Subscriptions & Retention

Comprehensive survey of churn prediction methods and evaluation metrics.

Survival analysis approach to churn with right-censoring and dynamic hazards.

The BG/NBD model for customer lifetime value with transaction data.

Simplified BG/NBD model that maintains predictive accuracy with easier estimation.

Shifted-Beta-Geometric model for subscription businesses with discrete renewal periods.

Structural model of freemium conversion analyzing optimal feature restriction.

Practitioner framework for freemium strategy and conversion optimization.

Empirical analysis of the relationship between customer tenure and profitability.

Links satisfaction metrics to retention and share of wallet outcomes.

Foundational paper introducing the Fogg Behavior Model: Behavior = Motivation × Ability × Trigger. The academic framework underlying 'Hooked' and most digital habit formation research. Over 4,500 citations; essential conceptual foundation for product designers.

Seminal literature review establishing empirical evidence framework for gamification effects. Finds gamification provides positive effects on engagement but is highly context-dependent. Over 4,000 citations; the definitive academic synthesis on gamification.

Rigorous RCT examining how user-generated content features affect app engagement. Finds content submission reduces session-ending hazard by 11% and app abandonment by 14%. Methodological gold standard for engagement research.

Large-scale randomized experiment (10,000 learners) showing contest-based interventions increase engagement by 45% and create persistent habit formation effects (75% higher engagement 12 weeks post-intervention). Cutting-edge field experimental evidence from leading economists.

Provides causal evidence linking gamification to outcomes mediated by behavioral changes. Validates Landers' theory of gamified learning with empirical evidence—important for understanding mechanism design.

Comprehensive Cox proportional hazards application to customer churn with time-varying covariates. Demonstrates how demographic, environmental, and behavioral variables predict retention. Over 600 citations; foundational for survival analysis in marketing.

Develops the shifted-beta-geometric (sBG) probability model for contractual settings. Accounts for customer heterogeneity in retention rates and provides forecasts for tenure, lifetime value, and customer base valuation. Essential methodology for subscription businesses.

Novel competing risks survival model predicting both churn propensity and churn reason. Incorporates text analytics (LDA) from customer service transcripts to distinguish service-provider-driven versus competitor-driven churn. Modern integration of NLP with survival analysis.

Combines machine learning (random forests) with competing risks methodology. Computes distinct risk profiles for different churn causes and identifies relationships between risks and customer behavior for targeted interventions.

Introduces the customer equity concept and decision calculus for balancing acquisition versus retention spending. Origins of viewing customers as assets. Over 1,000 citations; bridges practitioner and academic perspectives.

Unified strategic framework enabling marketing options to be traded off based on projected financial return, operationalized as change in customer equity. Key contribution linking marketing decisions to shareholder value through customer equity.

Demonstrates how CLV can value firms including high-growth firms with negative earnings. Shows 1% improvement in retention improves firm value by 3-7%, while margin improvement yields only ~1% increase. Validated on Amazon, eBay, Capital One, E*Trade data.

Landmark study using data from 1 million cellular customers showing exposure to a defecting neighbor increases churn hazard by 80%. Establishes that social contagion affects retention similarly to adoption. Foundational paper for network effects on churn.

Extends social effects research by analyzing directed networks. Finds that only contacts the customer initiates communication with influence churn decisions. Demonstrates importance of relationship directionality and recency of neighbor churn.

Uses gym membership data to study bidirectional peer effects in usage and churn. When high-centrality users churn, cascade effects increase average un-subscription by 7.2% within 6 months. Provides targeting guidance: short-term strategies should target low-centrality users; long-term should target high-centrality.

Comprehensive review from 12 leading CRM scholars covering social connectivity's role in retention. Synthesizes literature on network externalities in multiplayer gaming, communications, and shared services. Identifies research frontier for network-oriented retention.