Friday, September 29, 2006

When the Laws of Networks Break Down

OK, this is my last set of notes about Dion Hinchcliffe's presentation to the Winnipeg chapter of CIPS.

After an interesting presentation, I had the chance to listen to Dion speak with some members one-on-one. After standing by listening for a while, I got the chance to ask him the question that had been bothering me: The obvious corollary to the power of networks is that small networks are not very powerful at all!

I work in a small office, and often the circle of people impacted directly by a project or initiative is 5 or so. Are there any network effects to be had here? When does "Social Computing" devolve back into just having a morning cup of coffee with the team to all stay on the same page?

To be honest, I thought Dion's answer skirted the issue a little bit. His first response was to look for ways to expand the community involved. That's kind of negating the premise of the question, but it is a challenge to think about things differently. I'm trying to look in my work for ways to widen the circle and bring more brains to bear on the problems and opportunities we see.

However, another part of his answer was important. In small teams, the lessons of social computing to make everything permanent and searchable and discoverable are absolutely critical. When one member quitting can erase 20% of the knowledge in the system, having everything based on the 9am bull-session over a cup of coffee is a very expensive mistake to make.

I mentioned that in my work we were experimenting with MS Sharepoint Services. Dion said that it was an excellent tool, and I'm inclined to agree. It makes setting up blogs and quizzes and wikis and document repositories quick and easy. I'm looking forward to setting up more team sites, project sites and meeting sites in it to see how much it can help the smooth and efficient operation of our small department.

SO, I think that's all I'm going to be writing about September's CIPS event. I'm looking forward to a light-hearted and interesting topic for October's lunch: It's Possible to Have a Career in IT and Still Have a Life.

Monday, September 25, 2006

Lessons for Enterprises from Social Computing

Before I get into my post, I want to mention Dion Hinchcliffe's ZDNet post from today, "Can Web 2.0 be adapted to the enterprise?". In it, he remarks on alot of the issues I mention here with a lot more eloquence and insight than I ever could. It seems that his whistle stop in Winnipeg is part of a larger story arc in his thinking this days.

For my second set of observations from Dion Hinchcliffe's September presentation to the Winnipeg CIPS members, I'll focus on the more interesting second half of his presentation.

After we got past his evangelizing about Web 2.0, Mr Hinchcliffe got into real lessons that companies can learn from the social computing movement of recent years.

The key to this goes back to the two laws I mentioned in the last post, namely:

  1. A large group of people is collectively smarter than the most intelligent single member, and

  2. Networks (and groups) become more valuable with more members.



The net effect of these two laws in combination is that large groups are better resources than individuals, and extremely large groups are orders-of-magnitude-better than individuals. Companies can leverage this knowledge by not relying on the expertise and efforts of a few people on the payroll inside the department.

So, the lesson for companies are to look for large groups they can harness. In web 2.0 terms, this generally means building a platform/service/tool that lots of people want to use, give it to them for free, and then harvest information from their behaviour while using the tool. You can then sell this information, use it in-house to improve products/processes, or flat out sell the users to advertisers.

In terms of specific recommendations, Mr. Hinchcliffe had some important notes for enterprises:

  • We need Google-class search for our companies now.

  • That means every important piece of knowledge/documentation in the company has to be accessible by a URI.

  • We should be discouraging the use of Office applications. Big MS Word document stores are the enemy of transparent knowledge in a company.

  • Social computing platforms like Wikis and Blogs are simple to set up, easy to administer, and solidify a company's knowledge into permanent, accessible, DISCOVERABLE formats.



So, for an example from the software development perspective, consider a word processing company. They are going to need the interface for this word processor designed and tested to make their tool easy, powerful, and intuitive for users to actually use.

The older model would have the company write up a job description, post it online and in newspapers, interview candidates, and eventually hire a User Interface Engineer. That person would sit as their desk and design the interface. Of course they are going to want to verify their design choices and they'll probably pay a group of potential users to come in and use the tool for a day. The UI Engineer will carefully observe and measure their use, and make adjustments as he sees fit.

That's a lot of overhead to get user input.

So, social computing says that all those users have the knowledge about the best UI in their head already. Let's just go directly to the source and get them to tell us what we want. We set up a web front end to our software, and record where the users click, which menus they open, and most importantly where they don't click or never go. We can accumulate a huge amount of data without paying for a dental plan for a UI engineer.

My first post about Mr. Hinchcliffe had a bit of tongue-in-cheek poking fun at the "Big-City" Hinchcliffe. But especially in this part of his talk, he showed himself to be a thoughtful, interesting man who had some real things to say about Social Computing in the enterprise. I'll be watching his writings more closely from now on.