Introduction
Volunteer crowdsourcing platforms, such as Food Rescue U.S. (FRUS), play a critical role in modern logistics by matching volunteers with recurring tasks like food pickups. A central design question for these platforms is how to manage volunteer commitment. Should a platform encourage "adoption"—where a volunteer commits to a recurring task indefinitely—or should it rely on a "spot market" where volunteers sign up for one-off tasks? While adoption provides immediate stability, this research reveals a hidden trade-off: it can stifle long-term growth by reducing the pool of highly engaged volunteers who drive the platform's expansion.
The Problem: The Double-Edged Sword of Adoption
On the surface, adoption appears to be an ideal solution. It ensures that donors have a guaranteed pickup and provides volunteers with a predictable routine. However, the authors identify a "double-edged sword" effect. Volunteers who adopt a specific match tend to stop checking the app for additional tasks. In contrast, "spot market" volunteers—those who browse and claim individual rescues—are significantly more engaged. They often complete multiple tasks per week and are more likely to pick up new or unadopted donations, which is essential for scaling the platform.
If a platform allows too many tasks to be "locked" into adoptions, the spot market thins out. This leads to a vicious cycle: new, unadopted tasks go unmatched, causing donor frustration and eventual dropout, while volunteers find fewer opportunities to engage, leading to their own attrition. The core challenge is balancing this immediate commitment with the long-term engagement necessary for growth.
Methodology and Data
To investigate this trade-off, the researchers conducted an extensive study using a proprietary dataset from Food Rescue U.S. (FRUS). The data spans nearly two years and includes over 35,000 donations across 30+ locations, involving 1,652 unique volunteers and 906 recurring donation types.
The authors combined this empirical analysis with stochastic modeling to simulate market dynamics where growth is endogenous—meaning successful matches today directly influence the number of volunteers and donations available tomorrow. A key technical contribution is the introduction of a novel matching function: $$\mu(a, b) = a + b - \frac{1}{c} \log(e^{ca} + e^{cb} - 1)$$ Unlike standard Cobb-Douglas functions, this model respects the physical constraints of logistics (you cannot have more matches than the smaller side of the market) while capturing the "increasing returns to scale at a diminishing rate" characteristic of growing marketplaces.
Key Insights: The "All-or-Nothing" Result
The most striking finding of the study is that the optimal policy for platform growth is almost always binary. Mathematically, platforms should generally either allow full adoption or disallow it entirely. Intermediate "partial adoption" levels—such as allowing only a certain percentage of tasks to be adopted—were found to be suboptimal and could even lead to marketplace decay.
The decision to switch from an adoption-heavy model to a spot-market model depends on "market thickness." In thin markets, adoption is necessary for survival to prevent donor dropout. However, once a platform reaches a critical mass of volunteers and tasks, moving to a "no-adoption" policy can supercharge growth by forcing volunteers into the more engaged spot market.
Results and Implications
The empirical data supports the value of the spot market. On average, spot market volunteers completed 1.09 donations in the subsequent week, compared to 0.94 for those in adopted matches. Furthermore, for volunteers not in an adopted match, every additional available donation they saw on the platform decreased their probability of dropping out by 2.85%.
Donors were found to be particularly sensitive to service failures; a single missed pickup made a donor 3.4 times more likely to stop using the platform. This explains why platforms are often tempted to prioritize the stability of adoption, even at the cost of long-term scalability. The study's "repeated myopic policy" (making the best choice for the immediate next step) was exactly optimal in 90.7% of cases, providing a practical heuristic for platform managers.
Conclusion
For managers of volunteer crowdsourcing platforms, the research suggests a strategic evolution. While adoption is a useful tool for stabilizing a nascent marketplace, it can become a barrier to scale as the platform matures. By understanding the engagement gap between adopted and spot-market volunteers, platforms can better time the transition to more flexible matching models that maximize long-term growth and impact.
Summary updated