Guide to Managing an Online Community

Submitted by admin on Mon, 09/26/2016 - 11:32

Online communities are persistent collections of people with common or complementary interests whose primary method of communication is the Internet (Preece, 2000; cited in Ren et al. 2012).

What Are the Major Differences Between Online and Offline Communities?

  • "Online communities lack the physical cues of the face-to-face world, allowing people to change their identities, and thus reducing the face-to-face world influence of norms on individual behavior” (Andrews, 2002);
  • “Offline communities tend to consist from more traditional networks, such as family and friends, work/study colleagues, members in hobby groups, and neighbors. Online communities, on the other hand, tend to be vaguer in their description. Typically, such communities consist of groups of users, who have agreed to interact and share information with each other through certain technical platforms such as Facebook, MySpace, Twitter or a more designated site for users sharing a particular interest” (Nasi et al., 2011);
  • “Online communities typically have no mechanism to evaluate a member’s potential before they join the community. Anyone can join an online community at any time, but the communities might impose processes to assimilate these new members slowly into the community, allowing them to become full active members only after an apprenticeship period (O’Mahony and Ferraro 2007; Preece and Schneiderman 2009),” (cited in Ransbotham and Kane, 2011);
  • “Online communities offer a common forum for communication among all participating members. As such, this context presents a simple network structure, and diffusion of opinions occurs quite rapidly because messages are nonexclusive. In other kinds of communities, networks build on the basis of dyadic relationships. Communication within dyads excludes other community members and slows the diffusion of opinions” (Miller et al., 2009); and
  • “[In virtual communities] every member enjoys equality in speech, in contrast with centralized systems or traditional, bounded communities that allow certain members greater reward, coercive, and legitimate power” (Tsai and Bagozzi, 2014).

Types of Online Communities

Online communities have orientation types (Kannan et al. 2000) and management types (Porter et al. 2013). Tables 2 and 3 illustrate four orientation and three management types, respectively, providing examples and a brief description for each type.

                                                                                                                                                                                                                                                                                                                     

    

Table 2. Online Community Orientation Type (Source: Kannan et al. 2000)

    

Type Description Examples
Transaction-Oriented Transaction-oriented communities primarily facilitate the buying and selling of products and services and deliver information that is related to fulfilling those transactions.  Amazon.com
Interest-Oriented Interaction-oriented communities focus on topics of the members’ common interest. These communities usually have chat rooms, message boards, and discussion groups for extensive member interaction. Thus they are characterized by significant amount of user-generated content. Café Utne and Motley Fool 
Fantasy-Oriented Fantasy-oriented communities allow users to create new environments, personalities, stories and role-play. ESPNet and Sony.com
Relationship-Oriented Relationship-oriented communities are built around certain life experiences that are usually very intense and lead to personal bonding between members. Cancer Forum

                                                                                                                                                                                                                                                            

    

 

    

Table 3. Online Community Management Type (Source: Porter et al. 2013)

    

Type Description Examples
Third-party-Managed Owned and managed by firms that connect members for the exchange of information.  Epinions and eBay
Firm-Sponsored Organized by a single firm, and often embedded on a firm-owned Web site, in order to foster relationships with customers or facilitate peer support and service. Nutella and Denver 
            Broncos
Customer-Initiated Organized by individual members in order to interact around a shared product/brand interest. Members typically establish an independent Web site or establish the community on a hosting service (e.g., Yahoo! Groups) because customer-initiated communities are not affiliated formally with the firms that are the focus of community interest. Apple Newton and Steelers Fever

 

Getting Started With Your Online Community

There are three major challenges in starting an online community. The first is to define the community’s scope (breadth and topics to cover). The second is to take action by promoting the community’s scope. The third is to quickly attract early members.

Defining Scope

Online communities fail to take off for many reasons. Sometimes, it is unclear whether the community offers “services” or “experiences” that are attractive to potential members (Kraut and Resnick 2012). As an example, consider a forum with posts about French cuisine, US politics, and Australian Rugby. While mixing the three topics may increase the total number of messages in the forum several visitors will soon discover that some of the content is uninteresting to them. As a result, the community is likely to shrink or fail.

Step 1: Do not choose a mixed-topic scope for your community.

Even when the topic-scope is not mixed, an ambiguous topic may make it mixed anyway.

Step 2: Do not choose an ambiguous-topic scope for your community.

In some cases, mixing different topics may not reduce the value of your online community. Consider assembling various football fan communities into a larger football forum. Is this likely to help the communities in general? Maybe. If people are there simply to talk about their own teams, then it is probably best to not integrate. However, if fans of different teams get value from interacting with each other, then it may be better to integrate these communities. The point is to not limit the breath of the community topic.

Step 3: Do not limit the breadth of your community topic.

Taking Action

Open-source platforms such as WordPress offer a depth of knowledge and shared code, making it much easier for you to tailor a community to your own needs, rather than starting to build a platform from scratch.

Step 1: Build your community on an existing open-source platform.

Most communities allow people to perform more than one activity within the community (Kraut and Resnick 2012). Limiting member activity (to one specific thing) will bore people. Such boredom may incentivize potential members to join other communities.

Step 2: Allow people to perform more than one activity in your community.

A common mistake that online community designers make is to create separate forums or chatrooms for each topic embedded in the community (Kraut and Resnick 2012). Initially, most of these spaces will be empty.

Step 3: Initially provide just one space for member interaction.

Many people will not carefully investigate your community, but will instead assess it based on short descriptions or reviews from others that do investigate. Given limited attention from evaluators, it is difficult to completely convey the value of your community. Instead, it is far more effective to identify one or to key elements. This may be a topic not covered elsewhere, a different set of participants (e.g., dating sites for the elderly), a different set of interaction (e.g., Snapchat), or different social norms.

Step 4: Convey a succinct selling proposition.

If the public is much more aware of your community than of C2, then even those who are aware of both know that many people only know your community. Therefore, they will expect your community to be more valuable. Advertisement and celebrity endorsement investments are likely to make your community more valuable.

Step 5: Invest in advertisement and celebrity endorsements.

Attracting Early Members

One way to attract early adopters is to import valuable user-generated content from elsewhere (Kraut and Resnick 2012). For example, MovieLens, a movie rating community, imported a database of ratings that had been gathered by another movie rating site, EachMovie, which was no longer operational. With the EachMovie data available, even the first MovieLens subscribers received useful movie predictions (after entering a few ratings so that the site could calibrate the user’s tastes).

Step 1: Provide access to valuable data.

Another way to compensate for an initially small community is for paid staff to participate and provide the benefits that will eventually be provided by members (Resnick et al. 2010).

Step 2: Request paid staff to participate in the community.

In some cases, software bots can substitute the value that would be provided by other participants. For example, FIFA 14 is a soccer game that can be played on both Xbox Live and PlayStation Network. At least two players are needed for a game session. A player who is tired of waiting for another player can invite bots to play. These might not be as fun to play against, but may be a better option for those not wanting to wait for real opponents.

Step 3: Use bots that simulate other participants.

Offering permanent benefits is another way to convince people to join your online community. For example, Cloudmark, an anti-spam service, offered its service for free and—more importantly—promised a special lifetime subscription rate to its early users when it went commercial.

Step 4: Offer permanent discounts to early adopters.

Early joining benefits may also occur through explicit status markers. “For example, just as American Express prints “Member since” dates on credit cards (and advertises this fact), eBay shows a “Member since” date on its user feedback profile” (Kraut and Resnick 2012, page 263).

Step 5: Promote status benefits to early adopters.

It may be possible to marker your online community as “undiscovered” with the implication that those who adopt early will be recognized by their friends as innovators. Exclusivity, or the appearance thereof, is one way to promote a site as cool. Google did this with Gmail, where users could allow register an account when invited by someone who already had an account.

Step 6: Market your site as interesting but yet undiscovered.

Another way to attract early adopters is to ensure that your community is professional. The prevalence of poor writing, spelling, and layout in scam attacks is one sign of community unprofessionalism, and one of the reasons why people generally don't fall for them (Egelman, Cranor, & Hong, 2008).

Step 7: Enforce a professional design.

In addition to signaling the ability to convey professionalism, it may also be useful to convince members that you’re committed to your community. Visible financial expenditures can be one credible signal of commitment to future investment because it shows members that you’re willing to spend money on the community even if it does not immediately take off.

Step 8: Continuously invest in your community to signal commitment.

Conveying a trajectory of growth in membership and activity is especially helpful in raising expectations about your online community. There are two different, yet complementary, ways to display community growth. One is to acknowledge every new member. Another way is to show the percentage growth of the community, which is better for starters since the total number of members is usually small at that stage.

Step 9: When your community is small, and growing slowly,
     Step 9.1: Acknowledge each new member.
     Step 9.2: Display percentage growth (instead of absolute numbers).

Increasing Contribution in Your Online Community

Requesting Participation

People will not be able to contribute to an online community if they are not aware of the community’s needs and if they do not have the skills and resources necessary to contribute. This is why several online communities publicize lists of needed contributions. For example, the GNOME open source development project announces what the project needs by creating and maintaining reports which count the number of bugs in each of the modules and classifies them by severity and priority. Developers can then follow prioritized lists when contributing. When requesting participation in your community site …

Step 1: Publish lists of needed contributions.

Table 2. Open GNOME bugs (as of September 1, 2014)
Table 2. Open GNOME bugs (as of September 1, 2014)

Some communities provide tools that reduce task-related monitoring costs. Facebook, for example, provides awareness features that show members changes in information generated by their friends (e.g., news feed). The more automatic these tools are, and the easier to use, the more effective they will be.

Step 2: Provide easy-to-use tools for finding and tracking work that needs to be done.

Research on offline volunteerism shows that effective appeals are appeals that match members’ interests. This principle works in online settings as well. Harper, Frankowski, Drenner, Ren, Kiesler, Terveen, and Kraut (2007) found that asking members to respond to posts mentioning movies they had already rated in the movie review portion of the site increased their likelihood of reading and responding to those posts, compared to asking members to reply to random posts.

Step 3: Ask people to perform tasks that interest them and that they are able to perform.

How one asks for contributions makes a difference. Asking a specific question rather than making a general statement or asking an open-ended question increases the likelihood of getting a response by 50 percent (Burke, Kraut, and Joyce 2010). In many cases, it is better to identify particular people and personally ask them to contribute. Requests for help are answered up to 50 percent faster when a recipient is addressed by name than when the request is broadcasted to everyone present in the chat room (Markey 2000).

Step 4: Ask specific people to make contributions.

Many of the requests in online communities involve actions and decisions that people don't care about and are therefore unlikely to evoke deep processing. This will be especially true for newcomers who haven’t yet become committed to the community and don't necessarily care about its welfare. Thus, when asking for small contributions, requests without elaborate justification may be successful for the casual visitor.

Step 5: Ask people to perform simple (rather than lengthy) tasks.

Depth of processing theory indicates that the more people care about the decision domain, the more they will be willing to go through an informal cost-benefit analysis when making a decision. Managers of online communities can use preexisting differences among visitors to their site (by examining participation logs) to differentiate more involved people from less involved ones and to develop different appeals for these members. For example, managers may only want to stress the benefits of a contribution to volunteers possessing high involvement since these members strongly care about the community welfare.

Step 6: Only stress the benefits of a contribution to highly involved members.

Alternatively, managers might use the nature of the request itself as a means to increase people’s contribution. Generally, messages containing a strong fear appeal are compelling (White and Allen 2000) because they cause people to take the decision process more seriously. For example, one public radio station in Pittsburgh had its most successful fund raising campaign in history, raising more than half a million dollars in ten days, when it announced its license was for sale, raising fears of its commercialization.

Step 7: Use fear appealing messages.

Psychologists have long studied the factors that lead to liking (Berscheid and Reis 1998) and have shown that most of the factors that lead one person to like another also increase their ability to persuade each other. People tend to like others who have power and who are more physically attractive (Eagly, Ashmore, Makhijni, and Longo 1991).

Step 8: Ask high-status and attractive people to send community requests. Goal setting can be strategically used to increase contributions in your online community. Beenen, Ling, Wang, Chang, Frankowski, Resnick, and Kraut (2004) show that members rated more movies when they were sent an email asking them to rate a specific number of movies than when the message asked them to do their best. Not only that, but assigning a specific deadline also boosts contributions (Kittur 2011).

Step 9: Provide specific and highly challenging goals.

Step 10: Assign precise deadlines for goals.

Increasing Members’ Motivation to Contribute

People often contribute in online communities because they are either intrinsically or extrinsically motivated to do so. In this section we provide designers with several steps for increasing both types of motivation. Prior research, for example, finds that posting activity is positively influenced by offline social interaction (Koh et al. 2007). Face-to-face meetings increase the social presence of community members.

Step 1: Encourage members to socially connect face-to-face.

Challenges raised by online community requests should match or slightly exceed members’ skills (Kraut and Resnick 2012). The most enjoyable requests are ones in which members feel barely in control. For example, crossword puzzles that are too easy will be boring and crossword puzzles that are too hard will be frustrating, but some puzzles will be enjoyably challenging.

Step 2: Assign challenges that match or slightly exceed members’ skills.

Feedback about one’s performance can be quite motivating (Kraut and Resnick 2012). Positive feedback might be especially motivating because of people’s desire for self-enhancement. Such feedback can be verbal, in the forms of comments from other people, or nonverbal, in the forms of quantitative performance measures (e.g., achieving a score of 5 on a comment). In fact, quantitative feedback may lead to more verbal feedback. Systematic feedback also seems to reduce turnover of sites, at least for people who ask questions (Moon and Sproull 2008).

Step 3: Provide frequent performance feedback.

Step 4: Specifically encourage systematic quantitative feedback.

Positive feedback only enhances people’s intrinsic motivation when receivers perceive the praise to be sincere—if not, they will not interpret it as a reliable signal about competence and may perceive it as controlling. Moreover, praise is likely to be seen as most credible if it reflects the judgment of the entire community (i.e., it comes from multiple independent sources).

Step 5: Do not provide positive feedback for performances that are undeserving.

Performance feedback that is comparative in nature can also be motivating. Some people enjoy the thrill of competition and the feeling of beating a contestant. Comparative performance feedback, however, can also decrease motivation, especially for people who feel ambivalent about their participation in the community and for those who think that they have done “enough” for the community. Additionally, members might be discouraged from even trying if the performances of others are seen as unattainably high.

Step 6: Provide comparative feedback if high performance is desirable and obtainable.

While intended to have a positive outcome, rewards sometimes create the wrong incentives. Prizes contingent solely on task completion, rather than on quality, are especially vulnerable to the counterfeit action of low effort. For example, in an effort to increase post counts, some users may contribute many short and not very informative posts. Yet, gaming incentives can be reduced through the implementation of a performance metric exhibiting higher scores for people exerting higher effort on the task.

Step 7: Use performance metrics when offering rewards for effort-contingent tasks.

Status and privileges within the community may induce less gaming than rewards that are valuable outside the community because status and privileges within the community may not be very valuable to people who do not make genuine contributions to the community.

Step 8: Offer status and privilege rewards instead of tangible prizes.

An alternative tries to limit gaming by making it harder for the attacker to find counterfeit actions, rather than by eliminating incentives for an attacker to choose the counterfeit actions. Imagine that an attacker is trying to get rewards by performing low-cost actions that do not actually contribute to the community. If the eligibility criteria are transparent, it will be easy for the attacker to find actions that will be eligible. Moreover, if the schedule is predictable, the attacker will get immediate feedback about whether particular action was successful in meeting the eligibility criteria and can thus learn quickly which actions to keep doing. By contrast, if the criteria are not transparent and the schedule is unpredictable, it will be harder for an attacker to find a set of rewarded actions that he or she can undertake at low cost.

Step 9: Set unpredictable reward schedules for performance- and effort-contingent tasks.

Step 10: Set non-transparent eligibility criteria for performance- and effort-contingent tasks.

Both psychologists and economists have argued that one should be careful about providing rewards and other extrinsic motivators for activities that people find intrinsically interesting, because doing so undermines their intrinsic interest in the task. Tangible incentives seem to undermine intrinsic motivation in part because they undercut people’s feelings of autonomy and competence (Deci, Koestner, and Ryan 1999).

Step 11: Do not offer monetary rewards for intrinsically motivating tasks.

People are more likely to contribute to a group if they think that their contributions make a difference to the group’s performance. One way to influence beliefs about the efficacy of individual effort on group performance is to create topic subgroups (Latane and Nida 1981).

Step 12: Create topic subgroups.

If people believe that their contributions are merely substitutes, then there is little reason to contribute because their contributions have little likelihood of influencing group outcomes. Conversely, if they think that their contributions are complementary and unique, they should be more motivated to contribute because their contributions are likely to influence the group.

Step 13: Encourage members to make opposite, rather than substitute contributions.

Step 14: Send “thank-you” notes to members who provide unique contributions.

Dealing With Problems in Your Online Community

Making Norms Clear

The first step in reducing problems in your online community is to develop and follow a set of appropriate behavioral guidelines, which can emerge from what managers and most members consider acceptable behavior. A guideline is helpful because it defines community boundaries as to what is considered acceptable behavior and what is not.

Step 1: Create and follow explicit appropriate behavior guidelines.

There are multiple different ways to highlight appropriate behaviors in your online community. One strategy is to simply display the guidelines on the community’s website. Another tactic involves providing examples of appropriate behavior. It is also good to publicize inappropriate behavior so that members know what to avoid. However, displaying many examples of inappropriate behavior on the site may lead members to believe such behavior is common and expected. Providing an explanation as to how these behaviors can harm your community may also be useful since it provides members with feedback. A fourth way to encourage appropriate behavior is to show statistics confirming that members in your community follow the norms. On your community site...

Step 2: Display your guideline alongside examples of appropriate behavior.

Step 3: Display a couple (not many) examples of inappropriate behavior.
     Step 3.1: Explain why inappropriate behaviors are detrimental to the community.

Step 4: Display statistics that highlight the prevalence of appropriate behavior.

Limiting Inappropriate Behavior

In asynchronous conversation communities, posts can and should be screened. In several forums, moderators often remove inappropriate messages after they are posted or move them to other forums where they may be more appropriate. The use of moderators limits the impact of inappropriate behavior because it reduces the number of “read inappropriate messages.” However, moderators’ decision must be consistent as a means to ensure procedural justice. Procedural justice considerations also have implications regarding who is chosen to make moderation decisions. A good approach is to allow community members to select, through a voting system, other members to be moderators. Members who are selected by the community are more likely to be perceived as “impartial.” The moderation system should also prevent a single moderator from exercising a “reign of terror.”

Step 1: Hire moderators to prescreen, degrade, label, or move, inappropriate messages.
      Step 1.1: Train moderators to be consistent in their decisions.
      Step 1.2: Hire moderators who are members of the community.
      Step 1.3: Ensure that moderators are impartial.
      Step 1.4: Select moderators who have limited power.

Providing quick-reversion tools that allow certain community members to repair damage are useful because these can also help reduce the number of “read inappropriate messages.”

Step 2: Provide built-in tools that allow members to revert any document.

Researchers have developed several algorithms aimed at detecting suspicious ratings exhibited online (Resnick and Sami 2007; Mobasher, Burke, Bhaumik, and Williams 2007). These algorithms are helpful in decreasing the impact of counterfeit ratings. The influence limiter (For more information on influence limiter, please read: The Influence Limiter: Provably Manipulation-Resistant Recommender Systems (Appendix)), for example, does not throw out suspect ratings but instead partially discounts them. The only issue with algorithms is their inability to always perfectly partial out trolls (Resnick and Sami 2008).

Step 3: Do not ignore trolls. 
     Step 3.1: Use a system that partially discounts suspect messages.

Quota mechanisms are one tactic to prevent large-scale troll damage. For example, chat rooms can automatically block users that post too many messages in a short spam of time. It is important for designers to not disclose their quota criteria (if they do trolls will game the system). Moreover, quota mechanisms can be designed to send alerting messages to users performing suspicious activities.

Step 4: Implement activity quotas designed to prevent large-scale troll damage.

Step 5: Do not disclose quota criteria.

Step 6: Design quota mechanisms to send alerting messages.

Require members to complete CAPTCHAs and identity checks to prevent damage from automated accounts. In addition, advise members to not follow suspicious links by announcing such links on your community site. Lastly, allow members to report a limited number of suspicious messages. If you allow members to report unlimitedly people may start using this feature for the wrong reasons (e.g., to eliminate other users because of personal motives).

Step 7: Prevent the creation of fake accounts with CAPTCHAs or identity checks.

Step 8: Advise members to not follow spam links.

Step 9: Allow members to report a limited number of suspicious behaviors.