LIXIL Corporation: Migrating a SAP-based ERP system to Google Cloud to optimize data analysis
About LIXIL Corporation
The LIXIL Corporation was established in 2011 following the integration of five major domestic building material and equipment manufacturers, such as Tostem and INAX, and it has since acquired global brands GROHE and American Standard. The Japan-based company now has over 56,000 employees spread across more than 150 countries worldwide.
Tell us your challenge. We're here to help.
Contact usOptimizing data analysis to achieve higher performance and reducing operating costs following migration to Google Cloud.
Google Cloud results
- Over eight months, LIXIL built three SAP systems, 30 landscapes, and about 120 SAP-related servers on Google Cloud
- Performing high-speed processing and producing detailed analysis for each salesperson with BigQuery
- Linking data spread across the world using Cloud Interconnect and IP/ MPLS services
- Realizing a serverless data pipeline system with Cloud Functions and Workflows
- 400 employees now use the LIXIL Data Platform and this is growing steadily
Democratizing data to boost performance and cut costs
LIXIL Corporation develops and manufactures leading-edge sanitaryware, kitchens, and building materials designed to improve living environments in more than 150 countries worldwide. In May 2021, it launched its future-proof data infrastructure renewal project, the LIXIL Data Platform (LDP). Designed to “democratize data use,” this involved migrating the group’s Integrated Enterprise Resource Planning (ERP) by SAP S4 / HANA across to Google Cloud.
“BigQuery on Google Cloud was exactly what we were looking for to optimize our data. It has overwhelming computing power and scalability, is easy to use, and is cost-effective. We decided it would be much easier to generate synergies if our entire system was moved to Google Cloud."
—Mr. Iwasaki, Senior Vice President, Leader of System Development & Operation, LIXIL CorporationThe LIXIL Group runs on International Financial Reporting Standards and previously relied on various in-house systems. The company operated its ERP via a large-scale on-premises data center that ran on mainframes. Explaining the decision to move from an on-premises solution to the cloud, Mr. Iwasaki, Senior Vice President of LIXIL Corporation, Leader of System Development & Operation, said: "We no longer wanted to be tied to fixed assets, such as mainframes. We wanted to bring the flexible, scalable benefits of cloud technology into LIXIL's business.
“We chose Google Cloud from a number of platforms because we value data first and foremost. Each employee produces hundreds of millions of pieces of data on a monthly basis, but with our old system, our use of this data was limited in some departments, such as marketing. BigQuery on Google Cloud was exactly what we were looking for to optimize our data. It has overwhelming computing power and scalability, is easy to use, and is cost-effective. We decided it would be much easier to generate synergies if our entire system was moved to Google Cloud."
The LDP is a cloud-based data integration platform with a big data core that uses data analysis to enable employees without specialized knowledge to make the most of company data. With the support of Google Cloud, this data is drawn from mission-critical systems such as SAP and the mainframe, web application servers, production equipment, and IoT logs.
The LDP architecture consists of three components: a data lake, which provides an interface for integrating various data into the LDP, a data analysis element, and a data warehouse that allows files to be uploaded to the LDP for analysis. And, in addition to this basic functionality, LIXIL's proprietary in-house web application includes the LDP Data Catalog. "The LDP Data Catalog is a mechanism that allows you to instantly search all the data stored in BigQuery," says Mr. Kurimoto, Digital Technology Leader, Digital Division at LIXIL.
LIXIL is steadily promoting "democratization of data utilization" by utilizing LDP, and the company is particularly pleased that BigQuery has made it possible to perform data analysis that was not possible before. “LIXIL developed guidelines a few years ago for promoting standard sales processes, but the results were not as good as expected. We ended up with large volumes of data that our existing systems didn’t have the processing power to analyze effectively,” says Mr. Kurimoto.
“By utilizing LDP, it is possible to perform high-speed processing and produce detailed analysis for each salesperson. Next, we plan to work on improving specific sales processes based on the results of this analysis.”
“LDP enables data to be linked directly to it from existing data sources worldwide by leveraging Cloud Interconnect and IP/MPLS services in each country. Using Cloud Functions and Cloud Workflow, we were able to realize a serverless data pipeline system in which data linkage is executed in an event-driven manner."
—Mr. Kurimoto, Digital Technology Leader, Digital Division, LIXILCentralizing global data
LIXIL's data sources are scattered around the world in on-premises systems, data centers, clouds, and factories. To aggregate this data into the LDP, it was necessary to create an appropriate network route and a scalable and easy-to-use, general-purpose data linkage mechanism. “LDP enables data to be linked directly to it from existing data sources worldwide by leveraging Cloud Interconnect and IP/MPLS services in each country. Using Cloud Functions and Workflows, we were able to realize a serverless data pipeline system in which data linkage is executed in an event-driven manner," says Mr. Kurimoto.
LDP has also developed its own LDP AIKit that makes it easy to handle machine learning from Google Sheets by utilizing BigQuery ML. “This is a proprietary Google Sheets add-on developed in Apps Script. It leverages the capabilities of BigQuery ML to support regression, classification, and clustering of basic machine learning patterns. It can be used for various purposes, such as inventory/demand forecasting, and advertising cost simulation. BigQuery ML doesn’t require any specialized tuning to obtain forecasted values, which is really helpful. I would like to expand its use cases for the company in the future," says Mr. Kurimoto.
"Next, we plan to utilize the data collected by ERP (SAP S4 / HANA) in BigQuery. And we plan to move other areas of the business to Google Cloud in the future so that data can be quickly used to add value.”
—Mr. Iwasaki, Senior Vice President, Leader of System Development & Operation, LIXIL CorporationCreating a data-linked architecture
BigQuery's performance also helps identify the impact of discontinuing certain LIXIL products. Mr. Kurimoto says: "It's easy to list what parts are used in a product, but vice versa, we're dealing with finding out how many products use a part. Due to the huge number of products, the system is heavily loaded. LIXIL needs to perform this process once every few months, but due to the processing capacity of the existing system, a lot of person hours are required each time. By migrating these processing infrastructures to LDP, the processing time was reduced to about one-fifth of the conventional level, and the person hours of the product development department were significantly reduced."
Mr. Furusawa, who works in the System Infrastructure Department of LIXIL’s Digital Division, handled the migration in May 2020, which he says went very smoothly. "With a standard public cloud, you often get disruptions due to maintenance, but with Google Cloud, you can use the live migration function to run the migration without interruption. I’m happy to run our core system on Google Cloud. And it also lets you build one segment in multiple zones, which is great for network configuration.
“Being able to use the same subnet across zones makes everything easier to design and build. We were able to build a database that was replicated for each zone synchronously."
With help from Google Premium Support and technical account management, the LIXIL team built three SAP systems, more than 30 landscapes, and about 120 SAP-related servers over the course of eight months. By May 2021, the SAP accounting production environment was operational and settling accounts.
Google Cloud enables the LDP to optimize data for everyone, from accountants to engineers, but it is still a work in progress. "Next, we plan to utilize the data collected by ERP (SAP S4 / HANA) in BigQuery. And we plan to move other areas of the business to Google Cloud in the future so that data can be quickly used to add value,” says Mr. Iwasaki. Mr. Furusawa adds: "We also plan to integrate the system with overseas group companies, such as GROHE and American Standard, and to that end, Google Cloud's global infrastructure is indispensable.”
Currently, there are about 500 LDP users and this is growing monthly and will rapidly accelerate in the future. “The immediate goal is to create a situation where 3,000 people, about 10% of all employees, use the LDP in some way every month. I want to create a fully data-driven environment," concludes Mr. Kurimoto.
Tell us your challenge. We're here to help.
Contact us
About LIXIL Corporation
The LIXIL Corporation was established in 2011 following the integration of five major domestic building material and equipment manufacturers, such as Tostem and INAX, and it has since acquired global brands GROHE and American Standard. The Japan-based company now has over 56,000 employees spread across more than 150 countries worldwide.