What to Do When an Online Community Starts to Fail

HBR.org 2013-10-31

Success breeds success online. Indeed, if there is one maxim that dominates the fate of digital communities it is that of network effects — the more users participating in a community, the more valuable it will be to new users. As networks like Twitter and Facebook scale, their advantage over nascent platforms becomes seemingly insurmountable. And yet scale is no guarantee of sustainability, as two recent examples illustrate.

In one case, Yelp faces a class action suit by users in California alleging that they deserve to be paid for their contributions, along with allegations by researchers that up to 20 percent of its reviews are fake; in another, MIT Technology Review reports on Wikipedia’s shrinking contributor base and the challenges it faces in attracting new and more diverse editors. These challenges are unlikely to be fatal, but they illustrate the fact that the issues facing mature online communities are both significant and distinct from those faced earlier in their life cycles.

Rewarding Top Users is Important, But Not Enough

Believe it or not, online communities have been the subject of study for at least 20 years at this point, much of which has focused on starting and scaling them. In a 2009 paper examining the life cycle of online communities, researchers Alicia Iriberri and Gondy Leroy of Claremont Graduate University summarize some best practices for communities in the “mature” phase:

Researchers advise that creators and managers facilitate the formation of subgroups, delegate control to volunteer subgroup managers, organize online events, and reward and acknowledge members’ participation and contributions.

This is good advice, but it won’t help Yelp or Wikipedia, both of which helped cement this as best practice in the first place. As the HBS case study on Yelp details, the company’s early success was based in part on the creation of the “Elite Squad,” a user group in the “low thousands” whose active participation was rewarded by invitation to parties in their cities. Similarly, Wikipedia hosts meetups around the world, along with an annual conference.

If catering to existing contributors can’t stem contributor unrest and abandonment, what can? Newer research suggests the key may be knowing when to let certain members go, in order to gain new ones.

Some Community Turnover Is a Good Thing

Contrary to the traditional emphasis on member retention, a 2012 paper by researchers at Boston College suggests that a certain amount of turnover within an online community is a sign of health. They write:

Some turnover may be necessary to allow new members to join. Online communities are not technically limited to a finite size, but people tend not to join once the membership or communication levels are perceived to be too high. Groups that are isolated from outside perspectives can develop biases and insular thinking that leave them susceptible to overconfidence about the group’s ability to collaborate effectively. Thus, some turnover might be necessary to create an influx of unique contributors with new ideas, skills, and information.

The authors test this hypothesis by looking at the likelihood that a Wikipedia entry gains “Featured” status, which is designed to promote the best entries. To approximate community turnover, they look at the average experience of editors who contributed to each entry. They find that there is indeed a sweet spot of experience which beats out entries by both less and more experienced teams. They interpret this as evidence that some degree of turnover translates into better content.

For Yelp, the implication is clear: try to keep your best users happy, but as for the ones so upset that they want to sue you, probably better to let them go. For Wikipedia, the lesson is that editors leaving the community may be an opportunity to bring on new ones, provided they approach it in the right way.

Existing Users Must Help Engage New Ones

The Boston College researchers’ advice for community managers does not bode well for Wikipedia. Managers “should encourage the community to remain open to outsiders,” they write. By contrast, here is Tech Review’s diagnosis of Wikipedia at present:

Since 2007…the likelihood of a new participant’s edit being immediately deleted has steadily climbed. Over the same period, the proportion of those deletions made by automated tools rather than humans grew. Unsurprisingly, the data also indicate that well-intentioned newcomers are far less likely to still be editing Wikipedia two months after their first try.

Compare that to this description of the culture at the community photo site iStockPhoto, as described in a case study by the University of Western Ontario:

The iStockphoto approach to building its collection was unique. Every image submitted was scrutinized by an iStockphoto inspector before being posted to the member-accessible portion of the site. The inspectors, respected members of the iStockphoto contributor community, reviewed the images to ensure they met the company’s standards of content and quality. They approved photos for posting, provided advice on improving images that weren’t accepted and generally mentored contributors new to the stock photo business.

(Emphasis mine.)

Online communities tend to develop a sophisticated set of norms that can serve to discourage new members. (Think hashtags and @-replies on Twitter.) The challenge for community managers is to tie authority within a community to the mentoring of new members, whether formally, through changes in the way the platform calculates authority, or informally, by encouraging established members to lead through example in engaging new ones.

If they take this advice, both Wikipedia and Yelp have their work cut out for them — Wikipedia because of its bureaucratic culture and Yelp because its anti-spam policies tend to marginalize new accounts. But over time both will need to better to balance policies aimed at retention with those aimed at recruitment and mentoring. As in the offline world, communities online can’t underestimate the value of a fresh face.