Social Media Participation and Customer Value

Motivation and Research Gap

• Firms rapidly invest in social media but struggle to quantify returns
• Extant research focuses on UGC (reviews, diffusion) using aggregate data
• Critical gap: no study connects individual customers’ social media participation to their actual purchase behavior
• Skepticism persists — is social media just “angst-filled adolescent” hype?

Research Questions and Hypotheses

Central question: Does customer participation in firm-hosted social media strengthen the customer-firm relationship?

• H1: Participation → (+) visit frequency
• H2: Effect amplified by higher social media activity (postings)
• H3: Stronger for high-spending customers
• H4: Weaker for customers with narrower buying focus
• H5: Weaker for deal-sensitive customers
• H6: Stronger for customers buying premium products

Also examines: impact on customer profitability

Data and Empirical Setting

• Large specialty retailer (wine & spirits), northeastern US
• Novel multi-source dataset:
  • Social media participation (Facebook “fans”)
  • Individual transaction data (prices + costs)
  • Survey data (demographics, attitudes)
• 394 treatment customers matched to 287 controls
• Time period: January 2008 – March 2011 (82 weeks pre-launch)
• Firm launched social media site: August 2009

Identification Strategy

Challenge: Self-selection — loyal customers more likely to join social media and visit more frequently

Solution: Propensity Score Matching (PSM) + Difference-in-Differences (DID)

• PSM: Match treatment and control customers on pre-treatment observables
• DID: Compare pre vs. post changes between matched groups
• DDD extension: 3-way interactions for moderating hypotheses

Main Results — Visit Frequency

• Treatment group visit frequency increases significantly post-launch
• Control group shows no significant change
• Elasticity of participation on visit frequency ≈ 5.2%

Moderating Effects (DDD Results)

Three-way interactions — TreatD × CParT × Moderator:

All six hypotheses supported at p < 0.001.

Impact on Customer Profitability

• DID analysis replicated with customer profits as DV
• Treatment effect: positive and significant across all specifications
• Profit elasticity of social media participation ≈ 5.6%

Robustness Checks

• Alternative matching: Greedy matching and Mahalanobis distance → similar results
• Varying time windows: 3, 6, 9, 12, 15-month windows → estimates converge and remain significant
• Phantom regressions: Placebo treatment dates → all estimates indistinguishable from zero
• Serial correlation: Aggregate pre/post estimation + Newey-West SE → results hold
• Additional covariates: Internet risk aversion, tech comfort, social networking attitude → results robust
• Cohort effects: Early vs. late participants → substantively similar

Managerial Implications

1. Nurture relationships through social media — a Facebook page alone is not enough; active engagement and community management matter
2. Not all customers are created equal — integrate offline transaction data with online social media insights for targeting
3. Maintain high activity levels — more postings amplify the participation effect
4. Create premium product communities — specialized forums for high-value products boost returns
5. Social media complements CRM — use transaction history (spend, deal sensitivity, focus) to prioritize social media engagement

Limitations and Future Research

• Single firm, single category (wine) — generalizability across industries and platforms
• Message valence not examined (90%+ positive in this context)
• Message content types (product vs. promotional vs. community) not disentangled
• Future work:
  • Cross-platform and multi-firm studies
  • Content design and cadence optimization
  • Long-term CLV pathways
  • Different social media strategies

Citation: Rishika, R., Kumar, A., Janakiraman, R., & Bezawada, R. (2013). Information Systems Research, 24(1), 108–127. https://doi.org/10.1287/isre.1120.0460