search the web according to who saw the page, and possibly when.
Well, this takes us to this diagram:
Pages I've Seen
In the middle of the diagram we have the set of pages I've seen. I've explicitly visited these pages in my browser, and when I search for them, I expect to find them based on page content, pulled from my surfing history.
Pages my Friends Have Seen
The next circle out includes the pages my friends have seen. This set of pages includes the one shared by my friend Steve about the John Battelle article. When I search for pages like this, I expect to be able to find them based on the name of the friend who shared it, comments made about it, or the page content.
Pages People in my Extended Social Network Have Seen
The next circle after that includes pages that people in my extended social network have seen. My friend Steve has a friend Bertha who has "Liked" a page in Facebook. That action should carry some weight, so that I can search through my extended social network, in Facebook and elsewhere, to find pages that rank highly according to the attention information of friends of friends, and so on.
Pages Some People, Somewhere, Have Seen
The next circle out includes pages that some people, somewhere, have seen. This is what Borislav Agapiev calls the "Attention Frontier" -- it's explicitly the set of pages that actual human beings have looked at, and voted for, with their limited attention budget. Searching inside this set returns pages that some human being thought worth visiting, even if you don't personally know them, and also don't know anyone who knows them (ad infinitum).
Pages that Exist on the Web
The final circle in the diagram includes pages that exist on the web, regardless of whether any person has ever read/visited them. These are sad and lonely pages, published but never read, pining for attention. The content farm industry has page-generation machines that crank out an astounding volume of pages on a semi-automatic basis.
Is This Perspective Useful?
The Find Again problem is just a special case of general search, where I know that I've seen a page and want to find it again. The Find Social problem is just like the Find Again problem, but I want to find pages that people I know (and can name) I've seen.
Stepping out through the circles of "seen-ness" moves me further away from what I know to be good and relevant. Some things I can find in my inner history circle, and others require me to loosen the social reins, so that I can find pages of ever-more-questionable social pedigree.
This picture tells us that all search is really social search, with one extreme being search of my own personal history, and the other extreme being search of what a large group of anonymous people are looking at on the web.
Ask yourself this: what use is it to search over a set of pages that nobody has ever looked at? In this view, if we remove the "social" from search, then that's all we have left, and that's not very interesting or useful.
Optimising Digital Marketing Campaigns with Search, Social and Analytics
At SES London (9-11 Feb) you'll get an overview of the latest tools, tips, and tactics in Paid, Owned, Earned, Integrated Media and Business Intelligence to streamline your marketing campaigns in 2015. Register by 31 October to take advantage of Early Bird Rates.