My 2005 Prediction of Wikipedia’s Failure By 2010 Was Wrong

By Eric Goldman

“Prediction is very difficult, especially about the future.”

— attributed to Niels Bohr

I’ve written a few notorious posts in my 6 years of blogging, but none more so than my December 2005 prediction that Wikipedia would fail in 5 years. Those 5 years are up, and this post admits that my prediction is wrong. In this post, I’ll also look at how Wikipedia has changed since 2005. Although these changes don’t validate my prediction, Wikipedia circa 2011 differs in important ways from Wikipedia circa 2005.

Why Did Wikipedia Beat the Odds?

My initial prediction focused on the threats posed by spammers and vandals. With Wikipedia’s free editability, spammers and vandals can easily manipulate content to advance their own interests. Standing between them and success are Wikipedia volunteers. With millions of financially motivated spammers and thousands of dedicated volunteers, the volunteers seemed outgunned. How have they been able to vanquish the spammers and vandals to date?

In my opinion, the biggest change was Wikipedia’s (re)adoption of Google’s nofollow tag in 2007. Prior to that, spammers had significant incentives to insert links into Wikipedia pages for the link juice, no matter how many Wikipedia readers clicked on the links. After implementing the nofollow tag, spammers’ payoffs for getting links into Wikipedia decreased substantially. There are still other ways that marketers can manipulate Wikipedia (such as editing their own entries), but these efforts are small potatoes compared to SEO bots.

In addition, the Wikipedia community has done an admirable job developing anti-spam technological tools available to its volunteers. Anti-spam efforts remain principally a manual effort, but improved anti-spam tools has increased the leverage of the remaining volunteers.

Other Ways Wikipedia Has Changed Over 5 Years

While these factors have been pretty effective at keeping the spammers and vandals at bay (so far), other changes to Wikipedia have been less salutary.

* Xenophobia. Over the past 5 years, the Wikipedia community has closed ranks and become somewhat insular. Wikipedia offers technological tools to the public at large to edit the database, but as I explain in my Wikipedia’s Labor Squeeze article, the practical capacity to make edits that stick is limited to a much smaller universe of contributors. Joining that club is relatively easy, in the sense that it merely requires a person to “show up,” contribute and get along with other community members. At the same time, club membership is hard because it must be a concerted effort over time. Drive-by edits by uninvested contributors are unlikely to stick.

Xenophobia poses a serious challenge to the community because it makes it much harder to recruit and absorb new power contributors. Thus, community xenophobia may be one of Wikipedia’s biggest strategic challenges.

As I’ve watched Wikipedia close ranks, I’ve started to wonder if xenophobia is endemic to UGC websites. I’ve seen the phenomenon occur with other UGC sites, so perhaps it’s unavoidable. An interesting question worth a more intensive study.

* Increased Wikipedia staff. This isn’t a negative per se, but it is noteworthy. Wikipedia’s staff has grown substantially over the past few years (I believe it’s approximately doubled in the last two years or so). I don’t believe the new staff members have displaced any key editorial services traditionally performed by the community. However, the headcount growth reflects an implicit decision that Wikipedia cannot expand solely on self-coordinated volunteer efforts. With its growth, Wikipedia’s staff is increasingly looking like the staff of an editorially controlled publication.

As an example, consider the Wikipedia Public Policy Initiative, which I’m participating in. In my Wikipedia’s Labor Squeeze article, I discussed how Wikipedia could source new content, and possibly new contributors, by working with educators who incorporate Wikipedia participation into their courses. (To be clear, many other people made that suggestion too). Without Wikipedia staff coordinating the effort, some educators might independently do this themselves. However, with the Public Policy Initiative, Wikipedia is allocating FTEs to fix the coordination problem. I expect Wikipedia staff intervention will produce more and better contributions, but notice that it becomes much more expensive content to produce given the FTEs’ amortized costs.

* Increased technological barriers to participation. On its home page, Wikipedia still uses the introductory tagline “the free encyclopedia that anyone can edit.” However, the site has struggled with free editability over the years. A variety of (mostly minor) technological restrictions have been implemented over the years to limit contributions from untrusted sources.

From my perspective, the most important technological control is “Flagged Revisions” (rebranded “Pending Changes”), which allows only trusted editors to immediately make public edits. All other contributions sit in a tank until a trusted editor approves them. Flagged Revisions is a substitute to Wikipedia’s traditional full protect/semi protect approach to restricting edits on problematic pages. As a substitute, Flagged Revisions is more flexible than protection because people can still contribute to an affected page (although the contributions aren’t immediately public and may never be approved).

Instead of being used only on problematic pages, Flagged Revisions could restrict editing site-wide. It’s used for that purpose in some Wikipedia versions, but not in the English language version. When Flagged Revisions is used to widely control editing, it undercuts the tagline that Wikipedia is a place anyone can edit. That still may be technically true, but making distinctions between editor trustworthiness means that untrusted contributors may not actually be able to edit.

Thus, if the English language version of Wikipedia had broadly adopted Flagged Revisions before the end of 2010, I was going to declare my prediction as mostly correct. Unfortunately for my prediction, widespread implementation of Flagged Revisions hasn’t happened yet. However, as I initially predicted in my 2005 post:

Wikipedia will have to change its open access nature. Instead, Wikipedia will have to lock down lots of pages from being edited at all. Or Wikipedia will have to install some reputational management system to limit who has the right to post or edit content.

I may not have gotten the timing right, but I wonder if this fate remains inevitable.

* Several measures of editor engagement have peaked. There are several possible explanations for why contributions and contributors are down from historical highs, including:

– the community’s xenophobia, freezing out newcomers

– the “old guard”—the initial cohort of power Wikipedians—may be progressively turning over or burning out

– Wikipedia may have reached a plateau in terms of coverage and quality, so less work still needs to be done

Whatever the reason for these numerical declines, they pose a serious challenge to Wikipedia’s freshness. Wikipedia currently does a very good job keeping highly trafficked articles up-to-date. For example, when a celebrity has a major new development, that celebrity’s page is often updated that same day. However, I routinely find long tail pages that are poorly maintained and conspicuously out-of-date. Stagnancy could seriously undercut Wikipedia’s credibility over time; if the site looks like a ghost town, readers will become less trusting of the site, and that will enhance the attractiveness of competitors.

Looking Ahead

Over the years, it’s become clear to me that searchers want and need an encyclopedic entry in their top search results. In many circumstances, a relatively objective factual source improves the iterative search process. Wikipedia has unquestionably met that need, which is one reason why it remains so popular.

However, it’s less clear that Wikipedia’s current labor model and site architecture is the only—or even the best—way to deliver encyclopedic search results. Other labor models or site architectures could usurp Wikipedia’s current approach while delivering the same basic value proposition to searchers. It seems likely that any successful encyclopedic site will rely on crowdsourcing to ensure enough support for long-tail topics. However, free editability may be less important to the end result.

Wikipedia remains a community that can be baffling and exasperating at times, but I also routinely find the site’s content useful and interesting. I visit it daily as part of satisfying my intellectual curiosity. Happy 10th anniversary, Wikipedia!

My Prior Posts on Wikipedia

* Catching Up With Wikipedia (Feb. 2010)

* Offering Students a Graded Wiki Option—My Experiences, and Some Lessons (Feb. 2010)

* Why More Wikipedia Editing Restrictions Are Inevitable, and Some Comments on Flagged Revisions for Living People’s Biographies (Aug. 2009)

* Wikipedia and Rules Proliferation (Aug. 2009)

* Decay Rates of Committed Online Community Members–an Epinions Case Study (Jan. 2009)

* Wikipedia Revisited: the Wikipedia Community’s Xenophobia (Jan. 2008)

* Wikipedia and Search Engine Marketing (SEM) / Search Engine Optimization (SEO) (May 2007)

* Wikipedia Will Fail in Four Years (Dec. 2006)

* Wikipedia Will Fail Within 5 Years (Dec. 2005)

Also see my 2009 law review article, Wikipedia’s Labor Squeeze and its Consequences