Knol: Good Attempt at Herding Knowledge Cats
After Peter Drucker told us in 1966 that we were becoming a knowledge economy, it was inevitable that big companies would spend lots of money on complex knowledge management systems. Most of those investment had very poor returns, because they were based on old command and control styles of management and that is not how knowledge workers operate; since the Internet gave us the power, we are all cats.
Modern knowledge management is all about herding cats. Ever tried telling a cat what to do? Even “kitty, kitty, kitty” calls are pretty ineffective. A bowl of milk is better. Google’s recently released Knol service shows that they understand this.
Knowledge cats typically want three things:
- Money.
- Attention (which may lead to more money or just goodwill).
- Feel good about your contribution to the world’s knowledge.
It may seem a tad mercenary to put them in that order, but that simply reflects Maslow’s Hierarchy of Needs.
Look at how the existing alternatives to Knol score on these three criteria:
- Wikipedia. Useless on 1 and 2, great on 3. Wikipedia has been a phenomenal success story and most commentators put Knols as a Wikipedia competitor. I appreciate the irony of using Wikipedia for almost all the background links for this post, including the one on Knol; that was not deliberate, they were just the best sources that appeared from a Google search. Wikipedia will continue to thrive in genuine long tail stuff, where there really is not any economic value but just the passion of a few people to set the record straight.
- Wikia. This is on much more dangerous ground. Contributing your time to Wikipedia for the good of the world is OK as nobody is raking in the $$$ from your efforts. The idea of seeing the Wikia founders and VCs on front cover of Fortune all based on your efforts…. My prediction is that Wikia will be absorbed into Amazon; they are already investors and have the most to lose from a Google monoculture.
- Mahalo. They claim to have enough capital to last over 4 years without making a profit, so it will be a long time before we see how effective this is. They do pay contributors (either salary or per contribution) so they score OK on 1 and 3. They show who created the content and who contributed; but this is bottom of the page and not prominent. The attention branding is for Mahalo and not for the author. I am a skeptic on Mahalo. They are going for the fat part of the tail. That is the province of specialist publishers (who can figure out how to get some user generated content).
These different attempts all matter, because they are one of the best routes to the more structured Web that is sometimes labeled Web 3.0 or Semantic Web.
Just for fun, here is my definition of Web 3.0:
“The combination of Web 2.0 mass collaboration with structured databases”.
Or as Jeremy Liew of Lightspeed Ventures calls it, “meaning = data + structure“.
That is simple to say - and incredibly hard to pull off technically. There are two things that are relatively easy:
1. Getting lots of people to contribute unstructured information for free - comments, forums, blogs, etc.
2. Getting lots of people to input data into forms where the structure is already defined in a relational database.
Two things are incredibly hard:
1. Getting computers to understand semantics/structure in unstructured text. This has been the aim of natural language search for a very long time. Lots of very smart firms such as Hakia and Powerset are investing a lot of money to get there, but on the evolutionary scale they look like the earliest life forms in the primordial soup compared to humans.
2. Getting a lot of humans without any technical training in SQL or data modelling to create structure “on the fly”, to let structure “emerge” from lots of input through some form of “collaborative data modelling”. It looks like Freebase is attempting something along these lines, although since their March announcement it has been very quiet so that is hard to say. Freebase looks like a really long play, potentially huge and game changing but not any time soon. Because this stuff is hard.
Google Knol looks like an intelligent way to motivate lots of people to add the structure that is needed.
Knol functionality is not hard to replicate. The barrier to entry for a competitor is low. The big question is whether Google will tweak their search algorithms to favor a Knol entry over any other content. I don’t believe they will do this. Firstly because they are riding high and don’t need to resort to such grubby tactics and secondly because that would trigger some serious Anti-Trust action.
The most likely source of a Knol competitor, IMHO, would be a plug-in for Wordpress MU. But that, as the saying goes, is another story.
Cat image: phantom kitty
Share
Comments
Subscribe to comments for this post OR Subscribe to comments for all ReadWriteWeb posts
I think it is extremely naiv not to see that the real motivation for Google to build Knol to expand its advertising empire to a new undiscovered land. The biggest problem for Google with Wikipedia that they are not displaying ads, in spite of their huge number of visitors. Google just cant leave it that way...
Bernard said...
Firstly because they are riding high and don’t need to resort to such grubby tactics and secondly because that would trigger some serious Anti-Trust action.
Anti-Trust law is stupid , because it is anti-property rights. Google (or Microsoft) is entitle by its own property rights (PageRank algorithm in its search engine) to do whatever it sees fit for running its business. The state or anyone else has no right to other people's properties (Google). If you don't like something (a service a or product), the solution is easy, just don't use (buy) it. The following is a good article of how stupid the anti-trust law is:
Drop the Antitrust Case Against Microsoft
I recalled that about 3 years ago, Google de-ranked a US company from its search engine, since they were doing web-farming (trying to deceive the PageRank algorithm in order to be ranked high in Google searches by repeating certain keywords in white texts from web pages). The case was thrown out. This company has no right to Google's ranking system in its search engine, so the judge was 1000% correct in reaching that decision.
But you've got sites like http://www.eioba.com where is "2" and "3"... and:
* you join "already build" community
* they ALREADY have huge database
I think there is a bit of a disconnect between your opening premise and the guts of your argument. Initially, it sounds as if you think Knol is a potential remedy to the complex knowledge management system problem that many companies face. I would argue, however, that your "knowledge cats criteria" is not necessarily relevant for the knowledge management problem within companies.
I do believe that Knol could potentially help with the knowledge management problem because of its emphasis on the author. This could enable people-centric, bottoms-up knowledge management, unlike the traditional command and control style you mentioned.
For more of my thoughts, you can refer to a blog entry I made on the day that Knol was announced -- http://abovethenoise.blogspot.com/2007/12/googles-knol-initiative.html
Bernard, you're spot on about “The combination of Web 2.0 mass collaboration with structured databases”.
Would love to chat about this sometime.
I think you're right on the 3.0 definition Bernard, clear but incredibly hard without a larger skill set on the part of the masses. Which leads me to believe that Knol is a good intermediate step in the right direction, whatever their motivation it pushes the evolution and innovation along. If Google could build the infrastructure for the average user to contribute to the web in a new way, dare I say it, like Myspace opened up html to the average user, that could be extremely helpful.
I think Knol could be very useful, the test will be how they manage to sort and filter competitive knowledge items. Wikipedia is great as each article goes through an open source type editorial process. With Knol it appears each author has total control over their content, therefore expect to see *a lot* of low quality content.
In the end it will be the sort and filter functionality that will be the test, but you would have thought Google might be pretty good at that...
Bernard, you say "Getting a lot of humans without any technical training in SQL or data modelling to create structure “on the fly”, to let structure “emerge” from lots of input through some form of “collaborative data modelling”." is very difficult.
I agree, and would add that there's a huge gap between the folks who can used SQL or other databases to organize their thoughts and knowledge (i.e. geeks) and those who have no knowledge or fluency in this work (i.e. civilians).
I also think the rise of services like Freebase and (somewhat) Google Knols are trying to bridge the gap.
We recently launched an app in September that approaches this from a another, user-friendly angle. It's called Listphile. It allows people to organize and collaborate with lists, atlases, and databases. Listphile provides non-technical controls over contributors, edits, permissions, and most significantly, a given lists's data model. My hope is that Listphile turns into a vast and yes, messy, compendium of human knowledge, collaboration and sharing.
We don't try to arbitrate the semantic structure that people infuse into the lists/atlases/dbs that they create -- we just want to give them simple and powerful tools to structure the information they want to store in our app.
Glad to finally see someone identifying these two critical differences between Knol and Wikipedia- authorship and revenue share.
I think the last points regarding scale are critical. This may be easy to copy but no copy is going to have the market footprint Google has, not even close. It's so huge that you simply can't build your way to it (sorry Mahalo). Execution is where they have to get it right, which has not been great in other Goog products.
I'd also like to see some kind of transparency policy so we don't have the endless SEO gaming that search is dealing with- tell us what goes and what doesn't, in detail. People will game this because of the reputation and revenue aspects- both are valued based on traffic.
Thank you Bernard Lunn for this post. It's good. My take is as follows. Early rounds of knowledge management technology fell short because there was no "technology of knowledge." The 90s approaches to KM gave us information technologies masquerading as knowledge technologies, combined with voodoo about implicit knowledge and knowledge becoming wisdom.
Yes, knowledge management is always a techno-socio collaboration, but what is different today is that we actually have knowledge technologies -- that is, knowledge representation that is interpretable both by humans and computers, and is sharable both by people and their computers. Standards for knowledge representation (and reasoning) help, but what is very different from before is that we are moving towards system architectures that learn. How do they learn? By integrating input from people into what the system knows, and by observation and reflection on experience. Today we see a confluence of machine learning, deep linguistics, and symbolic reasoning based on knowledge models (aka ontologies). These approaches are supplemented by human contributions and curation.
"Today we see a confluence of machine learning, deep linguistics, and symbolic reasoning based on knowledge models (aka ontologies). These approaches are supplemented by human contributions and curation.
"Posted by: Mills Davis | January 3, 2008 8:53 PM"
Could someone please translate the foregoing into English? Thanks in advance.