Do Good, Stay Longer? Temporal Patterns and Predictors of Newcomer-to-Core Transitions in Conventional OSS and OSS4SG

This study compares conventional Open Source Software (OSS) with mission-driven OSS for Social Good (OSS4SG), finding that OSS4SG projects have significantly higher contributor retention and core transition rates, and that taking time to learn a codebase before intensifying contributions (the "Late Spike" pattern) is a more effective strategy for achieving core status than immediate intensive involvement.

The Gist

The Big Idea: "Joining the Club"

Imagine you want to join a local community club. There are two types of clubs:

1. The "Tech Hobbyist" Clubs (Conventional OSS): These are like high-end photography clubs. People join to sharpen their skills, show off their best shots, and maybe build a professional portfolio. It’s competitive, fast-paced, and people often come and go quickly.
2. The "Helping Hands" Clubs (OSS4SG): These are like community garden clubs. People join because they care about a cause—like feeding the hungry or protecting the environment. They aren't just there for the "gear"; they are there for the mission.

The Problem: In almost all these clubs, most people show up once, do one small thing, and then never come back. This "broken pipeline" makes it hard for clubs to survive because they can't find new leaders (the "Core Members") to take over.

Researchers wanted to know: Does the reason you join a club change how likely you are to become a leader?

---

The Findings: What the Researchers Discovered

1. The "Sticky" Factor (Mission Matters)

The researchers found that the "Helping Hands" (Social Good) clubs are much "stickier."

• Analogy: If a photography club is like a revolving door, the community garden is like a magnet.
• People in social good projects stay much longer and are 20% more likely to eventually become leaders. Because they care about the cause, they don't just quit when the work gets hard.

2. The "Map" to Leadership (Pathways)

How do you become a leader?

• In the Tech Clubs: There is basically one "main road" to leadership. You follow a very specific, rigid path: do a task, get it approved, prove you're good, and boom—you're in. If you wander off the path, you're likely out.
• In the Social Good Clubs: There are many different "side streets" and "scenic routes" to leadership. They are more welcoming and give you more ways to prove your worth, such as being given direct access to the "tools" earlier on.

3. The "Slow Burn" vs. The "Flash in the Pan" (Timing)

This is the most surprising part. Most people think that to become a leader, you should start working like a superhero on Day 1. The researchers say: Don't do that.

• The "Early Spike" (The Flash in the Pan): This is the person who arrives, works 100 hours in the first week, gets exhausted or overwhelmed, and then slowly fades away. It takes them a long time (about a year) to become a leader, if they ever do.
• The "Late Spike" (The Slow Burn): This is the person who spends their first few weeks just "looking around the garden." They explore different areas, learn how things work, and don't do too much at first. Then, once they understand the ropes, their activity starts to ramp up.
• The Result: The "Slow Burn"ers become leaders twice as fast (in about 5 months) compared to the "Flash in the Pan"ers (who take a year).

---

The "Cheat Sheet" for Success

If you want to become a leader in a software project, the paper gives you two pieces of advice:

For the Newcomer (The "Joiner"):

• Find your "Why": Pick a project that aligns with your values. You'll be more likely to stick around.
• Be a Tourist first, a Resident second: Don't try to build a skyscraper on your first day. Spend time exploring the "neighborhood" (the code) first. A slow, steady increase in effort is much more effective than a frantic burst of energy.

For the Maintainer (The "Club President"):

• Don't just look at the "Superstars": The person doing 1,000 lines of code in week one might burn out. Look for the person who is exploring many different parts of the project—they are your future leaders.
• Build better "Welcome Signs": Create guides and easy tasks that help people explore the project without feeling lost.

Technical Summary

Technical Summary: Do Good, Stay Longer?

1. Problem Statement

The sustainability of the Open Source Software (OSS) ecosystem is threatened by a "broken pipeline": while many newcomers join projects, most become inactive after their initial contributions, failing to transition into core contributors (the trusted developers who maintain long-term continuity). This creates a dual challenge:

• For Newcomers: A lack of clear guidance and predictable pathways to leadership roles.
• For Maintainers: A lack of reliable early signals to identify and cultivate promising talent.

Furthermore, existing research often treats OSS as a monolithic entity, ignoring how a project's mission (e.g., technical/commercial vs. social good) might fundamentally alter community dynamics and onboarding success.

2. Methodology

The researchers conducted a large-scale comparative empirical study between Conventional OSS and Open Source Software for Social Good (OSS4SG).

• Dataset: 375 projects (190 OSS4SG, 185 OSS) comprising 92,721 contributors and 3.5 million commits.
• Core Definition: Used the 80% Pareto rule (the smallest set of contributors responsible for 80% of commits) to identify core status, recomputed weekly to ensure sustained engagement.
• Research Framework:
  • RQ1 (Structural/Outcomes): Analyzed community metrics (Gini coefficient, Bus Factor, retention rates) and used Survival Analysis (Kaplan-Meier & Cox Proportional Hazards) to measure the probability of achieving core status over time.
  • RQ2 (Predictors/Pathways): Used Machine Learning (Logistic Regression, Random Forest, Gradient Boosting) to identify early behavioral predictors from a contributor's first 90 days. Used First-order Markov Chains to map milestone sequences (e.g., Pull Requests, issue comments, direct commit access).
  • RQ3 (Temporal Patterns): Applied Dynamic Time Warping (DTW) clustering to normalized weekly contribution intensity time series to identify engagement "shapes" (patterns) and ranked their effectiveness using Scott-Knott clustering.

3. Key Contributions

1. First Systematic Comparison: The first study to contrast newcomer-to-core transitions specifically between OSS4SG and conventional OSS.
2. Predictive Framework: Developed a model identifying early behavioral signals (Breadth, Commitment, Momentum, and Scope of Impact) that predict core achievement.
3. Temporal Pattern Discovery: Identified specific contribution "shapes" (e.g., "Late Spike") that minimize the time required to reach core status.
4. Evidence-Based Guidance: Provided actionable insights for both newcomers (how to contribute) and maintainers (how to onboard).

4. Key Results

• OSS4SG Superiority in Retention: OSS4SG projects are "stickier." They retain contributors at 2.2× higher rates and contributors have a 19.6% higher probability of achieving core status compared to conventional OSS.
• Predictors of Success: The strongest predictor of core status across all projects is early broad exploration (modifying many different files and contributing significant code volume in the first 90 days).
• Divergent Pathways:
  • Conventional OSS is highly concentrated, with 61.62% of transitions following a single dominant technical pathway.
  • OSS4SG is more flexible, offering multiple pathways and 4.2× higher rates of Direct Commit Access (trust-based access) during the transition.
• The "Late Spike" Advantage: Contrary to the intuition that "hitting the ground running" is best, contributors who follow a Late Spike pattern (low initial activity that increases over time) achieve core status significantly faster (21 weeks) than those who follow an Early Spike pattern (high initial activity that tapers off, taking 51–60 weeks).
• Pattern Flexibility: OSS4SG supports two effective patterns (Late Spike and Low/Gradual), whereas conventional OSS only rewards the Late Spike pattern for speed.

5. Significance

This research shifts the understanding of OSS sustainability from a "one-size-fits-all" model to a mission-aware model.

• For Newcomers: It suggests that finding a project aligned with personal values (OSS4SG) and focusing on broad exploration rather than immediate high-intensity output is a more efficient strategy for leadership.
• For Maintainers: It provides a roadmap for identifying talent through early "breadth" of activity and suggests that creating diverse onboarding tasks (cross-module issues) can accelerate the transition of newcomers into the core team.