Wednesday, March 3, 2010

Knol: The State of Play
By Roger Browne

Knol is a Google service that allows anyone to publish an article about almost anything.

When Google’s VP of Engineering, Udi Manber, announced Knol in December 2007 it was widely regarded as a “Wikipedia killer”. If Knol’s users could replicate the success of Wikipedia, the pundits proclaimed, then Google would gain an enormous commercial benefit. No longer would users leave the Google’s search results to find their answers at Wikipedia; they would go to Knol instead.

Google, meanwhile, was promoting Knol as a place for named authors to write authoritative articles to share “units of knowledge”, or knols.

Google’s announcement was accompanied by a sample Knol page, an article about insomnia by Rachel Manber (Udi Manber’s wife) who is an insomnia specialist from Stanford University. The article was comprehensive, carefully-written, well-illustrated, neatly formatted and very readable.

Knol conducted a private invitation-only test until the gates were opened to the public in July 2008. There was an immediate flurry of interest, encouraged by the potential to earn AdSense income by allowing Google to place ads on knols.

Unlike many of Google’s other ventures, where Google’s team can be quite secretive, the Knol team was remarkably open, engaging directly with the authors. They solicited improvements, delivered them, discussed problems, helped those who showed potential, and even submitted collaborative edits to many articles (including one of mine).

The first hundred thousand knols

Less than six months later, Google proudly announced that the 100,000th knol had been published. The announcement was met with cries of “So What?”. The media has a short attention span, and those who remembered the launch of Knol didn’t think much of where it hand ended up. Slate magazine hadn’t even waited that long before writing off Knol.

Matt Cutts thought it was weird that the same people who had hailed Knol as a dominant player were now the ones now dismissing it. Matt blogged that, in his opinion, Knol was doing just fine, but the comments posted to Matt’s blog were overwhelmingly negative.

Platform development

Meanwhile the Knol team slogged away, improving and promoting Knol. It seems to me that they’re taking a long-term view of this project. A steady stream of new features, improvements, and marketing activities was delivered:

• Knol debates, with opposing knols pitted head-to-head (in contrast with Wikipedia’s “neutral point of view” policy)
• articles permitted in any language, and eight languages for the Knol interface
• version control with rollback
• various awards and badges for “top viewed” and “top pick” knols
• a very comprehensive advanced search system
• display of related knols in the sidebar
• table and caption editing tools
• email notifications and activity feeds
• integration with Google Analytics
• facilities to help embed knol content into other sites
• a competition in conjunction with dummies.com with a $1000 prize
• a new kind of knol called a collection, which works like a directory
• expanded name verification options (but still for the US only)
• the ability to embed images, equations, slideshows, video, calendars, maps, documents, presentations, spreadsheets, gadgets and thousands of special characters and symbols
• audio playback of knols
• “dofollow” links for trusted contributors
• pageview statistics
• comment rating
• a promotion for knols about holiday crafts
• crowdsourced categorization of knols
• a ten-point review scorecard for each knol
• a standard citation format
• a twitter interface
• ten million professional stock photos licensed for use at Knol via PicApp
• and many, many more enhancements

20 months later

Now, almost 20 months after its launch, the service still sports the beta tag. Google hasn’t announced an updated count of knols, but a search for [site:knol.google.com] shows 163,000 results, although there may be more. Those results are a real mixed bag.

There are lots of spam pages which have no purpose other than to drive traffic to dubious commercial websites. If you’re interested in acai berries, you’ll find plenty of knols about them.

Next are the crank pages. Knol provides a soapbox for any author to pontificate about whatever they like, so it’s not surprising that you can find frauds, lies and delusions. Some of the more egregious of these, such as the “run your car on water” knols, are being flagged and removed. Other knols just reflect the obsessions of their authors. If you want to read articles like “Horniness is a spreading epidemic that must be stopped”, knol is the place to go.

There’s still much copying of content from Wikipedia, despite great efforts to stamp it out (helped by the “similar content” links generated in the sidebar of the knol).

There are thousands of pages about Knol itself. Many are full of navel-gazing and pontificating about the service, rather than real content. Around these pages have formed several “organizations”, informal affiliations of knol authors who give out awards, place their stamps of approval on articles, or conduct lobbying campaigns. It tends to be a small clique of knol authors who get involved with this, and the whole process seems to me to be somewhat self-congratulatory.

A “currency” of sorts has evolved, called “knol translation points”. The idea is that you earn these points by translating knols into other languages. You can then spend the points to have your own knols translated, or you can “award” your points to others in recognition of their good work.

There are some interesting pockets of light. A number of eHow authors have contributed over six hundred knols under the joint username eHowKnol. The quality is variable and unexceptional, but nevertheless above average.

More significant is the use of Knol by the Public Library of Science, an open access publisher, as a platform for “open communication and discussion of new scientific data, analyses, and ideas about influenza research”. Under the title “Influenza: Currents” new articles are drafted and collaboratively edited. They are then subjected to stringent vetting by team of qualified moderators, after which the successful articles are included in PubMed.

Knols can be sorted by rating or by page views. Looking at the top knols, it’s clear to me that the most-viewed knols are, on average, much higher quality than the top-rated knols.

Only ten knols have had more than 100,000 page views over the past 20 months, and some of those are knols written by Googlers back in 2008. The majority of knols have had far fewer page views, mostly below a thousand, and sometimes only several dozen. It seems that Knol is a great platform for writers, but not so much for readers.

A killer?

How has Knol fared as a Wikipedia killer? Google Search Trends provides a dramatic answer:

(I couldn’t compare the sites using Google Website Trends, because that service does not accept queries for google.com and its subdomains.)

The Knol team no longer seems to be using the phrase “authoritative content”, and I can’t help thinking of something else Matt Cutts wrote after the 100,000th knol was posted:

My personal conception of Knol is that when you want to write a quick article or put some information on the web, Knol is a great place to do it.

Is that it then? Is Knol destined to be nothing more than the next Geocities?

Knol: The State of Play by Roger Browne | Comments (13)

>> More posts

Advertisement