Marketing to mobile and social customers

Customer experience management marketing based on behavior analysis in Korea

By Soonsung Bae, Senior Manager of Platform Business, GS Home Shopping

GS Home Shopping is one of the largest TV home shopping companies in the world, as well as the leading multimedia retailer in Korea. The company provides shopping services via TV, Internet, catalogs and new media platforms, including digital cable TV, IPTV (Internet protocol television), smartphones and tablets. In 1995, it became the first company to introduce home shopping to Korean consumers, and it receives 800,000 unique visitors to its online shopping sites on a daily basis. Also, the company’s mail-order catalogs are delivered to 2 million households every month.

Using the channel data, customer data and logic, we have created a recommendation service, which has been imparting a meaningful impact on sales. Moreover, we have ascertained that as more customer experiences are created, the customer responses improve.

Retail business trends

GS Home Shopping has consistently analyzed customer behavior across channels. In 2013, the retailer found that customer behaviors had noticeably changed on Internet and mobile channels. In particular, customers were rapidly moving over to mobile and social shopping.

Characterized by a small number of items yet high sales volumes, social shopping is a business model optimized for mobile. With fewer products as compared to other shopping channels, more targeted curation for each product can be achieved. Social selling ensures competitive pricing through high volume, predictable sales.

Meanwhile, mobile shopping is becoming increasingly personalized, and its sales window is becoming smaller. Also, a large number of mobile shoppers participate in mobile shopping via apps.

Customer experience management

Customer usage patterns became complicated due to mobile device usage and changes in customer characteristics. It seemed that preferred customers were moving away from the PC; however, analysis showed that usage pattern had changed to shopping via mobile devices instead of PCs.

With these changes in customer experience underway, GS Home Shopping is implementing activities, such as:

  • Integrating the customer experience centered around mobile.
  • Strengthening personalized recommendation and scientific store management.
  • Promoting integrated channel push marketing based on customer segmentation.

Customer behavior analysis platform

GS Home Shopping’s existing platform for customer behavior analysis is comprised of a data warehouse where data such as order history, customer information and product information reside along with web log data.

Traditionally GS Home Shopping used this data for marketing activities such as channel-specific customer segmentation and targeting centered around product visit, purchase experience and push channel-specific marketing. However, this type of push marketing was relatively simple, and used individual- and channel specific target marketing techniques for the customer experience.

Push marketing now entails loading and consolidating data from all customer experiences, refining customer targeting and optimizing communication channels. The data for determining which channels — such as mobile, push SMS or email — evoke a response is now managed in a separate system. The response is communicated via the marketing channel that is most optimized for the customer.

In the past, pull marketing for PC and mobile was operated independently. In addition, stores sold a large number of items at small sales volumes, and the customer experience data was most relied upon for things such as product display. But from now on, mobile and PC customer experiences will be integrated. There’s an emphasis on carrying a small number of products, providing individualized recommendations and conducting A/B testing.

To support these changes, GS Home Shopping has built a big data platform that uses much of the existing platform structure but now centers around the customer activity log. The platform supports big data, recommendations and real-time analytics.

Based on this, we have segmented integrated channel customers, optimized customer-specific preferred channels, and personalized content and offers for push marketing. On the pull side, we have established a foundation for integrating customer experience in real time, refining personalized recommendations, and scientifically operating stores.

Big data platform and recommendation engine

The big data platform is comprised of:

  • An area for collecting customer data from a variety of devices such as PC and mobile.
  • An area for storing collected data in Hadoop.
  • An area for analyzing how to best arrange and utilize the data.
  • An area for real-time data processing in order to enhance the customer experience.
  • An area for creating recommendation algorithms. The data stored in Hadoop is loaded into the database and used for the recommendation service.

Customers receive product recommendations on the main page and separate pages on the PC. Recently, this service has been extended to mobile shopping, and even when channels other than mobile or PC are visited, customer data is integrated so that the visitor can be seen as one customer. Using the channel data, customer data and logic, we have created a recommendation service, which has been imparting a meaningful impact on sales. Moreover, we have ascertained that as more customer experiences are created, the customer responses improve.

Using SAS®

GS Home Shopping needed to manage the entire purchasing journey of a customer in order to understand the customer experience, and respond immediately to customer behaviors in accordance with changing trends.

To that end, SAS has been introduced for the design and implementation of an overall campaign architecture, along with in-memory visualization capabilities for market sensing, customer insight enhancement and new marketing opportunities.

In addition, we previously operated one recommendation model through a Hadoop-based recommendation engine and an in-house product recommendation method. To improve upon this, GS Home Shopping has been testing a variety of recommendation models using SAS for data mining. Without creating a separate data set (or data mart) for SAS within Hadoop, SAS is used to access the Hadoop (Impala) module directly.

GS Home Shopping will continue to carry out real-time integration of the customer experience, while developing and expanding algorithms for personalized recommendations, and optimizing an infrastructure to utilize big data and recommendation services.


Read More

Get More Insights

Want more Insights from SAS? Subscribe to our Insights newsletter. Or check back often to get more insights on the topics you care about, including analytics, big data, data management, marketing, and risk & fraud.