Social search - KM thinking

Social search is touted as the next big thing for improving information retrieval, relevance and awareness. Let's take a look.

What exactly is Social Search?
There is no clear answer as the field is emerging and changing at a rapid pace. Here is one early definition: "..a collection of Internet wayfinding tools informed by human judgment. That judgment takes place in the form of tags, click-through activity, search history, and other actions". Source These technologies are being applied to bookmarks, images, tags, blogs, bibliographies.

Social search comes in many favors. New engines are riding the web2.0 wave making it difficult to evaluate progress in this heaving landscape.

Subscribing to a Flickr, del.icio.us, diigo or technorati tag via RSS - allows you to connect to a community, annotate, scan for recency, popularity or some rating measure as applied to posted images, bookmarks, URL links or blog posts. This brings new finds directly into your aggregator helping keep you up-to-date and raising your awareness.

Scanning or searching Digg or Wink - helps you quickly zero in on news, posts and items others have rated as interesting, worthwhile or can be used for finding experts.

Social search engines such as Eurekster, Prefound, Searchles,  Gravee, Collarity, Zimbio,  .... claim to use collaborative filtering, relevance rankings, community activity & behavior, 'collections', unique ranking scores to improve search returns, provide a 'personal touch', guide inquiry and add 'meaning' to those coded search algorithms the big boys use.

Affordances
What do we need to make social search really work?

• Dedicated community - people who share your interests, are sincere, active, honest and helpful - not always easy to find and maintain.
• Visual help and tools to refine a search - Quintura looks interesting with their interactive keyword clouds.
• Permanent URL - so individual searches can be stored, shared and updated.
• Ranking or scoring mechanism - simple but intelligent enough to prevent obvious spamming and gaming.
• RSS feeds - so you are alerted when rankings change, a repeated search yields new findings or friends provide annotations.
• An intuitive back-channel and community directory - to converse around results, rankings and relevance.
• Critical mass - there is a tipping point when social search offers greater value, improved relevance and increased awareness - no single engine is there yet.

So is your search social- yet?

Comments

[David Locke]

So KM boils down to this, "I don't know what I am, so I'll make sure you don't know what you are." Or, "I can't do what I promise, so hey, look at this...."

Try defining KM as if no infrastructure is necessary at all, and then maybe, we'll get there. The only problem is that we wouldn't have anything to sell, and no reason to go there. So sell everything and don't ask any real questions.

[David Locke]

By social, you mean centralized. I've come across a paper on address-free architecture that emerges from diffusion gradients and flooding.

In diffusion, one node expresses a want to those nodes known to it. Those nodes either provide the wanted, or forward the want to those nodes known to them. This goes on until the want can be serviced, then that service is sent back via the path that it travelled. The layers constitute a gradient.

Flooding is similar. Yell out into a room "Sort yourself out with A's by the door. And, in less than n squared time, the room is sorted. No centrality involved.

This stuff is both faster, cheaper than centrality. True distribution is yet to be achieved.

There is a long way to go. And, Web 2.0 is utterly a distraction. But, KM must go on distracting itself, as it is a distraction.

[Garsett Larosse]

I like Swicky and Frappr.
The challenge, it seems to me, is to integrate social search in the organisation's social assets.
For example, we've added Swicky-like functionality to a competitor research extranet. Frappr is a good tool to integrate to any geographically distributed organisation.

[Bryan Alexander]

What about using Google search strings (like this one) in other web 2.0 apps, like blogs or wikis? The social affordances of the latter then add social power to the former. Perhaps this is second-order social, in your terms, Denham.