What Community Builders Can Learn From Research

Posted by thogenhaven

Two weeks ago, Tom Critchlow suggested that we work to close the gap between inbound marketing and content marketing communities. It's time to build bridges again, this time between inbound marketing and research. In this post, you'll find research on participation patterns, how to spot high-value users, seeding content in a new community, how to bring new life to old content, and a little bit of gamification.

Some research is already being shared with the inbound community. Bill Slawski from SEO By The Sea does a great job reading and condensing patents from the search industry. But there is so much more research waiting to be tapped.

I am currently in a PhD program and therefore attend academic conferences. They are different to MozCon, SearchLove, SMX, Blueglass and the other conferences we all usually go to. And different means different perspectives. Last week at CSCW, 160 researchers from private companies and universities presented a paper. Topics include social media analysis, collaboration, gamification, incentives, recommender algorithms and online communities. For better or worse, I did not attend 160 presentations. So this will be a very limited summary, focusing on online communities.

Why Should You Care?

Universities and private companies like IBM, Microsoft and Google do some legit research. Being familiar with this research is a competitive advantage and will help generate new ideas.

In this post I focus primarily on community building. At SearchLove last year, Rand had a slide stating a 34% growth in 4 months, primarily from Q+A, YouMoz, the blog and user profiles. Add to this that community members are some of the best link builders you'll ever find. Getting community right is a huge win.


Who Participates In Online Communities?

Previous research offers two perspectives on participation patterns in online communities:

  1. Some people contribute, and others do not. It is an inherent, personal trait like hair color.
  2. Lurking is a development stage toward being an active member. All people potentially contribute, after the learning/socialization phase: users lurk for a while before participating.

Michael Muller from IBM presented fascinating research on a study on 8,711 online communities covering diverse topics with 224,232 unique users. The insight of the research shows a completely different pattern than the conventional wisdom above: 84 % of those users who participate in one or more community, lurk in others. However, the majority of members' lifetime contributions are in the beginning on their membership. Thus, many users start off contributing like mad, then stop. This means retention is key.

(Graph is printed in Muller, 2012. See references in the bottom of this post).

Design implications: Do whatever you can to grasp new members. There are many ways to do this: Make sure they get encouraging feedback to their initial comments/contributions. Assign them a mentor. Send them nice emails. Reach out to them on social media.

Spotting Talent

Despite the overall participation trend identified by Michael Muller, some people are more likely to contribute more to new communities than others. In fact, only few people end up participating in the first place. Google+ VP Bradley Horowitz once wrote about 90-9-1 principle, describing how 1% of community members are creators, 9% are synthesizers, and the remaining 90% are users/lurkers who do not directly add anything to the community.

Rosta Farzan and colleague from Carnegie Mellon University and University of Minnesota developed an algorithm to identify potential high-contributing members. The algorithm uses the following metrics to spot a potential high value member.

  • Quality
  • Motivation (quantity, frequency, and commitment)
  • Ability (knowledge, trustworthiness, and politeness and clarits)

Those identified as potential high-contributing members participated 10 times more actively than those not classified.

Design implications: sometimes the gold is right in front of us, but without our knowing. Identifying high potential members early on can help us reach out and retain these creators.

Starting A New Community

In inbound marketing, one often hears the advice: go build a community. Yes, we'd all love to have flourishing communities, right? But how to get critical mass? One solution often used is seeding a site with (third party) content. This is supposed to show that the community is lively and thereby encourage users to contribute. Jacob Solomon and Rick Wash from Michigan State University tried this form of bootstrapping when starting a new wiki.

The results show that users contribute more when they are given a blank page, than they do when they see a seeded page. This makes sense, as there is more work to do on a blank page. However, contributions made on a blank page tend to be unstructured. If the users see a page with some content (e.g. headers, text chunks, objective content, opinionated content etc.), they tend to contribute content similar to the seeded content.

Design implication: If you want users to create a special kind of focused content (e.g. replies of a certain length or with a special focus), seeding can be good. The bad news: seeding content is not a shortcut to start a community as it might actually reduce contributions. Two weeks ago, Rand and Dharmesh launched Inbound. When the site was launched, it was already seeded with many good articles. According to this paper, this seeding reduced contributions, but made them more focused on the kind of articles Rand and Dharmesh want. Sounds plausible.

New Life To Old Content

This one might require a bit engineering power. But it is really neat. Aditya Pal and colleagues from University of Minnesota created an algorithm to detect expired content on a Q&A site. The algorithm uses metrics such as

  • TF/IDF
  • Reference to a specific time (e.g. date, month)
  • Fixed vs relative time reference (ago, after, before, today, tomorrow)
  • Reference a date in past
  • Tense of the question

Design implications: Such algorithms are not only useful on Q&A sites. On enterprise websites, it can be used to flag content that ought to be updated, removed, rel=canonicalized or 301 redirected to new content. This creates better and fresher content on websites, as well as help avoiding old and irrelevant pages rank in Google. It can also help scale some of Cyrus Shepard's advices on fresh content, and help you rank for QDF keywords.

(This illustration is made by Dawn Shepard for Cyrus' post mentioned above)

Gamification Over?

Gamification has been a hot topic in the last couple of years. For many websites, the question is no longer if gamification systems should be implemented, but if it should be kept. Jennifer Thom and collaborators from IBM studied the removal of gamification points from IBM's internal social network. The researchers found that removing the points system made users contribute significantly less than before.

Design implications: You might (also) be tired of hearing about gamification. But it kinda works... So you might want to take a look at these gamification slides from Richard Baxter:

Curious for more?

The ACM Library is very good. In fact, so good that Matt Cutts blogs about it. To access the articles, you might have to go to a library or a university. But many researchers are happy to share their research, and link to it directly to their own work from their personal websites (The authors have the rights to share their own articles for free). So a little Googling can usually provide the article.


Michael Muller (2012): Lurking as Personal Trait or Situational Disposition? Lurking and Contributing in Enterprise Social Media. Proceeding to CSCW 2012

Aditya Pal, James Margatan, Joseph Konstan (2012): Question Temporality: Identification and Uses. Proceeding to CSCW 2012

Jacob Solomon, Rick Wash (2012); Bootstrapping wikis: Developing critical mass in a fledgling community by seeding content. Proceeding to CSCW 2012

Rosta Farzan, Robert Kraut, Aditya Pal, Joseph Konstan (2012): Socializing volunteers in an online community: A field experiment. Proceeding to CSCW 2012

Jennifer Thom, David Millen, Joan DiMicco (2012): Removing Gamification from an Enterprise SNS. Proceeding to CSCW 2012

Sign up for The Moz Top 10, a semimonthly mailer updating you on the top ten hottest pieces of SEO news, tips, and rad links uncovered by the Moz team. Think of it as your exclusive digest of stuff you don't have time to hunt down but want to read!

By: [author-name]

This article was syndicated via RSS from: http://feedproxy.google.com/~r/seomoz/~3/mbWo-SRRYiw/what-community-builders-can-learn-from-research

No comments yet.

Leave a Comment

You must be logged in to post a comment.