Online Algorithms for Matching Platforms with Multi-Channel Traffic

Vahideh Manshadi,Scott Rodilitz,Daniela Saban,Akshaya Suresh
DOI: https://doi.org/10.48550/arXiv.2203.15037
2023-08-02
Abstract:Two-sided platforms rely on their recommendation algorithms to help visitors successfully find a match. However, on platforms such as VolunteerMatch (VM) -- which has facilitated millions of connections between volunteers and nonprofits -- a sizable fraction of website traffic arrives directly to a nonprofit's volunteering page via an external link, thus bypassing the platform's recommendation algorithm. We study how such platforms should account for this external traffic in the design of their recommendation algorithms, given the goal of maximizing successful matches. We model the platform's problem as a special case of online matching, where (using VM terminology) volunteers arrive sequentially and probabilistically match with one opportunity, each of which has finite need for volunteers. In our framework, external traffic is interested only in their targeted opportunity; by contrast, internal traffic may be interested in many opportunities, and the platform's online algorithm selects which opportunity to recommend. In evaluating different algorithms, we parameterize the competitive ratio based on the amount of external traffic. After demonstrating the shortcomings of a commonly-used algorithm that is optimal in the absence of external traffic, we propose a new algorithm -- Adaptive Capacity (AC) -- which accounts for matches differently based on whether they originate from internal or external traffic. We provide a lower bound on AC's competitive ratio that is increasing in the amount of external traffic and that is close to (and, in some regimes, exactly matches) the parameterized upper bound we establish on the competitive ratio of any online algorithm. We complement our theoretical results with a numerical study motivated by VM data that demonstrates the strong performance of AC and furthers our understanding of the difference between AC and other commonly-used algorithms.
Computer Science and Game Theory
What problem does this paper attempt to address?
### Problems the Paper Attempts to Solve The paper attempts to solve how a two-sided matching platform can effectively integrate traffic from different channels to maximize the number of successful matches. Specifically, the paper focuses on volunteer service platforms like VolunteerMatch, which attract volunteers not only through the platform's internal recommendation algorithm but also from external links (such as social media or email notifications). External traffic typically visits specific volunteer service pages directly, while internal traffic can be guided to multiple volunteer service opportunities through the platform's recommendation algorithm. The main research question of the paper is: **How to design a recommendation algorithm that allows the platform to efficiently utilize both internal and external traffic to maximize the number of useful successful matches?** ### Background and Motivation 1. **Multi-channel Traffic**: The platform attracts traffic through various channels, including internal traffic (users entering through the platform's homepage) and external traffic (users directly entering specific pages through external links). 2. **Impact of External Traffic**: External traffic poses challenges to the platform's recommendation algorithm because this traffic is usually only interested in specific service opportunities and cannot be redistributed by the platform. 3. **Data Observation**: Through the analysis of VolunteerMatch data, the authors found that external traffic accounts for a significant portion of the total traffic, and there are significant differences in the number of check-ins and sources among different service opportunities. ### Research Objectives - **Optimize Recommendation Algorithm**: Design a new recommendation algorithm that can maximize the number of useful successful matches in the presence of external traffic. - **Theoretical Analysis**: Evaluate the performance of different algorithms through the competitive ratio and propose a new algorithm—the Adaptive Capacity (AC) algorithm. - **Empirical Study**: Validate the performance of the AC algorithm on actual data through simulation experiments and compare it with existing algorithms. ### Main Contributions 1. **Model Construction**: Proposed a framework for the online matching problem, considering the situation of multi-channel traffic. 2. **Algorithm Design**: Designed the AC algorithm, which effectively deals with external traffic by dynamically adjusting the remaining capacity of each service opportunity. 3. **Theoretical Guarantee**: Proved that the AC algorithm can achieve an approximately optimal competitive ratio under certain parameter settings. 4. **Empirical Results**: Demonstrated the superior performance of the AC algorithm in practical applications through a case study based on VolunteerMatch data. ### Conclusion Through theoretical analysis and empirical research, the paper demonstrates how to design an efficient recommendation algorithm in the presence of external traffic to maximize the number of successful matches on a two-sided matching platform. This research has significant practical implications for volunteer service platforms similar to VolunteerMatch.