Driving Accountability and Improving Workforce Reliability in an On-Demand Staffing Marketplace
Introduction
Anthor is a workforce marketplace, delivering on-demand staffing solutions to grocery and CPG retailers. However, the seller-driven model created a significant challenge: workers were frequently canceling jobs, providing poor service, and missing critical deadlines. This lack of accountability was not only damaging user experience but also putting the platform's reputation at risk.
Upon analyzing user performance data, we identified a pattern—80% of operational failures stemmed from a small cohort of workers. These recurring issues, including last-minute job cancellations and no-shows, were significantly impacting client relationships and trust in the platform.
To tackle this, we implemented a performance-driven accountability system. By introducing penalties for unsatisfactory behavior—such as limiting access to future gigs for repeat offenders—we aimed to reinforce professionalism and improve the overall quality of service. This approach not only reduced operational disruptions but also boosted customer satisfaction and platform reliability.
Example of penalty algorithm.
My contributions
I led the problem analysis phase by collaborating closely with key stakeholders—operations, engineering, and customer success—gathering critical insights from user behavior data, performance metrics, and direct feedback from both clients and gig workers. This deep dive into data allowed me to identify the root causes of operational failures and define a targeted approach to drive accountability.
Based on these insights, I developed a clear and actionable roadmap with a five-week implementation timeline. I coordinated with the engineering team to integrate the penalty system into our existing infrastructure, ensuring seamless technical execution without disrupting ongoing operations. I also set measurable objectives to track performance improvements and developed comprehensive training materials and rollout guides to align internal teams and prepare users for the system changes.
After launch, I continued to monitor the system’s performance, implementing continuous improvements based on real-time feedback and data analysis. By refining penalties and incentives, we were able to maintain high levels of worker engagement while driving substantial improvements in reliability and service quality.
Results achieved
Increased gig delivery success rate from 75% to 85%, even during peak periods of daily gig requests, significantly improving platform reliability.
Boosted average worker reliability from 40% to 60%, directly addressing the issue of accountability and improving overall workforce quality.
Reduced daily gig cancelations from 400 to 130 average, achieving a 1:3 ratio between cancelations and completions over the following months, enhancing client satisfaction and operational efficiency.
Graph compares the total number of gigs (missions) finished each week and total number of gigs canceled in the same time frame. Our solution was released early August, where you can see a clear reduction in cancelations.
