Recommendation Systems

Our research in Recommendation Systems (RecSys) is empowering businesses to deliver smarter, more personalized customer experiences. By harnessing data-driven insights, we are developing AI solutions that enhance decision-making, optimize engagement, and unlock new growth opportunities across diverse industries.

Bidding Optimization for DSP

Bidding Optimization for DSP is a project focused on developing a fast and efficient bidding system for demand-side platforms (DSP). Using Machine Learning, the system maximizes KPIs and optimizes ad purchases within a 10ms latency. Key achievements include improving ROAS across major campaigns, handling millions of bids per second, and optimizing costs. The project also explored Dynamic Creative Optimization (DCO) and expanded to global SSPs.

Corca and IGAWorks: Redefining DSP Ad Optimization

Project Objective
The project’s core objective was to engineer a predictive algorithm that could accurately model user responses to Bid Requests, setting the stage for strategic bid price determination.
This ensures advertisers reach target audiences at optimal moments, maximizing budget utilization and ROI. Additionally, it was crafted to support rapid and cost-effective bidding within Real-Time Bidding (RTB) environments.

Key Features and Technological Advancements

Real-Time AI-Powered Bid Optimization: Leveraging advanced AI, the algorithm dynamically interprets user behavior in real-time to create precisely tailored bidding strategies. By evaluating user response rates (CTR), it calculates bid prices with precision, amplifying Ad effectiveness.

Multi-SSP Compatibility: Custom-tailored algorithms for diverse SSP platforms allow for finely tuned bid strategies that are uniquely aligned with each platform’s structural nuances, driving optimal ad performance.

Ultra-Responsive Processing: Bid determinations occur within a 100-150 millisecond range, facilitating rapid and accurate real-time bidding in high-speed RTB environments.

Resource Optimization: Designed with a meticulously balanced architecture, the infrastructure achieves high processing speeds with minimal memory footprint, achieving peak operational efficiency.

Technical Challenges and Solutions
This project overcame formidable challenges due to Apple’s iOS ATT (App Tracking Transparency) policy, which restricts cookie-based tracking. To address this, Corca implemented sophisticated user-matching technology and developed platform-specific bidding strategies, securing optimal ad performance across a wide array of channels.

Data-Driven Strategy Dvelopment
Corca synthesizes data from numerous touchpoints—including Bid, Impression, Conversion, and Purchase metrics—to create rigorously optimized ad strategies. This comprehensive data-driven approach empowers advertisers to make strategic, evidence-based decisions, maximizing the impact and efficiency of their advertising campaigns.

Future Research and Expansion Horizons
Looking ahead, Corca aims to build upon its current bid optimization capabilities by developing Audience Prediction AI and Creative Recommendation Algorithms to further elevate targeting precision and ad impact.

Audience Prediction AI: Through in-depth user behavior analysis, this technology will enable advanced audience segmentation, significantly enhancing targeting accuracy.
Creative Recommendation Algorithms: These algorithms will deliver optimized creative recommendations to users, amplifying ad engagement and effectiveness.

Corca’s DSP solution leverages state-of-the-art AI to open transformative opportunities for advertisers aiming to maximize ROI. This solution transcends basic automation, positioning itself as a cornerstone innovation within the digital advertising landscape. By continually enhancing advertising efficiency and delivering value across the ecosystem, Corca is setting a new benchmark in digital ad technology.

Conclusion
The collaboration between Corca and IGAWorks has transformed digital advertising through AI-powered bidding algorithms, enabling smarter and faster campaign decisions. By focusing on data and real-time user behavior, this innovation has made advertising more relevant and efficient, benefiting both advertisers and users. This success has reinforced Corca's commitment to advancing AI technology in advertising, positioning the company to lead future innovations in the industry.

Promotion Optimization Project

Promotion Optimization Project focuses on using AI to enhance promotion planning and operations. By developing a sales prediction model, the project automates event planning to maximize revenue while ensuring the target profit is exceeded. It suggests the optimal products, discount rates, and event strategies to achieve the highest sales. The AI-driven events have consistently maintained an average revenue increase of 104% compared to traditional event methods.

Promotion Optimization Project: Data-Driven Personalized Customer Experience Research

Project Overview

Corca's AI-Based Promotion Optimization Project leveraged advanced AI technology to enhance promotion planning and execution, delivering tangible results. By analyzing customer data, the project automatically determined optimal discount rates, product selections, and timing, successfully providing personalized benefits that achieved both customer satisfaction and revenue growth.

Research Objectives

Data-Driven Promotion Strategy:
AI was used to analyze customer purchasing patterns and preferences, allowing for personalized event proposals that increased customer engagement and event efficiency.
Sales and Profit Growth:
Monthly sales and profits in key product categories increased by over 5%, supporting sustainable growth.
SKU Efficiency Enhancement:
Based on sales data and customer response, SKU count was streamlined by up to 60%, focusing on core products.

Key Research Areas

AI-Powered Promotion Optimization:
Customer data was analyzed by AI models to optimize promotion timing, product selection, and discount rates, maximizing the effectiveness of each promotion.
Predictive Performance Modeling:
AI predictive models assessed the impact of discount rates and price adjustments on sales and profits, supporting data-driven decision-making. This enabled the development of optimal promotion strategies tailored to seasonal trends and store-specific needs.

Future Research and Expansion Potential

Building on the success of this promotion optimization project, Corca plans to further leverage AI and customer behavior data to advance customer segmentation models and promotion performance forecasting systems. This approach aims to create systems that deliver increasingly personalized experiences at each customer touchpoint.

Conclusion

Through its AI-BasedPromotion Optimization Project, Corca has successfully implemented a data-driven personalized event strategy based on customer behavior and preference analysis. The company has transitioned from intuition-based approaches to evidence-based event planning by implementing real-time AI solutions, introducing predictive models that enable precise targeting and optimized event timing. Through SKU optimization, Corca achieved innovative results by reducing SKU complexity by 60% while maintaining variety and customer satisfaction, and these data-driven strategies have significantly contributed to improving customer experience and operational efficiency. Moving forward, Corca will build upon this success to further advance its AI technology, enhance customer segmentation and performance forecasting, ultimately leading retail innovation by creating meaningful and strategic interactions at every customer touchpoint.

AlphaDom Project

In the AlphaDom Project, we partnered with Cociety to develop cutting-edge AI solutions for the 'Worker Shop' wine store. The project featured personalized wine recommendations powered by AI, customized receipt sketches, and auto-generated messaging, showcasing the versatility and innovation of retail media technology. This project serves as a key milestone in creating an Omni-channel Retail Media Testbed, effectively blending online and offline experiences, and demonstrating significant potential for scaling in the retail sector.

AlphaDom: The Future of Omni-Channel Retail Media

Where AI Transforms Shopping into Smart Experiences

Project Overview

The AlphaDom Project, a collaborative initiative with Cociety, explores innovative potential in retail media* solutions. We established an advanced testbed at the Worker Shop, an omnichannel wine store located in Pangyo Tech One building. The project's primary objective was to maximize retail media effectiveness and enhance consumer interactions through seamless integration of online and offline channels.

(* What is Retail Media? : Retail media turns shopping spaces into dynamic media channels. From digital displays to AI-powered personalized content, it creates meaningful connections with customers at every touchpoint in their shopping journey. This emerging channel is reshaping how brands engage with customers at the moment of purchase.)

Convergence of Retail Media and AI Technology

We leveraged AI-powered technology as the driving force to organically connect digital and physical channels. By combining this with solutions that enrich the customer journey, we designed a new form of retail engagement that spans the entire purchasing process.

Retail Media Innovation: AI-Powered Wine Recommendation Model

The wine recommendation model, the cornerstone of this project, demonstrates the transformative potential of modern retail media. While traditional approaches typically offer one-way information delivery, our AI-based system provides personalized product suggestions through real-time customer interaction.

This sophisticated model analyzes multiple factors including customer preferences, visit purpose, food pairing requirements, and ambiance considerations to recommend the most suitable wine. In the complex world of wine selection, our AI effectively digitizes sommelier expertise, making it accessible to everyone. The system considers not only taste preferences, aroma profiles, and price points but also food compatibility and occasion-specific requirements to deliver optimal suggestions.

This personalized approach represents a revolutionary advancement in retail media, transcending traditional product displays and one-way advertising to enable individualized, two-way communication. By combining the tangible product experience of physical retail with data-driven personalization, it exemplifies the future of retail innovation.

Enhancing Customer Engagement Through AI

Building upon the wine recommendation system, we employed additional AI technology to transform routine transactions into memorable interactions. The AI Sketch Model converts customer-uploaded photos into artistic renderings, transforming traditional purchase documentation into personalized digital memorabilia. Additionally, our AI Text Generation Model creates customized messages based on user-selected keywords, adding meaningful context to each wine purchase. These technologies elevate the standard transaction into a unique, individualized engagement for each customer.

Conclusion

The AlphaDom Project has demonstrated both the innovative potential of retail media and AI technology's capacity to create more personalized and meaningful commercial interactions. Particularly in the wine sector, where individual preferences and contexts play crucial roles, we've proven the capabilities of AI-powered solutions. Combined with experience-enhancing elements, this project opens new possibilities for the future of retail.

This initiative serves as a significant milestone in advancing retail technology and demonstrating how innovation can enrich human interactions. It shows that the integration of AI in retail can go beyond efficiency improvements to create more meaningful and personalized customer engagements.