Innovation on SAP HANA Platform – Advanced Personalized Recommendation
I recently wrote about blog how SAP HANA helped resolve a serious inventory reallocation problem at Hong Xiang. But SAP HANA – supported by and optimized for Intel’s latest processor technology – is being used in plenty of situations where there aren’t any major crises per se. Instead, these enterprises are simply looking for new ways to extract better insights and value from their existing technology investments.
Nippon Paint China is a great example of what I mean. Founded in 1881 and entering the Chinese market in 1992, it’s among the oldest paint manufacturers in the world, and relies on a large number of SAP® solutions to drive its business. They’re in a highly competitive market that’s jammed with a glut of low-to-mid level products, so the company continually seeks innovative ways to differentiate itself from its many rivals. To give you an idea of how they strive to stand out, in 2011 they developed the iColor website – an interactive platform tailored to foster communication among three essential communities: customers, renovation designers, and construction companies.
iColor showcases outstanding examples of art work, decoration, and construction, along with offering case studies and free downloads of specialized design tools. Customers are welcome to share pictures, provide comments, and establish immediate contact with vendors to assist with their upcoming projects. Take a look at it for yourself – it’s very impressive (iColor website).
Although iColor was a hit from the beginning, Nippon Paint China realized that it held a tremendous amount of untapped potential. In particular, they looked for ways to conduct correlated cross-system analytics that fused structured data produced by SAP applications and stored in SAP HANA along with semi-structured/unstructured data such as website log data (containing huge quantities of invaluable iColor site usage details, such as browser, user region, color and design preferences, to name just a few) that were flooding into Hadoop. The resulting knowledge about market trends and customer demands would be integral in their never-ending work to offer the most compelling products, maintain optimal inventory levels, and ultimately outpace competitors.
As I alluded to earlier, Nippon Paint China was already using lots of SAP solutions (like SAP Enterprise Resource Planning (ERP), SAP Customer Relationship Management (CRM), and SAP NetWeaver Business Warehouse (BW)). In addition, they’d already migrated BW to SAP HANA to use its real-time, in-memory processing to help speed up key queries and reports. Given the great results they’d been having with SAP technologies, it was no surprise that they looked to SAP HANA as the cornerstone of their ambitious undertaking.
The new system populates SAP HANA with iColor website log information that’s been stored in Hadoop, along with customer and product details extracted from SAP enterprise applications such as ERP and CRM. Hadoop is also tasked with conducting some preprocessing on raw website visitor data (like basic log aggregation) to come up with preliminary summary information. SAP Smart Data Access returns these previews to the application, while SAP Data Services is responsible for loading larger amounts of raw Hadoop data into SAP HANA for deeper processing that fully takes advantage of the advanced vector extensions (AVX) offered by the Intel® Xeon® processor family.
The new unified analytics solution has delivered on its promise of breaking down the barriers that blocked users from fully comprehending the interplay between enterprise information and unstructured website visit data. As a result, Nippon Paint China and its community of customers, designers, and construction companies have each realized significant benefits.
Business users can now ask complex questions that federate log and SAP application data to return comprehensive results. For example, it’s possible to determine color and painting style preferences for specific users simply by associating their website interactions and corresponding entries from the SAP enterprise applications. This knowledge helps product designers confidently develop solutions that will be in tune with customer requests and market momentum. Simultaneously, with much sharper views of upcoming demand patterns, the company’s marketing team can respond more quickly with accurate messaging and skillfully targeted campaigns, and the resulting stronger interactions with clients have helped boost website traffic and associated revenue. Best of all, customers are presented with better, more precisely personalized recommendations to help them achieve their decoration goals.