Cox Automotive is a leading provider of vehicle remarketing services and digital marketing and software solutions for the automotive industry – primarily serving consumers, dealers, and manufacturers. Cox Automotive unites more than 20 brands in this space, providing an end-to-end solution to transform the way people buy, sell, and own cars.
In 2014, Cox Enterprises established Cox Automotive to bring its wholesale and retail automotive businesses under a single operational entity. In doing so, Cox aimed to usher in new levels of efficiency and innovation by delivering new customer and consumer experiences through greater data insight.
The challenge was that its existing data environments were not scalable or flexible enough to support the new enterprise data strategy or the rapid data ingest needed to achieve the company’s goals. For instance, the legacy environment couldn’t allow data schemas to be set up on the fly, incorporating new data as it flows into the system and giving BI and analytics users access to that data in real time.
Also, particularly leading up to the 2014 Super Bowl, Cox needed to be able to provide data and analysis on website activity as a result of TV ads in minutes or hours, in order to allow automotive to make on-the-fly adjustments. This would have been impossible using legacy technology.
As Cox Automotive was formed, its executives sought to implement an enterprise data strategy that would enable the company to ingest, enrich, blend, and expose rich automotive data from across its brands. Cox Automotive saw Apache Hadoop as the ideal foundation to launch the company into its next era because of Hadoop’s capability to store an abundant amount of data from multiple sources and allow for much faster analysis, and decided to adopt Apache Spark in particular – as opposed to using Apache Hive – to ensure the new environment could meet the organization’s need for real-time performance. They selected Cloudera’s platform based on the company’s vision, proven experience in the marketplace, performance, and manageability.
Cox Automotive has created five enterprise data hubs – one for each of its assets: audience, vehicles, customers, products and users - enabling better management of diverse data sets from across its brands. The EDHs total more than 1 PB and incorporate 200 million rows of data daily. Cloudera provides the discovery and analytics environment to enable common access, storage and processing of the data.
Before, analytics following Super Bowl ads required a 10-hour batch processing job that would run overnight. The legacy environment couldn’t scale to accommodate growing data volumes, increased granularity or the faster processing times required to satisfy business and client demands.
In 2015, analytics were provided hourly for the Super Bowl, and will be available in near real time in 2016. Further, the insights provided are much richer, extracted from a bigger and more granular dataset. Cox can provide insights to its clients – auto manufacturers -- that they can't find anywhere else; looking at the web traffic for a particular make and model before a Super Bowl ad airs, and comparing it to the traffic directly afterwards; reporting on how much lift can be attributed directly to the Super Bowl ad and how that compares to other Super Bowl ads.
By incorporating Spark, Cox is now building executive dashboards that will showcase core metrics in real time, allowing Cox and its clients to make changes that improve behaviors and results.
With Cloudera, the team has achieved its goal of delivering new, better customer experiences through greater data insight, while improving operational efficiencies.
For example, Cox greatly improved its speed of analysis for car ad campaigns that were ran during the Super Bowl. Previously, the company couldn’t understand how the Super Bowl ads impacted online searches and other website activity until the following day, meaning the website’s interactions with consumers had to be anticipated in advance and analyzed posthumously to improve next year’s placements – proving practically useless, as many variables impacting the car shopping experience can change over the course of one year.
With its new enterprise data hub environment, Cox was able to get hourly updates during the 2015 Super Bowl, enabling it to make recommendations to auto makers and implement changes to the website based on activity driven specifically by that evening’s commercials. Better yet, for the 2016 Super Bowl, Cox will be able to deliver real-time updates, which would have been impossible to even consider in the past due to infrastructure restrictions.
The Spark solution, in particular, literally shows the time lapse between when a commercial airs to when the audience starts searching on the site.
While improving its products and services through better, faster analytics, Cox has reduced the total cost of ownership of its data environment by 50% per terabyte (TB).
Cox Automotive uses Cloudera Enterprise, Data Hub Edition for its Hadoop platform. Additional Hadoop components used include Apache Hive, Apache Hue, Apache Kafka, Apache Spark Streaming, Cloudera Manager, and Impala. Cox has also integrated visualization and analytics tools Datameer and Platfora into its EDH environment to help analysts better understand the company’s data.
The biggest game-changer for Cox is that its EDH environment is helping to provide auto dealers with new insights. One example is on the inventory management side – Cox can look at the cars for sale in the market, understand the consumer demand and easily see and report if there is too much supply. The company can now also provide pricing and sales recommendations to help dealers get the best value from a particular car in the quickest amount of time.
Cox Automotive analysts, executives and IT staff are now thinking and using the data generated in the new environment in previously inconceivable ways. Cox Automotive aims to bring products and analytics to market faster than anyone and with the help of Cloudera, is delivering on that mission.
