First to Market, or First to Rate?

In the world of business the success of your idea depends hugely on word of mouth and feedback, especially with the Internet. The same is true of music, with new bands scrabbling to get any bandwidth with already saturated music lovers. In such a situation we often rely on music taste makers that we know and trust to point us towards new music we might like. This was a subject that Malcolm Gladwell discussed in his excellent book ‘The Tipping Point‘, where he described such people or groups as ‘super influencers’ who have an effect on many people’s opinions and can even make or break new products.

Well it seems not everyone believes Gladwell’s view of how information is dispersed. In a recent article Fast Company talks with Duncan Watts, a Network Theorist, on his own views of how information travels. He feels that information sharing in tightly grouped areas such as music depends much more on who gets the first attention, rather than who is referenced by the most connected people. This was found in an experiment where he set up new songs on a music sharing site, and then asked different groups to rate the music – but some of the groups had a social aspect, and others didn’t. What transpired was that the top rated list of songs was totally different for each group, there was no shared ‘best song’. The result for the ‘social’ world seemed to be driven more by who got the first votes, not who was actually best.

This result has some intuitive merit. In today’s social information networks, information that is tagged most rises to the top of display lists such as ‘Recent Top 10’ or similar. Other people then see these lists and give their own opinions on this music, perhaps biased by its position in the list, instead of going round all available music and seeing what they like. In this case the influencer is the social group en masse, rather than a single super influencer. Either way it seems that getting out there first with your product into a new area is critical, although as Watts says – “In general, the ‘best’ songs never do very badly, and the ‘worst’ songs never do extremely well, but almost any other result is possible&rdquo – so make sure your product is worthy of any attention is receives and you can’t go far wrong. Or if it’s not great, then make sure you can get feedback from this failed attempt to make sure the next one is worthy.

[From Boing Boing]

Squidoo – Crowd Sourced SEO Friendliness

For some reason that escapes me now I started to spend a few hours playing around with Squidoo to create my own ‘lens’ on Smorkin’ Labbits, those cute little collectible rabbits that smoke. Squidoo was created by Seth Godin, amongst others, to enable experts (and fans) with minimal internet experience to create, manage and write their own spiel about the subjects they love. It’s sort of a crowd-sourced About.com, with the added fun of having to identify spam type posts.

The overall experience of creating my lens was straight-forward and powerful, although I found that one limitation of Squidoo is that it doesn’t like you uploading pictures to make lists – something that I really wanted to do to list out the colourful little toys. The solution: create a false Flickr account, upload my pictures there, then hand transfer every image URL since the ‘Plaxo’ image grabbing code seems not to be working right now. Ug. What’s interesting now is whether or not I make any money off this activity, as Squidoo borrows another crowd sourced site idea (ePinions’) to share any affiliate revenues they make with us, the creators. Somehow I think the percentage payback on a $5 plastic toy ain’t so great, but guess that’s why all the camera focussed lenses are already highly developed.

Photosynth Presentation at TED

Thanks to Mateo for sending me this great presentation of the PhotoSynth work emerging from Microsoft’s Live Labs.

What’s interesting to me about this is that way back when I read David Siegel’s excellent Futurize your Enterprise (from 1999) he discusses many of these kinds of applications made possible by fully described semantic data. This led me to the concept of being able to search the web to pull up all photos from a physical location, say around the Empire State Building, and then stitch them all together into a single interaction point blending everyone’s photos. What really threw me about this presentation is that the way they zoomed round Notre Dame was almost identical to how I’d thought it might work. Now the great thing about this is that they don’t rely on positioning data, but instead visual reference points that they automatically detect and stitch together, so I’m certainly looking forward to the model they create of everyone’s photos from the top of the Empire State.

Update: Having played around with the online demo (which is stunning and intuitive) I was struck by the similarity to the sequence in Blade Runner where the main protoganist Rick Deckard zooms into a slightly holographic photograph of one of the replicants he’s chasing. The way the screen panned as he focussed on sections and ‘looked round’ shower curtains is totally how PhotoSynth feels in use. The future is here again.