I hadn't noticed this feature on amazon.com before - although it seems as if it's been there since April. Product pages on amazon.com now show the aggregated purchase behaviour for similar products. My initial reaction was that this was pretty cool - a great way of using crowd wisdom to inform a purchasing decision. I've blogged a little bit before about the problem of too much choice, and this seems like a great way of narrowing down the options. Which is great - for stuff like cameras, and flash cards, and hard disks. It's not great for books and music, even if, on first glance, the amount of choice seems far larger. But actually, it isn't, or not in the same sense, as the choices aren't mutually exclusive - once you buy one camera, you won't be buying another one for a while. Buying a book, however, doesn't prevent you from buying other books - even similar books. Amazon have, of course, worked this out, as they don't show this information for books and music. Nice feature, and given Amazon's focus on measurement and testing, it's a reasonable assumption that this is having a measurable effect on purchasing. [tags]wisdomofcrowds, crowdwisdom, choice, amazon[/tags]

Cogenz is a social bookmarking service (like del.icio.us), but explicitly designed for enterprises: instead of the del.icio.us model of a huge vat of shared tag soup (ignoring private bookmarks), cogenz pours the soup into different company bowls and only allows you to drink from your own bowl. OK, enough with the soup analogy - basically it allows companies to have a private version of del.icio.us and not worry about installing and maintaining scuttle or similar). Niall Cook noticed my attempt to sneak in and register before they officially opened up the beta, and was gracious enough to let me be the first external person to access the beta site, and I've been playing with it for a couple of days. After some problems caused by a slight mismatch in our shared understanding - read his post to be clear on what cogenz is and how it works, I think I can post some initial thoughts. Short summary - pretty good, especially for a first release. Clean and simple design, and does most of the simple bookmarky and taggy things one might expect. Everything is feed enabled, and the site throws up related users. If the site stays as it is and no more functionality was added, I think it would be of value to many companies wanting to experiment and start using social bookmarks without having to do it in public. So, now you can tag intranet pages and interesting external content without worrying about what is being revealed to the outside world. Of course, you still have to worry about who's reading it from inside your company, but that's a separate problem ! Of course, nothing is perfect, and here are some quick thoughts I jotted down (Google Notebook really is good for this kind of activity). Nothing horrific, and probably in the plan for future releases.

• Should be able to search for users
• Can't see how to tag a user (although people can tag themselves)
• Permissioning model may be too simple (although I could be trying to make it too complex)
• currently it looks binary (either part of company and therefore able to access all bookmarks, or not part of company and thus not able to read any bookmarks).
• But, how do we share a limited set of content with an external partner.
• Should the system warn if trying to share bookmarks that really are internal (ie point to an intranet that is inaccessible outside the firewall)

See my post for a link to a talk by Joshua Schachter of del.icio.us - where he talks briefly about groups and networks, and this post (in fact the whole blog) from Steve Eisner talks in more detail about some of the issues with permissioning tags and groups. Anyway - nice work guys, it will be interesting to see where this goes, and what the pricing model is ! [tags]cogenz, tags, social, bookmarks, delicious[/tags]

I'm watching John Seely Brown's plenary address at the 2006 Collaborative Technologies Conference, which is a wide ranging and fascinating talk on ways and forms of collaboration. Lots of good stuff and plenty of blog activity around this and the other talks - but in many ways echoing (in a good way) what is rapidly becoming the conventional wisdom in our brave new world...mashups, Second Life, "honouring the emergent", etc. Something that did make me think though, was his description of how an Asian apparel company (Li & Fung) has a large number of suppliers (10,000+) that it has organised into loosely coupled supply and innovation networks where knowledge, practice and process are shared between companies. Hmm...now the way I've generally thinking about collaboration, sharing, security, etc, has been fairly simplistic: there are essentially 2 situations - you're inside a firewall, or you're on the outside. Most available tools focus on the external world (technorati, del.icio.us, etc). Some companies are starting to think about replicating these tools and services within organisations (an external service probably cannot subscribe to, tag or bookmark internal content). But how do we facilitate sharing and collaboration across a number of different companies ? The partners don't have access each other's internal space, and using an external service reduces their ability to reuse internal content (unless they want to copy it to an external service, and then keep it up to date - yuck). It also maybe implies a need to agree on a common set of tools (good luck !) - and suddenly they're not very loosely coupled anymore. So, is there (or will there be) a need for something like a reverse proxy architecture that allows proxying of interesting internal content (feeds, web content, web service calls) to authorised external services ? Or am I making up solutions to problems that don't exist ? Or, even worse, am I misunderstanding something ? [tags]ctc2006, collaboration, internal, external, collaboration, proxy[/tags]

Exciting times ahead...

I played around with Mog yesterday - a music based social network that tracks what you play etc, lets you connect with people, blogs - all the usual kind of stuff. Some of the coverage:c|net, zdnet, Red Herring.

Underwhelming. As a reasonably long-term last.fm subscriber, I'm not sure what mog has that would make me want to move. Let's see:

• Track what I'm playing ? Got that.
• Blogs, yup;
• forums - covered;
• tagging, likewise;
• networks - whether explicit or emergent - also there.

Actually, mog doesn't do emergent networking (creating neighbours based on similarity) - or at least I don't think it does. I would check but I seem to have become un unmogger - the site won't let me log in, won't mail me my password as it doesn't recognise my email address, and won't let me sign up again with that email. Humph. Once feature that last.fm doesn't have is that you can upload your entire music collection, but you'll need to be patient...I left it running this morning having processed about 70 songs in an hour, and I have no idea what it's going to do with the information.

Anyway - not a chance I'll move away from last.fm to mog. Tom Coates sums it up aptly:

"It's got some attention because it was linked to from BoingBoing, but I have to be honest, at first glance it looks a lot like a crappy last.fm - a site that they completely haven't acknowledged in most of their press"

[tags]last.fm, music, collaborative, mog, audioscrobbler[/tags]

Every so often I decide that I need to skim through my keyword feeds in Bloglines (more to get my number of unread posts down below 4 figures than in the expectation of finding anything really useful). This morning, I was skimming through one of my ROI feeds, following a couple of links, much of the same stuff - yes ROI is probably important, yes we can work out how much paper is being saved, probably, no we can't get into anything more convincing.

And then, like a 40 watt bulb flickering to life I had an epiphany. Given that calculating the ROI of the squishy stuff is really really really hard, and generally it's the squishy stuff that's the most important (more engagement, better communication, accelerated innovation, more reuse of existing information, talent retention, productivity increase, etc), then why not use a tool that seems ideally geared towards arriving at a close approximation of reality without needing or being able to measure directly? I am, of course, talking about crowd wisdom.

Here's how it could work:

• Implement your hard-to-measure-the-ROI-of initiative (for example, redesign your intranet to be more usable, or implement social bookmarking, or an internal blogosphere, or whatever)
• Set up a mechanism to aggregate the wisdom of the crowd - maybe a prediction market (although aren't they usually associated with events that have a measurable outcome such as "Who will be the next President", "Who will win the World Cup, etc" ?). Maybe all that's needed is a simple survey asking people how much they thought had been saved, or how much extra revenue earned.
• Add and divide, and obtain average. I'm sure there are plenty of slightly more rigorous statistical techniques here for processing the numbers (eg throwing away outliers, etc), but essentially, that's it.
• The hard part is now trying to work out whether that number bears any actual resemblance to reality - but maybe that's easier to judge when there's a number in front of you.

I admit that this doesn't let you know the ROI up front (would that be worthwhile, asking employees to predict the ROI before implementing ? Is that too complicated a question to ask ?), but it might give you some justification of impact.

What do you think ? Stupid idea, or just impractical ?

[tags]ROI, measurement, fuzzy, squishy, metrics, aggregation, crowdwisdom, wisdomofcrowds, predictionmarkets[/tags]