CJ Olive Young

CJ Olive Young scales omnichannel retail recommendation and personalization for customers on Google Cloud Spanner

Results on Google Cloud
  • Boosts combined peak transaction capacity by 160x

  • Unifies omni channel data hub to power recommendations and personalization with zero-downtime synchronization

  • Cuts database query response latency by 90%

  • Eliminates downtime during quarterly "Olive Young Sale" peaks

CJ Olive Young unified omnichannel retail data with Cloud Spanner, boosting throughput 160x and cutting latency 90% at peak.

Evolving from legacy systems to support omnichannel shopping at scale

CJ Olive Young is South Korea’s leading health and beauty (H&B) retailer, serving tens of millions of customers across an extensive network of physical stores and a massive digital commerce app. To maintain its position as a global K-beauty lifestyle powerhouse, the company connects online shopping with in-store visits to deliver a continuous, member-centric retail experience.

However, this massive popularity created immense technical pressure during major promotional events, such as the quarterly "Olive Young Sale (All-Young Sale)". During these peak campaign periods, shopping traffic instantly surged by tens of times compared to normal shopping days. Millions of concurrent members rushed to the Olive Young app to check real-time order history, track deliveries, write product reviews, and verify loyalty tier benefits.

The company’s legacy architecture, a monolithic system running on traditional on-premises MySQL databases, reached its physical scaling limit. Traditional relational databases (RDBMS) can only scale "vertically" by adding more CPU and memory to a single server machine. When peak traffic hit, CPU and disk utilization skyrocketed to 100%, causing query bottlenecks, application slowdowns, and customer friction on the app’s "My Page" order screen.

During our major sales events, traffic explodes by tens of times. Cloud Spanner provided the infinite horizontal scalability we needed to handle massive peak transactions without compromising strict relational data integrity or uptime.

Kwangjin Lee

Team Lead / Data Platform Team, CJ Olive Young

Furthermore, member order data was fragmented across separate systems, making it difficult to deliver unified cross-channel recommendations or recalculate customer tier statuses quickly. CJ Olive Young needed a modern database capable of horizontal scaling (adding multiple machines seamlessly) while strictly preserving relational ACID data integrity for mission-critical financial orders. After evaluating legacy databases and NoSQL alternatives, CJ Olive Young selected Cloud Spanner, Google Cloud’s globally distributed database that offers unlimited horizontal scale-out with strong transactional consistency.

Building a unified omnichannel data hub with zero-downtime synchronization

To eliminate data fragmentation and guarantee low-latency customer interactions, CJ Olive Young built a Data Service Platform on Cloud Spanner. The platform consolidates data scattered across individual services into a standardized data model, then delivers it to each service in the way that service requires—in real time or near-real time, and in the form best suited to its workload. Sitting at the center of the customer backend on Google Cloud, the Data Service Platform—with Cloud Spanner as its storage foundation—acts as the single source of truth for both online and offline shopping activity.

The platform directly powers high-demand personalized app features in real time:

  • Unified purchase history: Instant order lookups for in-app review writing on "My Page"
  • Rolling personalization: Frequent product recommendation algorithms based on a rolling one-year purchase history
  • Automated member tiers: Dynamic customer loyalty tier calculations calculated over six months of transaction values
  • Omnichannel store context: Tracking offline store visit history over the preceding three months and caching active shopping carts
Thanks to Cloud Spanner's auto-scaling, our initial bulk data migration completed significantly faster than anticipated. Schema updates were applied seamlessly in the background without database locks or user interruption.

Beomjun Park

Data Engineer / Data Platform Team, CJ Olive Young

Migrating historical transaction data without disrupting active shoppers required a rigorous, non-blocking synchronization strategy. First, the engineering team extracted legacy snapshots and leveraged Spanner's elastic auto-scaling to parallel-load initial bulk data, cutting load time by 70% compared with the equivalent process on the legacy system. Next, the team deployed real-time Change Data Capture (CDC) pipelines to replicate ongoing transactional updates continuously. To ensure 100% data reliability, CJ Olive Young developed an in-house real-time validation program that verified zero discrepancies between legacy records and Spanner per member ID.

Crucially, schema adjustments (DDL updates) were executed dynamically in the background without database locks or downtime, allowing engineers to transition live member services smoothly with zero service interruption.

Technical architecture diagram featuring Google Cloud Spanner serving pipeline for CJ Olive Young

160x transaction capacity boost and a foundation for global AI commerce

Migrating to Cloud Spanner delivered immediate business stability and massive technical performance improvements. During the high-traffic June Olive Young Sale, the platform processed peak traffic without a single second of downtime or performance degradation, completely defending against potential revenue loss from infrastructure bottlenecks.

The quantitative gains transformed the customer shopping experience:

  • 160x throughput surge: Combined online and offline transaction processing significantly increased beyond legacy constraints, dramatically expanding standalone offline capacity
  • 90% latency reduction: Query response latency plummeted to near-instantaneous levels
  • 100% data consistency: Real-time synchronization unified online and in-store purchase records under unified member IDs
Automating database scaling and schema updates has removed the heavy operational burden of manual DB management, giving our engineering team the peace of mind to focus on creating new customer value.

Junseo Kim

Data Engineer / Data Platform Team, CJ Olive Young

Operationally, the transition eliminated the psychological stress of manual infrastructure management. Database scaling is now managed with simple button clicks, and automated maintenance relieves engineers from high-risk manual database tuning, dramatically boosting developer morale and productivity.

Looking ahead, Cloud Spanner provides the foundational substrate for CJ Olive Young’s next stage of global growth. The company is exploring the use of Spanner's native multi-region active-active replication to expand K-beauty retail worldwide without re-architecting data systems. Additionally, CJ Olive Young is considering connecting Spanner Change Streams directly to Gemini platform and Gemini pipelines. This will enable real-time intelligent commerce, reflecting a user's in-store or online purchase across all personalized recommendation and search algorithms within one second.

CJ OLIVE YOUNG is South Korea’s leading health and beauty omnichannel retailer, offering curated lifestyle and cosmetic products across digital platforms and physical stores worldwide.

Industry: Retail

Location: South Korea

Products: Cloud Spanner, Google Cloud

  • Google Cloud プロダクト
  • 100 種類を超えるプロダクトをご用意しています。新規のお客様には、ワークロードの実行、テスト、デプロイができる無料クレジット $300 分を差し上げます。また、すべてのお客様に 25 以上のプロダクトを無料でご利用いただけます(毎月の使用量上限があります)。
Google Cloud