B2B · UX research · UI revamp · Design strategy
Company
Coupang Play
Timeline
Jan 2024 - Jul 2024
In H1 2024, Coupang Play needed to manage 30,000 titles and 500,000 files by July to launch individually subscribable content channels. The CMS, built without a design-driven approach, had become increasingly complex, creating workflow inefficiencies.
To simplify workflow and reduce manual effort, I designed Magic Studio, an AI-powered poster generation tool, and CMS Project Sheet, a web-based spreadsheet replacing manual Excel updates.
These tools streamlined operations, enabling efficient large-scale content preparation.
CMS product had been developed without a formal design process, leading to inconsistency in design across the projects. Since the operation team got bigger and was working on 3-4 different ongoing projects, aligning design vision and outcomes across the team became essential.
To address this, I initiated to tackle three key areas:
Setting design principles
I led brainstorming sessions with the design team to define Play Operations' core design principles. We distilled these into three guiding words: Simplicity, Innovation, and Productivity, ensuring consistency across projects and alignment with business goals.
Creating a design system
As the front-end team began adopting the 'Tremor' open-source library, I observed low usage in design files and inconsistency from three different libraries. To ensure consistency and scalability, I adapted Tremor for CMS needs, updating components for production and dark mode compatibility. I established bi-weekly sessions with developers to align efforts and track adoption across teams.
Setting Operation Design KPIs aligned with the 2024 H1 business goal
I defined three key design focus areas: Workflow Improvement, Supporting New Business, and Internal Communication. Success was measured through Internal Team Efficiency, Tool Adoption, User Satisfaction, and UX Improvement.
For 2024 H1 projects, I established trackable KPIs aligned with the goal of releasing 30,000 titles and 500,000 content files by July 2024. Additionally, I initiated quarterly feedback surveys to monitor user satisfaction.
My user interviews identified key bottlenecks in the content metadata workflow, revealing why users relied on inefficient methods.
These insights shaped the 2024 H1 business goals and scope:
These insights led to two major design solutions, aligning with the goal of improving team efficiency to scale for 30,000 titles and 500,000 content files by July 2024..
Problem
Content operators relied heavily on importing/exporting CSV files to manage metadata. Despite frequent errors, they preferred Excel due to its bulk editing, customizable filtering, and compatibility with partner-provided files, which the CMS lacked.
Solution
Reduce manual CSV file exporting/importing
To address the reliance on manual CSV exporting and importing, I introduced a solution that made the CMS fully compatible with spreadsheet files.
This new feature offers Excel-like functions directly within the CMS, allowing content operators to manage and edit metadata without leaving the platform. By integrating familiar spreadsheet functionality, this solution minimizes input errors and significantly reduces the total lead time for content creation and updates.
Impact
Problem
Despite the tool's potential to reduce manual poster creation work, graphic designers were hesitant to fully adopt the AI-powered poster generation feature, citing low trust in the machine learning logic.
Solution
To drive adoption, I rebranded the feature from "poster auto-generation" to Magic Studio, positioning it as an intelligent, AI-driven tool. I introduced user-friendly language and gradually educated users within the app to emphasize the "AI" aspect, enhancing trust in its capabilities.
Impact