PreFound: Hyper-Relevance through the Extension of Social Search Tools into Social Networks
The social search roundtable I’ll be tonight for PreFound is called Extending On-line Social Communities into Pure Search.
In a recent post on the roundtable Steve Mansfield, PreFound’s CEO described it as an event to explore “extending social networks into social search.”
Re-Framing the Conversation
In some ways I think it makes more sense to frame the conversation as extending social search into social networks, in as much as the social search tools he’s offering must adapt to the needs of a given network/community to be truly useful. I’ll be sure to spring that one on Steve during the round table
This post is intended to introduce you to PreFound by way of investigating what we’re going to be discussing at the round table tonight.
Here’s Steve’s premise: “at their level of scale, online social communities have the ability to gather, organize, recognize, reference and tag related online content. This has the potential to far surpass the abilities of existing algorithmic processes for demographic relevance compared to traditional Search.”
(also read Yahoo Social Search, Act II by Ross Mayfield. I wish I had before writing this article
That’s Just Personalized Search, Right?
What Steve’s essentially saying is that algos that utilize social networking data – with pages ranked by community relevance and community usage – have a stronger chance of returning relevant search results than a purely algo-ranked method.
Applying community data to search relevance is not a new premise – Eurekster for example enables a site owner to develop a fairly unique index and the algo adjusts sites’ rank in the results with click stream data.
Many major search engines too have enabled users to benefit in some way from group searching.
So… how would this be different from personalized search then?
This is where we have to begin talking about PreFound and the tools they’ve developed for their searching community.
Turning Searchers into Algos and Crawlers
At the core of what makes PreFound most interesting to me is PFfinder, the search tool that enables the PreFound community to save and tag any kind of data so that this data will appear for given search results.
In addition, PreFound makes it easy for its “Finders” to upload its bookmarks, which are uploaded on the assumption that a given person’s bookmarks are more than likely highly useful sites. Finders can also just type their sites and tags in directly.
So it’s PreFound’s resource sharing emphasis that, in my mind, could really push a given online community’s results towards hyper-relevance (check out ReadExpress for another model in hyper-relevance). Especially if you worked in a Eurekster-style click-stream measuring community-oriented auto-correcting algorithm.
Spam vs. Community
So let’s assume that hyper-relevance is possible through extending social search into social networks and people decide to join networks solely to improve their search results.
What happens in the Mesothelioma community? How about the click fraud lawsuit community? (I predict that “click fraud lawyer” will soon be a VERY EXPENSIVE keyword phrase
Will it be possible for communities that legitimately focus on topics with high click-value key phrases to keep commercially-oriented content (…spam) at a minimum? Will it be worth it for participatory marketers to join specific communities to influence that community’s algo relevance?
What would prevent the auto creation of accounts and uploading of algo-busting bookmarks in order to spam the community results?
Can Hyper-Relevance Scale?
Besides participatory spam, my other concern for PreFound is this: it’s my understanding that Steve intends to take on mainstream search through the development of a hyper-relevant community-based engine.
I think that – almost by definition – a given community’s relevance will not deliver mainstream relevance.
Hyper-Relavance as Discovery Tool
Further, the randomness of what people, even within communities, consider relevant makes the PreFound experience more of a DISCOVERY experience, rather than the more utilitarian RECOVERY/NAVIGATIONAL/TRANSACTIONAL experience that users typically seek on mainstream search engines (yes I’ve been reading The Search).
Focus on Enabling Hyper-Relevance Everywhere
I think PreFound should let mainstream search alone and focus solely on the helpful, sharing individuals that exist within EVERY community and enable them to shape search for that community through tagging and other SERPs interactions.
Also, why not simply give searching and tagging tools to existing vertical sites so that their pre-existing community can tag their own results and generate hyper-relevant results for others through their searches and click throughs. Roll this out to the web’s verticals and then give them a cut of the paid search profits.
If PreFound’s model really can deliver hyper-relevance they’ll get far more money in the long-tail of the web’s verticals than they could from their own site (if I’m correct in believing that PreFound can’t become mainstream search due to its focus on community relevance).
If they could also enable a given vertical with the community messaging (soon to be) offered to existing PreFound members you’re likely to see increased site involvement similar to what happens when vertical sites create forums (and you’re likely to see the same kinds of political jostlings and battles that happen in forums too ;).
The Man Behind PreFound
I think that the main reason I’m excited about meeting with Steve and participating in the panel tonight is that I see in him a similar passion for search that Louis Monier had at Alta Vista. Steve believes in the power of the community to develop and define its own relevance and index of content, that the community can be both crawler AND algorithm.
Further, I find his work with tagging compelling – he’s almost convinced me that community tagging CAN scale.
I hold that PreFound is a true Web 2.0 site, from the high-level harnessing the collective intelligence perspective. And, from my perspective, it’s the PFfinder tool that, for now, sets PreFound apart from the likes of Yahoo’s MyWeb because it enables PreFound users to better share their findings.
And finally, I see in Steve’s commitment to engaging the conversation about PreFound, no matter WHAT people are saying, a model for CEOs everywhere: Prefound is Very Web 1.0
Other PreFound Coverage: