Seradigm: Knowledge Management

It was late 1998. A colleague waved me over to his desk. “Look at this” he said, “it’s a new search engine”. He was pointing to a page that was mostly white background, an ugly multicoloured logo with a stupid sounding name, and a single text box. I was sceptical. This was totally different to the information dense ‘portal’ style pages currently in use by Alta Vista, Excite, Infoseek, Yahoo and others. “That’s ugly” I said. “Yes, but it delivers really good search results” he replied. And that was that. I started using Google. It was a better mousetrap. Not a fancier, frillier, more trendy mousetrap, just a better one. This is how disruptive technology happens. Sure, marketing is important, but only if its promoting a product that is actually better.

Someone just recently introduced me to Cuil. It’s a new search engine, founded by one of the architects of Google’s large search index, and a computer scientist from Stanford. Their staff are well qualified, and they’ve got good investor backing. Could they be the new Google?

I don’t think so. Here’s the reason. It’s got nothing to do with Google’s $4B annual profit, or the more than $1B they invest into R&D every year. It’s a question of philosophy. Not the “treat your employees really well” philosophy, but the “how to make search work well” philosophy.

Here’s Cuil’s claim about their point of difference:

“Rather than rely on superficial popularity metrics, Cuil searches for and ranks pages based on their content and relevance. When we find a page with your keywords, we stay on that page and analyze the rest of its content, its concepts, their inter-relationships and the page’s coherency.”

What they’re saying, if I understand correctly, is that semantic analysis is more powerful than distributed cognition. One of the key things that powered Google’s success was that they based part of their relevance testing on the number of links that pointed to a particular page. They analysed the Web, as well as the content in each page. They figured, if lots of people link to this page, it’s probably useful.

Don’t get me wrong, I love semantic analysis. I’m an etymology geek. I get fascinated by morphemes, taxonomy, and topic maps. It’s just that I agree with James Surowiecki when he argues that (to paraphrase the Wikipedia article on ‘The Wisdom of Crowds’) “Market judgment, can be much faster, more reliable, and less subject to political forces than the deliberations of experts, or expert committees”. Distributed cognition just seems to scale better than semantic analysis. The problem with semantic analysis is that someone has to write the algorhythms that do the analysis. That person has to make judgments about how meaning is constructed in language, and in so doing they arbitrarily close off parts of the probability space. To me, within an information domain as diverse as the Web, there’s just too much variation between signifier, signified, and referent, as used by the billion or so authors and searchers, to enable semantic analysis to work as an exclusive technique.

I’ve heard Dave Snowden say many times¹ that semantic analysis has its place, but like Newtonian physics, only within boundaries. I wonder that Google’s success is because they do this, harnessing the power of both approaches.

I’ve also tried Cuil on a number of different searches, and it’s just not even close to being as good as Google at finding what I was after. Regardless of which theory of language/cognition resonates with you, in the end, it’s results that count.

[1] See this article and various podcasts on the Cognitive Edge site.

Yesterday I gave a talk on web2.0 and social networking systems to a group of scientists at a Crown Research Institute. There were about forty people in the room, and another six or so videoconferencing in from other sites. I asked for a show of hands on questions like “who’s heard of web2.0″ (about half), “who thinks they could attempt an explanation of what web2.0 is” (one person), “who reads blogs” (two thirds), “who has a blog” (none), “who has used a wiki” (five or so), and who has an account on Facebook (none). During my talk they asked a lot of very sensible questions about privacy, digital identity, using Web 2.0 and SNS tools in the enterprise, and their utility in distributed research collaborations.

Today I got a request from a colleague to join a new SNS/feed aggregator called FriendFeed. I almost screamed. Another web2.0 technology to learn, another user account to create, another set of social relationships to map, another thing to keep up to date? That’s the last thing I need!!!!! Given that I really trust the opinion of the person who recommended it, and that I have a professional interest in this area I went ahead anyway and created an account.

It made me think that there should be a word for being overwhelmed by all of the rapidly emerging new ways to collaborate and keep in touch with people. ‘Feedtigue’ seems like an appropriate term to me. It also made me think that if I’m feeling this, and I’m a passionate early adopter of such things, what must it be like for the scientists, and other non-IT people?

We’re in a space where the technology is developing so fast, it’s enabling a myriad of changes to the way we interact with people, the size of our social networks, the frequency of our communications. We can’t predict what will work, and what won’t. To me, the web2.0 boom is an evolutionary process. Many things will be tried and will fail. Some things will work and will stick. A lot of the “try everything and keep what works” has to be done by the early adopters, so the majority don’t have to expend the effort, and can wait until the really useful things stablise.

Once I had a look at FriendFeed I was quite impressed. Its main purpose seems to be to aggregate feeds from blogs, Flickr, del.icio.us, twitter, and to distribute them to your social network. It’s not so much another thing to keep up to date, but a way of gathering the existing things together to reduce the effort and friction. It’s a bit like Sxipper (a tool to manage identity and logins across many web sites), and Netvibes (a personal portal) in that it’s infrastructure that helps people glue a whole range of web.20 services together and make them easier to use. I’m hoping we’ll see more of this sort of thing in the future.

We had the Cognitive Edge Accreditation workshop in Wellington this week. It’s the first time we’ve had this in NZ since 2004, so it was great to have Dave Snowden back. Viv Read from Australia co-facilitated. It’s fantastic to see how much the methods have evolved in the last couple of years.

The following are my notes (just rough snippets I captured) on some of the ideas, concepts and examples from the workshop. If you’re familiar with Dave Snowden’s work, some of the below might resonate. If you’re not, have a look at the web site for some intro articles and podcasts.

Ideas

Blogs are a huge source of narrative material.

Preconditions for innovation are starvation, pressure and perspective shift (e.g. people will die in Apollo 13). Creativity is a symptom of innovation, not an input

The way that people recall things they know are not the same as the way they actually know in the context in which they know it. Human application of knowledge in the field is fragmented, not structured.

Communities of practices work within the first year (due to novelty), and then they’ll fail. So set them up with the expectation that they’ll be taken down in 18 months time. If at that time they want to keep the community going, let them do so, but they have to pay for it.

See-Attend-Act are separate tasks, rather than a natural response. People see more than they pay attention to, and they attend to more things than they act on.

Efficiency vs effectiveness (not and) – efficiency is stripped down to be optimised for the current circumstances, in a sense it’s in tension with effectiveness (which requires some redundancy/slack in the system to be adaptive to changing circumstances).

Language and the brain have co-evolved together. Languages have evolved be able to be learned by people who haven’t yet learned how to learn.

The cynics are the people that actually care about the organisation. Choose people 15 years or closer to retirement who’s careers have been impaired by undue cynicism, people who have been there for less than 5 years who are going the same way.

Creating environments where people can see the way other people see them.
If you ask for individual information items and you have a clear need people will give them to you (Just in time KM). If you ask people to give you everything just in case it’s needed in the future, they won’t (Just in case KM).

Quotes

“Religion is part of the sense making capability of the human condition.”

Terry Eagleton – Holy Terror – if you believe terrorists are evil or mad, you’ve lost. If you can’t see through their eyes you can never defeat them.

“Stories are a parasitical life form, warping lives in the service only of the story itself” – Terry Pratchett. Dominant myth structures will influence people’s behaviour.

Examples

Leadership development – IBM, 300 people identified as the future leader’s group. Each year 50 people will drop off the group.

Italian sales people – no longer had to write sales reports. Instead record a five minute audio 5 minutes before, and 5 minutes after a sales meeting.

Methods

Create a body of paradoxical stories where there isn’t an obvious right answer. Then get particular groups to index the stories (e.g. theology students and MBA students – to get the spectrum of good and evil . Then get other groups to index the stories to see where they as a group fits in that range.

Use anecdote circles where people have a common experience and need to communicate that experience. You won’t get objectivity in an anecdote circle in the same way as with individual capture. It’s important to ritualise capture so you get coherent anecdotes. Remember though that groups norm very quickly, so break the groups up after half an hour.

Population sampling – e.g. 5 demographic bands based on length of service. Chose 50 people at random from each group. They got an email from the CEO telling them they’d be getting a digital tape recorder, and asked them to find one person who represented the values of the future for the organisation, and one person who represented the value of the past, and ask them these 5 questions, and get them to index their responses.

Naturalistic vs idealistic approaches. Idealistic – you can train facilitators to be non biased. Naturalistic – humans are inherently biased so the process has been designed to minimise the impact of this.

Participative observation – hiring students to gather stories by working with the workers in an apprenticeship style model, working shifts with the most junior apprentice.

Naive interviewing – use people who the interviewees consider naive, e.g. 1st year engineering students for oil rig engineers (building the abstraction level into the method). Or using school children to interview old people.

Resilience vs control is a trade off in probe design (experiments) in the complex domain

This started as a comment on Dave Snowen’s Sassy Red blog post but it got a bit long and I wanted to expand on it a bit more (and save it in case Dave legitimately rejects the comment because of its length).
To put this in context Earl Mardle while mostly agreeing Dave’s post on IT education in school took issue with the statement that: In any complex system you can never replicate outcome, but you can replicate starting conditions.

It got me wondering whether there’s a useful distinction here between ‘starting context’, and ‘starting conditions’. It’s of course nonsensical to try to replicate the starting context in any learning situation (or any other situation for that matter). There will be different people, a different physical environment, different weather, there might have been national media announcements since the last event that have affected people’s mood (or in the case of a school announcements by the principal).

Starting conditions you can replicate because they’re much simpler. I’m fairly sure that if I put a soccer ball on the ground near a group of 6 year old boys that some kind of ‘soccer-like’ behavior has a chance of emerging. It can be a different group of boys from last time I did it, different playing surface, different weather. I can also fairly easily set boundaries by saying things like “remember, no punching”. If some ‘soccer-like’ behavior does emerge, I can wait to see if other interventions to stablise a ‘soccer-like’ pattern might be necessary. I might encourage two of the kids to take of their jerseys and make goal posts. There might end up being two sets of goals or just one. I might decide to say “remember, no hands”, or I might wait to see if some form of rugby/soccer hybrid (or some other game entirely) emerges.

Now to Earl’s objection to “In any complex system you can never replicate outcome, but you can replicate starting conditions”. I think he’s saying that you can’t replicate the ‘starting context’ for the teacher, because the teacher has already experienced the original outcome and process. The teacher is changed. They are no longer in exploratory/experimenter mode, they are well on the path to process entrainment and ritualisation (an important survival tactic for overworked teachers).

For me it comes down to outcomes again. If the desired outcome is ‘an exact replica of what happened last time’ then the change in the teacher’s starting context might lead them to impose structure to force the desired outcome. You might then end up with similar outwardly visible behavior as last time, but nothing like similar learning experiences taking place. I saw this often in my time at school, during my teacher training, and in the very brief time I actually spent teaching high school.

A good teacher (or parent, facilitator, manager) should however be able to let go of the exact, outwardly visible outcome, and focus on allowing the desired type of inward learning to take place. If the desired outcome was something like ‘the kids will learn some more physical coordination, how to interact with each other in a game, and how to negotiate and stablise adaptive boundaries themselves’ then as a teacher the fact that I’ve done something similar before shouldn’t be a hindrance. It should make it more likely that if I can identify what the important starting conditions were (as distinct from the irrelevant contextual elements), then I should be able to have something vaguely within the range of the learning experiences I’m after, occur.

It’s like good birthday party management. I’ve run 16 kids birthday parties and I’ve gotten fairly good at it. Not because I have a particular set of rituals or overt processes I follow, but because I know what to look for, when to intervene, and what kind of catalysts and boundaries to prepare in advance. The starting context is always different, but useful starting conditions can fairly easily be replicated.

I think we often have the same problem in organisational management, in particular poorly implement ‘management to outcomes’ approaches. In socially complex situations (which most workplaces are), managers try to achieve particular outcomes by replicating irrelevant aspects of starting context, or by having preconceived notions of how the process should work because that’s how they saw it unfold last time. They end up forcing behavior through extrinsic motivators or bullying, imposing arbitrary process, getting overt compliance with the process but tacit rejection of it, and in so doing miss enabling the creativity and myriad of interactions that led to successful results emerging in the past.