Week 5

Network dynamics and platform ecosystems - case discussion

Published

December 10, 2025

Prepare

Note📚

📖 Review Week 4 materials on dyads, triads, and network dynamics

📖 Read the case study: “SoundCloud: Tuning the Algorithm of Discovery” (available in Moodle’s case studies section)

📖 Revise key concepts: network effects, multi-layered networks, and creator-listener dynamics

Participate

Note👥

🙋 Case Discussion: “SoundCloud: Tuning the Algorithm of Discovery”

This week we analyze SoundCloud’s recommendation system challenge through the lens of network science. CEO Michael Weissman must choose between three strategic options for overhauling the platform’s music discovery engine:

  1. License existing technology from Google, Amazon, or Spotify
  2. Build a proprietary network analysis system leveraging creator-listener relationships
  3. Develop a hybrid human-algorithmic approach combining community curation with machine learning

Key Discussion Questions:

  • How does SoundCloud’s multi-layered network structure (creators ↔︎ listeners ↔︎ content) create unique opportunities for recommendation systems?
  • What network metrics could identify emerging artists and musical trends before they become mainstream?
  • How should SoundCloud balance algorithm-driven discovery with maintaining its creator community culture?
  • What are the strategic trade-offs between speed-to-market, competitive differentiation, and technical risk?

Practice

Note💻

Hands-on Analysis: Network-based recommendation systems

📊 SoundCloud Network Data - Exploratory Data Analysis

Explore how network structure informs content discovery and recommendation algorithms using real SoundCloud data.

  • Analyze creator collaboration patterns and influence propagation
  • Identify taste-making nodes and community structures
  • Evaluate different recommendation approaches using network metrics
  • Examine user engagement patterns and platform dynamics

The EDA provides a comprehensive analysis of the SoundCloud dataset covering users, tracks, social connections, and engagement patterns from 2019-2020.

📊 SoundCloud Network Properties Analysis - Reciprocity & Transitivity

Test for fundamental network properties in the SoundCloud follow network:

  • Calculate and interpret reciprocity (mutual following patterns)
  • Analyze transitivity/clustering (triangle formation tendency)
  • Compare observed properties against random network baselines
  • Conduct rigorous statistical tests (Erdős-Rényi and CUG tests)
  • Understand what these patterns reveal about community structure

This analysis uses both classical random graph comparisons and Conditional Uniform Graph (CUG) tests to rigorously evaluate whether the SoundCloud network exhibits structured social patterns.

Perform

Note📝

No Assignment This Week

This week focuses on case discussion and understanding how network dynamics shape platform strategy. No performance assignment is required.

Ponder

Note🤔

Platform Ecosystems and Network Effects

Reflect on how network dynamics shape digital platform strategy and value creation:

  • Multi-sided markets: How do platforms balance the needs of different stakeholder groups (creators vs. consumers)? When do their interests align or conflict?
  • Network effects vs. quality: Can platform growth dilute content quality? How can recommendation systems address the signal-to-noise problem at scale?
  • Path dependencies: How do early strategic decisions (e.g., SoundCloud’s creator-first approach) create both opportunities and constraints for future growth?
  • Make vs. buy decisions: When should platforms build proprietary technology versus licensing external solutions? What factors should drive this choice?

Consider other platform examples: YouTube (creators and viewers), Airbnb (hosts and guests), Uber (drivers and riders), or GitHub (developers and projects).

Further Reading:

  • Platform strategy: Parker, G., Van Alstyne, M., & Choudary, S. P. (2016). Platform Revolution. Norton.
  • Network effects in tech: Eisenmann, T., Parker, G., & Van Alstyne, M. (2006). “Strategies for Two-Sided Markets”. Harvard Business Review.
  • Recommendation systems: Ricci, F., Rokach, L., & Shapira, B. (2015). Recommender Systems Handbook. Springer.