Sunday, 20 December 2009

HP software predicts outcome of social content

The boffins at Hewlett Packard's Social Computing Lab reckon they I have come up with a way to predict how 'viral' or popular a piece on content, such as a YouTube vid, would be. The blog post here explains it very well. In summary, they looked at 7000 YouTube videos and 60 million stories on Digg to work out what becomes popular and what doesn't, as part of the testing of a new product called iCatcher.

I'm sure the new software will be snapped up by Ad-land to help predict ad revenue. However, it could also be invaluable in the world of earned media to plan and predict the outcomes of social media campaigns.

Tuesday, 15 December 2009

Conversation tracking in Virtual Worlds

Last week Habbo Hotel launched a brand tracking tool that allows you to monitor conversations about brands and products in the virtual world.

I think this is a very smart move by Habbo as one of the barriers to brands getting involved in virtual worlds is the fear they can't track what is being said about them. Hopefully, this will encourage more companies to engage with people in their own spaces rather than trying to drag them to their corporate website.

Perhaps this move by Habbo will be the kick off for other virtual worlds and communities to launch similar tools.


Sunday, 13 December 2009

Manchester tells its story via social media

I really like the idea that the people at Salford University and friends have come up called New Mornings Old Streets - that is encouraging people to show everyday life in the North West using social media. The initiative is looking for 500 people to submit short videos, audio commentary and other multi-media content describing everyday life in the NW.

It is based on a 1959 BBC production, A Morning in the Streets, that focused on working class life in the region.


The project is open to anyone living or working in the Salford and Greater Manchester area.

Monday, 7 December 2009

Coke Facebook facial recognition app

I came across a nifty Facebook application by Coke that uses facial recognition technology to find people on Facebook that look like you. I believe it's got some fudge factor built in that stops you just finding loads of piccies of yourself.

It's a really cool app that uses cutting edge police-grade tech, but feels a like bit pointless, IMHO. Or maybe I've just lost my sense of frivolity.