Heavy Lyfting

Read the case study on Lyft and answer the questions that follow.

Lyft joined the ride-sharing market in 2012, three years after Uber started the industry in the U.S. However, Lyfts decisions regarding the use of technology has helped the organization overcome some of the struggles Uber faced as a start-up. The use of Big Data, the cloud, data mining, web mining, and data and cloud analytics have played major roles in the success of these ride-sharing companies. However, they still face many challenges related to the use of these technologies.

Real Big Data:

Consider the amount of data needed to run an operation like Lyft. Lyft, while one-fourth the size of Uber, has approximately 1.5 million drivers and over 25 million users in its network. Lyft needs to maintain updated information on each driver and each user. Lyft needs to maintain data on each vehicle that each driver uses in order to classify it as Lyft, Lyft XL, Lux, Lyft Black, or Lux Black XL. Lyft also offers scooter sharing, so it must track the location of each scooter, as well as its usage and battery charge level. Lyft operates in 644 cities, so it needs updated maps that include road closures, traffic issues, detours, and numerous other constantly changing data points. Additionally, Lyft needs to track a multitude of other information, such as items lost in its vehicles, ratings information, and even song preferences.¹

Obtaining, maintaining, and protecting all of this data is just the beginning. Lyft needs to be able to quickly access the data in order to receive requests, match requests with available drivers, inform the rider of the driver, the drivers information (vehicle description and license plate number), and then stream a moving map of the drivers location to the rider. Simultaneously, Lyft informs the driver of the identity of the requestor and the requestors rating, the passengers pick-up location, and the passengers destination. Then Lyft must stream step-by-step audible directions as well as a moving map to the driver. When the ride is complete, Lyft instantaneously sends a bill to the customer, and provide an opportunity for comment and to provide a tip. Lyft then credits the driver with the payment, and provides the driver with an opportunity to review the rider.

In order to compete successfully with competing taxi services and Uber, each Lyft ride must be executed nearly flawlessly. With so many variables out of Lyfts control, for example, weather, major events, driver and passenger attitudes, road construction, etc., Lyft cannot afford to experience any technical hiccups. With all this in mind, consider that Lyft performs nearly 2 million rides each day.

Technology challenges:

Early on, Uber hosted the infrastructure to support its big data on-premises and experienced challenges because of it. In 2014, Uber data base administrators discovered a data breach that disclosed information involving approximately 50,000 drivers. In 2016, a data breach released personal information on 600,000 Uber drivers and 57 million customers. In 2018, Uber admitted that it failed to provide reasonable security for consumer data and paid a $148 million fine.² Learning from Ubers struggles, computer programmers Logan Green and John Zimmer chose to use a cloud-based infrastructure when they created Lyft in 2012.³ By using a third party database storage provider, Lyft minimized its initial database infrastructure expenses by foregoing the need to install its own database hardware and software. When Uber did switch to a cloud-based infrastructure, it combined services from several smaller cloud service providers, while Lyft chose to utilize the services of cloud computing powerhouse, Amazon Web Services (AWS).⁴ Lyft contracted to pay AWS $8 million per month (nearly $100 million/year!) for its services.⁵ Data from internal applications, the web, social media, and more are combined into large data sets that can by analyzed to help meet organizational needs. For example, analyzing data from numerous public and private databases provides insights related to areas such as weather, parking, traffic incidents, bus and taxi ridership, and much more are helping ride-share companies predict demand for their services more effectively and thereby improve customer and driver experiences.⁶ Analyzing mined data is even being used to help drive innovation. In fact, Lyft released some of their autonomous driving data to sponsor a research competition to help advance self-driving technology.⁷

Continued Challenges:

While the ride sharing industry has found valuable advantages in the use of big data and data mining technology, it is not without its struggles. For example, the cost of this technology is significant, and some question if the benefits are worth the cost.⁸ Both Uber and Lyft have also experienced cybersecurity issues with third parties gaining unauthorized access to company data.⁹ Furthermore, even though they have used big data and data mining to help improve the experience of drivers, the industry is faced with lawsuits¹⁰ and increased regulation¹¹ regarding the treatment of these important contributors to the organizations success.

In order to remain competitive and promote innovation, Lyfts managers will need to find a way to overcome these challenges, and continued use of big data, data mining, and web mining may help them do just that.

What are the most significant challenges facing Lyft? How can big data, data mining, and related technologies help them overcome those challenges?