Study Notes on the Success of Online Communities

Are All Successful Communities Alike? Characterizing and Predicting the Success of Online Communities

Authors and Contact Information

  • Tiago Cunha, University of Michigan (tiolivei@umich.edu)

  • David Jurgens, University of Michigan (jurgens@umich.edu)

  • Chenhao Tan, University of Colorado Boulder (chenhao.tan@colorado.edu)

  • Daniel M. Romero, University of Michigan (drom@umich.edu)

Abstract

  • Growth of online communities provides opportunities to examine mechanisms explaining group success.

  • Most research defines community success through a single measure, mainly focusing on the number of members.

  • Proposal of multiple definitions of success is essential for comprehensive understanding.

  • Identification of four core success measures:

    1. Growth in the number of members.

    2. Retention of members.

    3. Long-term survival of the community.

    4. Volume of activities within the community.

  • Findings reveal low correlations among these measures, suggesting distinct types of success that require differing attributes and behaviors.

  • Practical implications concern the design of new online communities and how platforms facilitate these communities.

Keywords

  • Online Communities, Success, Group Dynamics, Reddit

Introduction

Importance of Understanding Community Success

  • Essential for building and maintaining vibrant online communities.

  • Enables user connections in shared interest areas.

  • Not all created communities achieve success; major datasets enable large-scale dynamic studies of community life cycles.

Problem Statement

  • Can we predict a community's success early based on attributes measured in its early lifetime?

  • Existing research often narrowly defines success based on user participation numbers.

  • Success metrics can vary significantly by community type (e.g., a social group versus a cause-based community).

Hypothesis

  • The richness and complexity of online communities necessitate a variety of success measures to understand community achievement better.

Community Success Measures

Proposed Success Measures

  1. Growth in the number of members: Total new users within a year after the community reaches a minimum size (k).

    • Distinction between growth in commenters (Gcom) and posters (Gposters).

  2. User retention: Average monthly retention rate within the first year after attaining k users (R).

  3. Long-term survival: Percentage of activity in the last three months of a 24-month time frame after hitting k.

  4. Volume of activities: Average number of posts and comments in the first year after reaching k members.

Measuring Community Characteristics

  • Formalization of measures through specific metrics involving user engagement levels for both posts and comments.

ResulCorrelation Between Success Metrics

Findings on Success Metric Correlation

  • Pairwise Spearman’s correlations across metrics reveal positive relationships but low correlation values.

  • Highlights diversity in metrics, showing that high numbers in one measure do not guarantee high values in others (e.g., growth does not necessarily indicate retention).

  • Specific examples illustrate varying performance across different community goals and types.

Characterizing Community Success

Defining Community Success

  • Emphasizes that community success varies widely between groups, influenced by nature and purpose.

  • A battery of studies supports the notion that successful communities may exhibit different desirable characteristics.

Detailed Success Definitions

  • Success measured over a fixed timeframe after k users join:

    • Growth:
      G{com} = | igcup{i=1}^{12} U{com}^{(i)} | G{posters} = | igcup{i=1}^{12} U{posts}^{(i)} |

    • Retention:
      R = rac{1}{12} imes igg( rac{ig| U{i o i+1} ig|}{|U{all}^{(i)}|} igg)

    • Survival:
      extSurvivalRate=racextActivityinFinal3MonthsextTotalActivityover24monthsext{Survival Rate} = rac{ ext{Activity in Final 3 Months}}{ ext{Total Activity over 24 months}}

    • Volume of activities:

    • Average posts:
      rac{1}{12} imes igg( ext{Total } T_{posts}^{(i)} igg)

    • Average comments:
      rac{1}{12} imes igg( ext{Total } T_{comments}^{(i)} igg)

Implications of Findings

  • Communities exhibit unique success metrics; success cannot be captured through a singular definition or measurement.

  • The heterogeneity in community behaviors offers insights into effective strategies for community organizers.

Predictive Features of Success

Feature Classification

  • Six categories of features predicting success based on established research norms:

  1. Volume and Speed of Activities: Analyzes the number of posts, comments, and activity timing to predict future growth and activity.

  2. Distribution of Activities: Measures Gini coefficient for content distribution across community members.

  3. User Composition: Prior user activity affects future success; considers experience levels in new members based on past statistics.

  4. Linguistic Style: Evaluates interaction language, sentiment, and adherence to community norms related to user satisfaction.

  5. Social Network Features: Constructs communication graphs regarding user interaction metrics.

  6. Parent Communities: Studies the genealogical structure of communities related to early adopters.

Application of Features in Predictions

  • Utilized logistic regression models to interpret success measures from the set of predictive features, revealing community engagements’ influence on various success aspects.

Results and Interpretation

Prediction Accuracy

  • Predicted success with AUC scores ranging from 0.72 to 0.84 depending on characteristics fit for particular measures.

  • Growth in community commenters and user retention seen as generally easier to predict versus survival metrics.

  • Analysis of feature effectiveness revealed key insights into which attributes most affect community dynamics and success rates.

Design Implications for Community Organizers

Community Founding Strategies

  • Early active participation (even from a small core group) correlates positively with long-term success.

  • Recommendations to maintain steady comment and discussion rates to foster ongoing engagement and participation, preventing burnout and stagnation.

Limitations of Study

  • The study’s control constraints concerning community topics and their associated success patterns.

  • The potential gap in applying findings in real-world community-building scenarios due to the observational nature of the study.

Conclusion

  • Diverse pathways to community success require a comprehensive understanding of user behaviors and engagement strategies.

  • Core findings emphasize the important role that early community actions have in forecasting their potential longevity and success across multiple dimensions.

Acknowledgments

  • Supported by National Science Foundation under Grant No. IIS-1617820.

References

  • Extensive reference list from various in-depth studies demonstrates the breadth of research interconnecting community dynamics, success, and prediction methods.