3.5 User-Generated Content
User-generated
content has been the key to success for many of today’s
leading Web 2.0 companies, such as Amazon, eBay and Monster. The community adds value
to these sites, which, in many cases, are almost entirely built on user-generated
content. For example, eBay (an online auction
site) relies on the community to buy and sell auction items, and Monster
(a job search engine) connects job seekers with employers and recruiters.
User-generated content includes explicitly generated content
such as articles, home videos and photos. It can also include implicitly
generated content—information that is gathered from the users’
actions online. For example, every product you buy from Amazon
and every video you watch on YouTube
provides these sites with valuable information about your interests.
Companies like Amazon
have developed massive databases of anonymous user data to understand
how users interact with their site. For example, Amazon
uses your purchase history and compares it to purchases made by other
users with similar interests to make personalized recommendations (e.g.,
“customers who bought this item also bought...”). Implicitly
generated content is often considered hidden content. For example, web
links and tags are hidden content; every site you link to from your
own site or bookmark on a social bookmarking site could be considered
a vote for that site’s importance. Search engines such as Google
(which uses the PageRank algorithm) use the number and quality of these
links to a site to determine the importance of a site in search results.
Collective Intelligence
Collective
intelligence is the concept that collaboration can result
in smart ideas. Working together, users combine their knowledge for
everyone’s benefit.
The first chapter of Wikinomics,
by Don Tapscott and Anthony D. Williams, tells
the Goldcorp story. Inspired by the
community efforts in Linux, the CEO of Goldcorp released
to the public proprietary geological information about the company’s
land. Goldcorp offered cash rewards to people who could use this information
to help the company locate gold on the land. The community helped his
company find 8 million ounces of gold, catapulting Goldcorp from $100
million in stock equity to $9 billion.1 Goldcorp reaped amazing benefits by sharing information and encouraging
community participation.
User-generated content is significant to Web 2.0 companies
because of the innovative ways companies are harnessing collective intelligence.
We’ve already discussed Google’s PageRank (Section 3.3),
which is a product of collective intelligence. Amazon’s and Last.fm’s
personalized recommendations also result from collective
intelligence, as algorithms evaluate user preferences to provide you
with a better experience by helping you discover new products or music
preferred by other people with similar interests. Wesabe is a web community where members
share their decisions about money and savings—the site uses the
collective financial experiences of the community to create recommendations.2 Reputation systems (used by companies
like eBay) also use collective intelligence to build trust between buyers
and sellers by sharing user feedback with the community. Social
bookmarking sites (Section 3.10),
and social media sites (like Digg
and Flickr)
use collective intelligence to promote popular material, making it easier
for others to find.
Wikis
Wikis,
websites that allow users to edit existing content and add new information,
are prime examples of user-generated content and collective
intelligence. The most popular wiki is Wikipedia,
a community-generated encyclopedia with articles available in over 200
languages. Wikipedia
trusts its users to follow certain rules, such as not deleting accurate
information and not adding biased information, while allowing community
members to enforce the rules. The result has been a wealth of information
growing much faster than could otherwise be produced. In 2005, an experiment
comparing 42 entries from Wikipedia and Britannica (a popular
printed traditional encyclopedia) showed only slightly more inaccuracies
in the Wikipedia articles.3 The Wikipedia entries were promptly corrected, though, whereas
errors in Britannica entries cannot be corrected
until the book’s next printing and will remain in already printed
copies.
Wikipedia, Wikia (a site for specialized wiki
communities about popular television shows, games, literature, shopping
and more) and many other wikis use MediaWiki
open source software (originally developed for Wikipedia). The software
can be downloaded from MediaWiki’s website (www.mediawiki.org),
where you can also find descriptions, tutorials, suggestions and more
to help navigate the software. Wikis are also used by many companies
to provide product information, support and community resources. SocialText,
the first wiki company, provides corporate wiki services. Many companies
have found that using wikis for project collaboration reduces e-mails
and phone calls between employees, while allowing the ability to closely
track a project’s changes.4
Collaborative Filtering
Though collaboration can result in a wealth of knowledge,
some users might submit false or faulty information. For example, Wikipedia has experienced instances
of people deliberately adding false information to entries. While moderation
(monitoring of content by staff) is sometimes necessary, it is time
consuming and costly. Many Web 2.0 companies rely on the community to
help police their sites. This collaborative
filtering lets users promote valuable material and flag offensive
or inappropriate material. Users have the power to choose for themselves
what is important. Examples of sites using collaborative filtering includeDigg, a news site where users rate
the stories (see Section 3.8),
and social
bookmarking sites such as del.icio.us,
where users can easily find popular sites (see Section 3.10).
Customer reviews on Amazon
products also employ collaborative filtering—readers vote on the
usefulness of each review (helping other readers to find the best reviews).
Craigslist
Craigslist, founded by Craig Newmark, is a popular classified ads website that has
radically changed the classified advertising market. Newspapers have
experienced a decline in classified ad sales,5 as revenues from help-wanted ads on Craigslist climbed to $50
million in 2006.6 Most ad postings on Craigslist are free, and it’s easy for
anyone to post ads. The site has gained popularity because of its job
and housing postings. In 2005, a documentary, “24 Hours on Craigslist,”
showed the diverse postings that occur on the site in a single day.7 Craigslist is built on user content, leveraging the Long Tail by connecting the unique
(often unusual) needs of its users. The site also uses collaborative filtering—users
are encouraged to flag inappropriate postings.
Wisdom of Crowds
Wisdom
of crowds (from the book of the same title written
by James Surowiecki) is similar
to collective intelligence—it suggests that a large diverse group
of people (that does not necessarily include experts) can be smarter
than a small group of specialists. The key difference between collective
intelligence and the wisdom of crowds is that the latter is not meant
to be a collaborative process—part of forming a reliable crowd
is making sure people don’t influence each other.8 For example, Surowiecki describes how calculating the average
of all submissions in a guessing contest (e.g., guessing the number
of jelly beans in a jar) often results in nearly the correct answer,
even though most individual estimates are incorrect and vary considerably.
When the U.S. submarine Scorpion sank in
1968, the Navy asked various experts to work individually assessing
what might have happened; their collective answers were then analyzed
to determine the accurate location of the submarine.9 Practical everyday applications of the wisdom of crowds can be
seen in sites employing collaborative filtering.