I know! Let’s build a social search Google killer
2007 has witnessed the most significant advances in online communication and connectedness. This was the year when marketers started to embrace and felt a little threatened by social networks. IT wasn’t really about technology – but the coming to fruition of many of the ideas and desires to connect.
But during this year, one thing that seemed to have stayed stagnant is search. At a time when people power dominated and shaped everything from how we communicate to make purchase decisions, to most people search engines seem to be lagging behind – leading some ‘experts’ to declare the death of Google.
So did Google stand still? Well I’m confident they didn’t. But before we can look at where Google is at and where they’ll likely end up – it’s worth looking at some Google history and some of the barriers to social search.
Most people think that Google’s primary contribution to search was the use of linkage data to judge reputation. But a bigger contribution, the one I believe is the reason that I’ll still be Googling to my grave is the technology and methods they developed to scale and efficiently deliver information to where ever you are. Add to this a ranking method which can be swapped in and out like a removable hard disk – and you have the infrastructure for longevity. Oh and let’s not forget Spam – not something to be taken lightly.
What’s up with the algorithm
Link reputation has served us well, but letting the published community define reputation and relevance over the people that consume that content has limited appeal. More about that here - From link spam o user data
As reputation dominates other factors in the ranking game – Is Google constantly pointing us into the ‘head’ of content on a given topic. As Chris Anderson pointed out in his groundbreaking book The Long Tail, what’s the point in having a tail if you can’t get to it?
User driven recommendation engines like Amazon and Last FM do this very well but can Google really reach those parts – or maybe that’s not what we want out of our reputation system? A navigation aid and nothing more.
Developments this year
People write search algorithms not computers. So you could argue that Google like all search engines were already influenced by people. It’s just that those people worked with the data available to them.
Personalisation ramped up
Using search history and other signals from user data such as bookmarks and content saved in your Google home page is a move towards employing more user data to judge relevance and more importantly query context or intention. But the use of data is confined to the individual.
Blended results – Google universal
On the face of it blended results seem more an aid to navigation than anything else. I.e. save people having to explore vertical indices to find what they’re looking for by attempting to understand and deliver vertical results into regular Google. Danny Sullivan introduced this concept more than five years ago and called it invisible tabs.
But the really interesting development her e was the ability to use popularity data in verticals in order to be able to identify when a vertical result is relevant for the masses. Google can identify if a news story is relevant for regular result in less than an hour of publication.
So we’re starting to see the use of user data in quite a big way already within Google.
Search 3.0 – blended and vertical http://searchengineland.com/071127-091128.php
Google Knol
http://searchengineland.com/071213-213400.php
Of course, Manber did say that Google could better tell which of the Knol pages were of high quality by looking at signals such as ratings.
Going all the way
So what stops Google from using user data to refine search results in a more direct way – Spam (remember the Direct Hit Algorithm) trust in everyone’s opinion is just one of the issues.
However Google is already using popularity data to determine search quality. All clicks on a results page are tracked and high recall rate can be used for further tweaking.
Toolbar data and even ISP data could be used to determine the relevance of a site to a given query. But if Open standards like OpenSocial and OpenID are adopted on mass then many of the Spam and trust issues could be overcome through the use of personal reputation scores like personal Pagerank. APML will no doubt play role in enabling people, content and behaviour to be connected in one seamless network or as Tim Burners Lee refers to as The Graph.
One thing is for certain, we are unlikely to see a set of open standards designed by committee and adopted by everyone to take us to the next generation of the web (Semantic web). We are more likely to see adoption based on the need for attention and economic growth. Until one day we’ll find that we’ve arrived. Just like when O’Rilley coined the phrase web 2.0 – no one planned it or named it – things just emerged that way.
2008 will be one of the most interesting years, not just for search but for the internet. Technology was never the limitation – people will ultimately decide how, why and who they wish to connect and engage with – marketing can only follow.
guns don’t kill people, rappers do
Did you enjoy this post? Why not leave a comment below and continue the conversation, or subscribe to my feed and get articles like this delivered automatically to your feed reader.
Comments
No comments yet.
Leave a comment