Uber, a billion-dollar ride-sharing company, is seeking to make choices for its latest product “Express Pool”. This ridesharing service is cost-efficient, but requires longer wait times and walking for customers to reach the pick-up location. The management team needs to decide if it should increase the wait time from two to five minutes and conduct further experiments before rolling out the product in different markets.

Uber's business strategy is to constantly innovate and improve their dynamic algorithms and systems. They follow a continuous improvement approach that forces their product teams to improve the customer experience, driver experience, and innovate their systems and algorithms. They run multiple experiments concurrently or in controlled isolation as needed to improve company’s financial performance.

To experiment and enhance their products, Uber uses Level A/B experiments, switchback experiments, and synthetic control experiments. These approaches allow a recalibration of the algorithm to make a cost-efficient product that is at least 20% cheaper than existing Uber Pool rides. Additionally, Uber uses historical data to model parameter changes' success rate. Demand curves and customer feedback help the management team calculate the impact of algorithm parameter changes. However, results of experiments may get affected due to external factors like weather and the type of city where the experiment was conducted.

The goal is to maximize the profit by finding the optimal wait time that reduces cost per ride for customers, but also limits wait time to a maximum of 15 minutes to prevent customers from canceling their ride.

Angel Anguiano – Brief Summary:

The case describes the launch and development of Uber's “Express Pool” feature, which aims to reduce ride costs by optimizing the route through a longer waiting time and walking to the pick-up point for customers. The feature was introduced in some cities to gauge their response. Uber has to decide whether to increase waiting time from two to five minutes. Uber was founded in 2009 and is the first ride-sharing company in the U.S. In various cities, Uber offers different ranges of products. Uber shows two sides of the app to its riders and the drivers. Customers can enter the destination in the app, and the app displays ride options with the expected fare. On the driver's side, they can accept or decline the ride. In August 2017, the company adopted two strategies: first, ask the riders to wait up to two minutes for matching them with co-riders, and second, walking to the pick-up and drop-off points. Later in November 2017, Uber launched the express feature in Boston and San Francisco to test their project. While San Francisco showed an increase in volume and cost savings with no degradation in marketplace metrics, Boston received negative feedback from riders who didn't like to wait in cold weather or walk to their pick-up point. The experiment increased cancellations but decreased per-ride cost. Despite some issues, Uber launched the Express feature for twelve US cities. The main debate is between a longer waiting time that may negatively impact customer service and the economic benefits of letting the algorithm choose the optimal ride.

Eduardo Alonso – Brief Summary:

Uber, one of the world's largest ride-hailing companies, desired to develop another revenue stream, which targeted carpooling. This was not an easy task, as several teams worked on different aspects simultaneously, causing a risk of contamination and spillover effects. Furthermore, the external factors were challenging to deal with.

Uber's data showed that many people traveled along similar routes at similar times, leading to the development of Uber POOL. The company needed to figure out if this product was necessary and how people would react to riding with strangers, other than the driver.

In 2014, Uber introduced a carpooling service which offered a reduced price to passengers willing to share rides with others. This service maintained its door-to-door convenience without requiring passengers to walk. The aim was to increase seat utilization and efficiency. Drivers were compensated based on ride time, distance, surge rates, and also received an extra fee for each additional passenger. However, the service remained unprofitable in 2016, forcing Uber to identify a new strategy. By 2017, two critical strategies emerged; waiting for up to two minutes while the algorithm matches riders to co-riders and walking a small distance to and from pick-up and drop-off points. These strategies continue as ongoing processes where data analysis is used to improve efficiency and ultimately, profitability.

Crystal Ballin – Overview:

Uber is a ride-sharing company that emerged in the late 2000s, with competition from other companies such as Lift and Wingz. Independent drivers provided services to individuals in need of a ride, in a highly competitive industry with flexible employment rules. Uber, initially known as UberCab, received a cease-and-desist order from the California Public Utilities Commission for operating without a taxi license. As a result, Uber removed the word 'cab' from their name and continued operating in multiple cities, offering different options such as Uber XL and UberLUX. Uber's innovative spirit continued with various features and product improvements, with numerous experiments like User-level A/B testing, Switchbacks, and Synthetic Control Experiments tested for improving the user experience.

With the addition of new services, including the Express Pool, Uber encountered issues with excessive wait times for passengers. Data scientist, Duncan Gilchrist, was tasked with exploring the balance between the ride's cost and the wait time. Data scientists played a critical role in Uber, focusing on scaling the business while solving complex challenges through data-driven decisions.

Question 3:

The success of Uber's expansion was due to their leadership team's decision to implement a new program. By gathering an abundance of customer data, they were able to make informed decisions regarding the project. They discovered that customers were likely to cancel their ride if they had to wait longer than two minutes, but the benefits of cost reduction outweighed the potential loss of customers. Additionally, their team was staffed with representatives from various areas of expertise, which allowed for comprehensive coverage of the project. Despite their successes, there were still areas for improvement, such as finding a more efficient way for drivers to travel and determining a better pricing strategy.

Question 4:

Increasing the wait time for riders from two minutes to five minutes resulted in a 7% increase in ride cancellations, whereas 2-minute wait times showed a higher rate of completed trips and fewer cancellations. Completed "Express" rides decreased for those riders who waited five minutes, while "POOL" trips increased. Overall, drivers received a higher payout for 2-minute wait times due to a lower rate of cancellations.

Conclusion:

Uber's innovation is fueled by feedback from riders, drivers, and experiments. Investment in data science has allowed Uber to gain a deeper understanding of how users interact with their platform. Their diverse range of experiments, including User-level A/B, Switchbacks, and Synthetic control experiments, allows them to improve their products gradually. Express Pool was one such project that utilized a team of data scientists, specialists, managers, and engineers in order to optimize the rider and driver experience. After analyzing the data, it was clear that a two-minute wait time was the best option, as it resulted in fewer cancellations and a higher rate of payout for drivers. Based on these findings, we recommend that Uber continue to implement a two-minute maximum wait time for their riders.