The Hidden Problem Driving Delivery Workers to Quit That Isn’t Just About Low Pay

At 6 p.m. on a rainy Friday evening, Maria’s phone buzzed again. Her gig app was sending her back to the same burger joint for the third time that night. Frustrated, she sighed and turned her car around. “I feel like I’m driving in circles,” she muttered.

Maria’s experience isn’t unique. Thousands of gig workers across the country are caught in a frustrating cycle. They take inefficient routes, backtrack to the same locations, and are overloaded with orders while other drivers sit idle.

Many believe gig worker frustration comes down to low pay or poor tipping. However, researchers have uncovered a deeper and more surprising cause: flawed delivery algorithms.

A new study reveals that the same systems designed to optimize deliveries may be sabotaging gig workers with unnecessary extra trips. These wasted journeys push workers to exhaustion, making the job far less rewarding than it could be.

Hidden within the problem is a potential solution that could reshape the gig economy.

Why Are Delivery Workers Quitting at Record Rates?

Delivery platforms like DoorDash, Uber Eats, and Instacart rely heavily on algorithms to manage orders. Each request is analyzed in milliseconds. The system balances factors like driver proximity, delivery time, and customer location to minimize delays.

On paper, this system seems efficient. However, researchers discovered a troubling pattern. These algorithms often overload some drivers with redundant trips, forcing them to retrace their steps, while other drivers are left waiting for assignments.

Computer scientist Hadi Hosseini from Penn State University explained that their research revealed how existing systems frequently push certain workers to take on far more tasks than others. The result is that some drivers are unintentionally forced to make redundant trips, even when better alternatives exist.

This imbalance stems from a critical flaw in how most delivery systems are designed. Traditional algorithms focus on minimizing travel time rather than fairly distributing workloads.

As a result, some drivers are bombarded with tasks, while others are barely given enough work to make ends meet.

This inefficient assignment problem becomes particularly noticeable in high-density areas. Multiple drivers are available, yet they still experience uneven workloads.

For example, Maria might be sent five miles away to deliver a meal. Meanwhile, another driver positioned just blocks from the customer receives no orders at all.

Worse, some workers are sent back to the same restaurant multiple times because the system fails to optimize task sharing.

The result is wasted time, extra fuel costs, and mounting frustration for drivers. All of these factors contribute to high turnover rates in the gig economy.

The Breakthrough: Fixing the Algorithmic Blind Spot

Researchers Hadi Hosseini from Penn State University and Šimon Schierreich from Czech Technical University were determined to find a solution. They began by analyzing the delivery algorithms used by major platforms like DoorDash, Uber Eats, and Instacart.

During their investigation, they discovered that the core issue stemmed from a failure to recognize when a delivery task could be reassigned to a more suitable driver. This oversight resulted in redundant trips, increased fatigue for some workers, and wasted potential for others who were left idle.

The researchers realized that fairness and efficiency were rarely addressed together in gig work systems. Algorithms often sacrificed one for the other. Delivery apps focused heavily on speed, often pushing some drivers to complete multiple trips while others received no work at all.

This pattern was particularly damaging in busy urban areas where high order volumes created chaotic conditions. In one simulated scenario modeled on real gig work conditions, a driver was sent back to a restaurant three times in under 30 minutes. Each visit added minutes of wasted travel time, while another driver waiting just blocks away remained idle.

The researchers identified this pattern as a fundamental design flaw. Hosseini explained that existing systems were not designed to prevent wasted trips, which convinced the team that a better solution was needed.

Their investigation led to a breakthrough called Non-Wastefulness, a method designed to ensure that no worker is burdened with extra trips if another available worker can handle the task without additional cost or delay.

In technical terms, Non-Wastefulness addresses a critical flaw known as the Price of Non-Wastefulness (PoNW). This measure calculates the efficiency trade-off between fair task distribution and optimal delivery speed. The study showed that balancing both could reduce redundant travel without slowing down overall operations.

Importantly, the algorithm is highly flexible. The researchers designed it to integrate easily into existing delivery systems without major structural changes. They found that transforming any current system into a non-wasteful model could be achieved in linear time. This achievement was rare in algorithm design and made the solution practical for large-scale gig work platforms.

Hosseini emphasized that turning any current system into a non-wasteful model is surprisingly simple. Although the adjustment seems small, the research team found that it could have a significant impact.

Unlike traditional approaches that focus only on minimizing delivery times, Non-Wastefulness prioritizes both fairness and efficiency. The algorithm identifies inefficient assignments and redistributes tasks in a way that balances workloads without sacrificing delivery speed.

How the Non-Wastefulness Algorithm Could Save the Gig Economy

Imagine two drivers, Alex and Sam, both waiting outside a busy pizza shop.

In a typical algorithm, Alex might be assigned three deliveries. One order is heading north, one south, and one is just a block away. Meanwhile, Sam remains idle.

The existing algorithm assumes that assigning all three deliveries to Alex minimizes overall travel time. However, this solution often leads to burnout for Alex, while Sam’s availability goes unused.

The Non-Wastefulness Algorithm would recognize that Sam could efficiently handle the nearby delivery. By offloading that task to Sam, the system prevents Alex from taking unnecessary extra trips while ensuring Sam stays productive.

In real-world testing, this adjustment significantly reduced travel inefficiencies. The algorithm consistently identified situations where tasks could be reassigned to improve both worker satisfaction and delivery speed.

In another simulation, researchers found that applying Non-Wastefulness reduced the number of redundant trips by 20-30%. This improvement resulted in less backtracking for drivers, shorter delivery times, and improved overall efficiency.

Even in complex delivery conditions, the algorithm performed well. In dense urban grids with multiple intersections, traffic delays, and high order volumes, the system maintained efficiency while distributing tasks more evenly across available drivers.

The researchers found that balancing fairness and efficiency was key. By combining the two, they created a system that prevented burnout without sacrificing customer satisfaction.

Another strength of the algorithm lies in its ability to handle complex scenarios involving junction points, which are locations where multiple delivery paths intersect. Traditional systems often become inefficient in these areas. However, the Non-Wastefulness Algorithm efficiently redistributes tasks to prevent drivers from crossing paths unnecessarily.

For example, if two drivers are near a busy intersection and one receives multiple orders that require backtracking through the same streets, the system identifies the optimal reassignment strategy to reduce overlap. This feature allows platforms to improve delivery times without overwhelming individual drivers.

The Real Impact on Gig Workers

For gig workers like Maria, the Non-Wastefulness Algorithm could transform their experience on delivery platforms.

In gig work, driver turnover is notoriously high. Studies show that up to 50% of delivery workers quit within their first year, often citing stress, fatigue, and unpredictable earnings as primary reasons. The Non-Wastefulness Algorithm directly addresses these concerns by improving task distribution.

In one scenario modeled in the study, researchers found that introducing the Non-Wastefulness Algorithm resulted in a 25% decrease in the number of delivery orders requiring backtracking or multiple visits to the same location.

By reducing these inefficiencies, drivers experienced shorter shifts with improved earnings consistency. For many workers who rely on gig work as their primary income, this improvement could mean the difference between financial stability and unpredictable wages.

The algorithm’s impact extends beyond improved schedules. Fewer wasted trips also mean reduced fuel costs and less vehicle wear and tear. For gig workers who depend heavily on their cars, these savings are significant.

Reducing wasted travel not only improves efficiency but also allows drivers to earn more without adding to their workload. By reducing fatigue, the algorithm minimizes the risk of accidents caused by stress or exhaustion.

Schierreich observed that fairer distribution meant that drivers could work shorter hours without losing income. Some even managed to increase their earnings by using their time more efficiently.

This solution is not just about improving worker conditions. By ensuring fairer task distribution, the system also enhances platform reliability. Customers benefit from faster deliveries, while businesses experience improved driver retention.

The researchers emphasized that improving fairness is not just an ethical decision. Companies that adopt the Non-Wastefulness Algorithm can improve efficiency, reduce driver burnout, and strengthen customer satisfaction all at once.

As delivery platforms face increasing pressure to address worker dissatisfaction and improve conditions, this algorithm offers a practical and effective solution.

Why Delivery Apps Haven’t Fixed the Problem Yet

Despite its clear benefits, most delivery platforms have yet to implement Non-Wastefulness.

One major reason is that companies are deeply focused on speed metrics. Delivery apps are locked in competition, where faster delivery times are seen as the ultimate measure of success.

As a result, these platforms prioritize delivery speed, sometimes at the cost of fairer distribution.

“There’s a perception that improving fairness might slow down deliveries,” Hosseini noted. “But our model shows you can improve efficiency and reduce driver burnout at the same time.”

Another reason may be algorithm inertia. Delivery platforms have spent years refining their systems. Changing these algorithms, even for the better, requires significant testing. Some companies may hesitate to pursue that.

“The tools are there,” Schierreich emphasized. “But companies need to see that fairness isn’t just about ethics. It’s about improving the system as a whole.”

The Future of Delivery Work: A Path to Fairer Systems

The gig economy thrives on flexible labor. However, the current system isn’t sustainable.

Studies show that 50% of gig workers quit within their first year, with poor workload balance being a leading complaint.

The Non-Wastefulness Algorithm offers a way to fix this. By improving fairness without sacrificing speed, platforms can reduce worker turnover. Meanwhile, they can maintain their reputation for fast deliveries.

For consumers, this could mean more reliable delivery services. For gig workers like Maria, it could mean a healthier, more rewarding way to earn a living.

What Happens Next?

For now, gig workers remain at the mercy of algorithms that may be unknowingly sabotaging their efforts.

But the solution is clear. Non-Wastefulness could redefine the way delivery work is managed. This improvement would enhance both fairness and efficiency.

As Maria pulled back into the parking lot for her third pickup that night, she couldn’t help but wonder, “If they fix this, I might actually stick around.”

The gig economy’s biggest flaw may not be low wages. It’s the hidden inefficiency in delivery algorithms. With a simple algorithmic fix on the table, the future of fair gig work may be closer than we think.

Do you think delivery apps should change their algorithms to protect gig workers? Share your thoughts!

Reference: “The Algorithmic Landscape of Fair and Efficient Distribution of Delivery Orders in the Gig Economy” by Hadi Hosseini and Šimon Schierreich, 21 March 2025, arXiv. DOI: arXiv:2503.16002