…All these issues affect how you want your information to be visible from one network to the next and over time. Behaviours emerge from networks, they are not predicted or calculated in advance…

During February 2010 Google announced another step into social networks – Google Buzz. At the time, the early adopter crowd raved about its potential. And then came the privacy concerns, as information people thought was private or shared only between a discrete group suddenly became public. To Google’s credit, they have been quick to address some of those concerns  but I am surprised how unprepared Google has been for the backlash over messing with people’s social networks. Is their worship of algorithms becoming their Achilles heel?

Back in 2005 Google co-founder Sergey Brin gave a talk to a class at UC Berkeley (no longer online)

At around 16:30 mins, Sergey makes the following comment:

Semantics and tagging are great as long as computers are doing it… I’m a believer in innovative algorithms that can extract structure and knowledge

Google’s approach to search has been all about algorithms and automation – don’t let people make the decisions. There will be a never-ending battle with the search engine optimisation companies trying to game the algorithms and insert results we don’t want or need, but that’s just commerce at play. Sergey is correct – when it comes to consistently classifying great globs of information across every imaginable subject, computers do it better than people. But is the same true when it comes to classifying relationships within social networks?

In 2005, Ross Mayfield presented the following table at one of the first conferences to explore social media – Our Social World, held at Cambridge University in the UK: The Ecosystem of Networks:

1 12 150 1000+
Individual node
Creative Networks
Social Networks
Political Networks

Small group networks have very different dynamics to large tribal networks. In a group of 12 or so, you are either in a peer group, family group or team brought together for some purpose. Chances are most people will know most others in the network. With Dunbar’s magic number of 150, you are in an organisational network. I first came across Dunbar’s number in an article some years ago describing a company that constructs all buildings with car parking for 150. When they run out of parking spaces, it’s time for a new building. Organisational networks are more about communication of a message than creating the message in the first place. Where creativity does happen, it is typically by clusters of group networks within the organisation. Organisational networks of 150 are still small enough to challenge the message if you don’t agree with it. Tribal networks are another matter entirely. When you head into the realms of a network containing 1,000+ people, it is inherently political in nature. There’s usually a celebrity at the top publishing their world view and the followers are there because they buy-in to the message. They’re not there to challenge it – that usually requires an opposing political network, the anti-.

As well as differences in size, networks have different types. Some networks are permanent(ish), like family, long-term friends and associates. Others are are temporary, planned or otherwise – building a house, working for a company, launching a product, challenging a court case, attending Weight Watchers, getting divorced… And temporary can be anything from a few days to several years. Your individual behaviour will depend on your role in a given network – you may lead some, participate in others or simply be following a topic of interest, and whether you are an introvert or extrovert by nature. It will also depend on the type of network – you might be very social in work scenarios but more private within family situations. And your behaviour will change over time, as interests come and go, confidence grows (or shrinks) and your style of communication evolves.

How does this all relate to Google Buzz? All these issues affect how you want your information to be visible from one network to the next and over time. Behaviours emerge from networks, they are not easily predicted or calculated in advance.

“In the same way manual search broke down, social has broken down… When I had fifty friends, I could mine the stream, but as I have 500 or 5,000 friends, that breaks down, and the concept of who is a ‘friend’ has expanded by several orders of magnitude. It’s a mishmash of information and it’s no longer fun to dip into that. It has become a Google scale problem.”

– Bradley Horvitz, Google Vice President

Our social networks have become a Google scale problem? That assumption is either incredibly naive or arrogant. Or hidden between the lines is a concern that we are beginning to rely more on people than search to find what we need… A situation that has serious implications for Google.

Last year I wrote a blog post ‘Does Search Matter‘ questioning whether we need a ‘SocialRank’ to replace PageRank. Relevance on the Internet has degraded over the past few years. When searching for information, you increasingly have to scroll through a few pages of results (usually up to 3 before trying again with a different search query) or, if it is information-related, switch to Wikipedia first. There are two faults I find with current search: 1. SEO – search engine optimisers are continually trying to disrupt search results and insert their clients’ content at the top regardless of relevance; 2. Social Media – it’s so easy to publish information, and so many are doing it, that the noise:signal ratio has grown very large, complicated by one person’s noise being another’s signal. There’s a single solution that works around both issues – I look at what other people are linking to (for information searches) or reviewing (for shopping searches). Neither involves Google (or any Internet search engine). Twitter and Amazon are often the starting points respectively.

Despite the furore and lawsuits cropping up, Google Buzz is ultimately a good idea. At the moment, conversations are fragmenting across multiple different networks. FriendFeed, at one point, showed potential to bring them all together and I thought they’d be snapped up by Google until Facebook got there first. It’s handling the public/private nature of those conversations that requires more thought. Messing with privacy is a risky strategy. Facebook did it once and got away with it, with their Beacon fiasco in 2008. Now Google. What next?


Related Blog Posts

iStock_networkFeatured image: iStockPhoto

, , , , , ,
%d bloggers like this: