Social Metadata and the Relevance Revolution
Consider this:
* Google now uses over 100 pieces of information gleaned from your online behavior to serve you the most personally relevant ads.
* Applications like del.icio.us and Flickr enable a kind of massively multiplayer information architecture - collaborative structuring of data that supports personal, local and global views of the information space.
* And social networking sites promise to increase the relevance of everything from your personal network (Linkedin) to your RSS feeds (Rojo)Social metadata - information about who we are, who we know, who we trust and what we do – is the fulcrum for new applications that leverage an understanding of your behavior to bring you meaningful content, products and services. It's the foundation of the collective intelligence systems we see at work in Amazon, Google and eBay.
When Tim O'Reilly says collective intelligence will change how we live in the next few years, he's talking about a relevance revolution built on social metadata. It means using our own information – clickstreams, purchases, preferences, blog posts, tags “ to help us find more of what we want and eliminate the billions of things we don't.
This revolution will bring us powerful new tools to discover and recover information, help us make decisions and maintain connections. We will also be confronted with important questions about authority, identity, security and trust.
This talk will explore how social metadata is changing our information landscape, and the opportunities and challenges it presents. We'll also look at how we can take advantage of social metadata and design systems that tap into this collective intelligence.
Speakers: Gene Smith
download the presentation at Gene's website.
My notes from the session:
IA is the structural design of shared information environments.
Shared design of semi-structured information environment.
Using the wisdom of crowds to solve the problems of IA
- Find, use and interact in information environment
Read the book - the wisdom of the crowd. Good book to have around to read.
Example:
Amazon
Wikipedia
And of course, flickr, and its tagging system. (the cluster system. Groups, contacts.)
- the entire experience of flickr is built around user contribution.
Google has this AI system that tracks what you’ve been reading in the passing hour and give you the best system.
Social search:
Rollyo
linkedin
(who ever has the most friends wins)
three ingredients of social IA
- capture user actions
- what does user actions mean?
- Things people do online that we can track. (web log file etc.)
- Building blocks for popularity, community, reputation. (it's important to have the popularity part in order to build a more attractive social network)
- Ignore higher goals & motivation
- aggregate and display
- feedback
trackback is dead. (why?)
it's pretty low engagement in terms of a social thing.
Last.fm sends to the server what you listen to.
Once again, low engagement.
(Though, to think about it, last.fm has just launched a new interface to accompany more social interaction for its users. my dearest internet friend was just blogging about it the other day)
Kinds of aggregation:
- listing
- count an action
- clustering
- collaborative filtering
- other algorithms
- example: ebay
- prototagging (basically the weird lingo ebay ues)
- ranking
- order them
- example: youtube navigation toolbar
- yahoo's most recommended photo
- (girls in bikini will never be most recommended but will never be most viewed)
- amazon
- netflix
- digg
- whole bunch of factors are being included.
A feedback loop is a system where outputs are fed back into the system as inputs, increasing or decreasing effects - Wikipedia
Exercise of positive feedback and negative feedback.
Positive: All the people around you – if anyone has the hand up you put up
Negative: only the front and left as indicators – you raise your hand after 1 second delay and only raise it for 3 seconds. (it basically creates a wave)
Positive feedback makes it grow
Negative feedback contains it.
Everyone loves to use digg as an example for the social network/community/IA
Places to intervene
- introduce delays
- modify the strength of feedback loops
- who has access to what information?
- Adjust incentives and punishments
- Change the sytem
Example: google. Aggregating links in a different way.
Challenges:
- spam
- gaming
- balance
- relevance
- unintended consequencesdesign principles:
- allow for different levels of engagement
- monitor and tweak feedback loops
- participate in larger ecosystem
-- youtube is viral
- design new actions, aggregators, display
