Elevating the home experience with customer-centric innovation
LG uses the Databricks Platform to power smarter home appliances with secure access to customer data
Reduction in existing total costs
Decrease in customer data unification costs
Reduction in processing times
LG Electronics, a leader in global home appliances, has addressed many technological challenges over the years to provide better value to consumers. One example is LG’s recently launched “UP Home Appliances” service, which can continuously upgrade the functions of home appliances with self-developed artificial intelligence (AI) chips. To start this initiative, the home appliance company focused on the efficient use of their customer data and established a customer-centric data system. LG faced many difficulties along the way, such as large-capacity data management, information security and system expansion. But with the Databricks Data Intelligence Platform, they were able to streamline customer data management — reducing both the time to unify customer purchasing data and overall costs by 50%.
Data silos impeded an accurate understanding of LG customers
LG Electronics is known for their innovative home appliances worldwide, with sales of their various products reaching approximately 100 million units every year. This volume generates logs from product devices that amount to several terabytes of data per day, which includes customer feedback, billing information and more. Efficient management of this data (e.g., product searches and postpurchase experiences) was essential for an accurate understanding of LG’s customers, but the existing environment didn’t allow for efficient collection, analysis and delivery. Hyeojun Moon, Team Leader, CDO | LG Electronics, CDO, Data Platform Team, explained, “There was great difficulty in understanding the meaning of database tables that existed in individual systems of various business departments. We were also struggling with updating the initial schema, which interrupted users from using the data efficiently.”
Due to complex usage procedures and multiple restrictions, there were numerous other obstacles, including limited availability of data, difficulty in sharing use cases and similar data being delivered repeatedly. To address the shortcomings of the existing data system and seamlessly connect all employees to the customer data, LG developed an independent customer-centric data system, Intellytics Customer 360 (IC360). However, the process of accurately matching and linking individual customer data records, known as “customer data ID pinning,” proved to be challenging and complex. This process involves connecting data points from different sources to ensure they all refer to the same individual, which was not easily executable with LG’s existing systems. Even when using an external customer data platform (CDP) that connected individual customer databases with tens of millions of domestic customers, there was a significant time investment.
Because an external software as a service (SaaS) company provided the CDP solution, information security issues could occur with customer data leakage. Adding fuel to the fire, the CDP service also required more than 40,000 executives and employees across the company to utilize the data system for customer analysis, regardless of their department. When thousands of data user queries occurred simultaneously, the performance of the CDP system significantly deteriorated. More than 10 simultaneous queries were not feasible, and other problems — such as user queries failing to be queued — led to delays in important system query work. Together, these problems prevented IC360 from being successfully released.
Scaling customer data security with a unified platform
LG Electronics introduced the Databricks Data Intelligence Platform to address challenges in their customer data system, including handling large data volumes, ensuring information security and achieving system scalability. By leveraging the Databricks Platform, LG was able to more efficiently process customer data using ID pinning and successfully launched IC360. The scalability of this data system was enhanced through Databricks’ automated cluster management and Delta Lake. An optimized storage layer, Delta Lake enabled LG to take advantage of ACID transactions, scalable metadata handling and unified streaming and batch data processing within the same platform.
Through automated cluster management, LG benefited from flexible resource operations, scaling up compute during peak times and reducing it during off-peak times. For example, during an internally hosted hackathon related to the customer data index, Databricks ensured stable compute resources even with high usage. Moon explained, “In managing various workflows, we utilized Databricks’ software development kit — enabling us to add, remove and stop thousands of API-based jobs — to make efficient data management and access possible for all roles.” This streamlined process allowed data users to bypass learning different data languages, like Python or SQL, while data scientists could easily view and utilize customer data through the UI and connect data via SQL queries on the back end.
Data security issues were also resolved with the Databricks Data Intelligence Platform. Connecting to Databricks via a private link improved data safety by blocking external access to the control plane, allowing only authorized users. Databricks SQL, an intelligence data warehouse with a high-performance query engine, then enabled efficient SQL queries on large datasets, integrating seamlessly with other Databricks services and providing a user-friendly interface for data analysis. LG now confidently maintains secure customer data to streamline operations, enhance scalability and ensure data security.
Reducing work time and costs with powerful computing
Since LG Electronics introduced the Databricks Platform, they have successfully built a CDP that manages their vast amount of customer data. Previously, it took about 10 hours to pin all data on an individual basis and process it into personal attribute data and behavioral data. By using Databricks, the time has been reduced from 10 hours to three hours — a 70% improvement — significantly improving operational efficiency. Additionally, data logic processing costs have been reduced by 30% due to more efficient compute consumption and automated workflows.
In addition to launching IC360, Databricks also enabled flexible resource management, optimizing data system management costs. Compared to the 24-hour usage cost of the preexisting SaaS-type services, nonpeak costs were decreased using Databricks. Flexible operation during peak times resulted in up to a 50% reduction in total data system management and compute costs.
Moon concluded, “LG Electronics is committed to harnessing the power of our new technological environment and CDP utilization, which will serve as an important stepping stone to developing customer-centric innovations today and into the future.” Moving forward, LG Electronics plans to apply AI to customer data systems through their continued collaboration with Databricks, enabling various data users to obtain the information they want and maximize data utilization through text-based SQL. By continuously integrating cutting-edge AI and data technologies into their workflows, LG Electronics is not just optimizing their operations but also paving the way for future advancements that other companies can emulate.
