CASE STUDY
Project Summary:
ACTT Data Cleanup Platform
Transforming Big-Data Chaos into an Intuitive, Enterprise-Grade Experience
ACTT is a high-performance big data scanning application designed to identify and categorize ROT (Redundant, Obsolete, and Trivial) data across massive company infrastructures. Facing a legacy system with a low usability score of 63.3 and stakeholder concerns regarding "incessant scrolling" and complex navigation, I led a complete UX/UI overhaul. By implementing a Dashboard-based interface and a proprietary Design System, I transformed a tedious administrative chore into a "sexy," modern, and highly efficient workflow that delighted users and exceeded business requirements.
Project Overview: The "ROT" Specialist
Cleaning Big Data with Precision and Style
-
The Product: A high-performance data-scanning enterprise tool that identifies and categorizes company-wide data into ROT (Redundant, Obsolete, and Trivial) probability.
-
The Goal: Empower data owners to declutter massive datasets through an intuitive, "low-friction" interface that makes data hygiene feel less like a chore and more like a streamlined task.
-
My Role: End-to-end Product Designer. I led the UX strategy, information architecture, wireframing, and the final high-fidelity UI/visual identity.
-
Tools: Figma (Prototyping & Design), Miro (User Flows).
The Challenge: Humanizing the "Data Swamp"
Enterprise data cleanup is notoriously painful. Legacy tools are dense, intimidating, and feel like punishment. Our challenge: design an interface for a complex big-data operation (scanning, ROT analysis, bulk actions) that felt inviting, sleek, and genuinely intuitive—not like enterprise software.
The client needed a big-data scanning solution that could process millions of records, classify them by ROT probability (Redundant, Obsolete, Trivial), and empower data owners to take action—without the usual enterprise software dread.
Specific constraints:
-
Complexity disguised as simplicity: Surface-level ease with powerful backend scanning
-
Trust at scale: Users needed confidence before deleting/archiving company data
-
Diverse technical comfort: From data scientists to department heads with minimal technical fluency
-
The "fun" mandate: Create an interface so inviting that teams would want to clean house
Research & Discovery
I dove into the purpose of the application - the core reason for developing it: Here's what I discovered:
-
The Business Pain: Companies like Fortune 500s are drowning in data, with over 30% of content reported as ROT. This leads to massive storage costs and legal/compliance risks.
-
The Technical Barrier: The existing system had a Heuristic Evaluation score of 63.3/100. Stakeholders were worried about "incessant scrolling," inconsistent terminology, and a total lack of undo/redo functionality.
-
Visual Personality: The prominent "wish list" item that all respondents had in common was that the application should have a "sexy, modern look."
-
To understand the complexity of ACTT, I interviewed six key stakeholders ranging from DevOps Leads to Executive Sponsors. A common theme emerged: The data review process was a 'black hole' of productivity. Users were intimidated by the scale of the data and frustrated by 'incessant scrolling' and a lack of session saving.
I also conducted a heuristic evaluation of the legacy system, which scored a mediocre 63.3/100 due to poor navigation, inconsistent terminology, confusing workflow processes and no data visualization.
My strategy was clear: we needed to stop treating data as a list and start treating it as a manageable workflow."
My Heuristic Evaluation Score:
63.3/100
Ideation & Design Evolution
Armed with my research data, I created a UX Strategy Canvas. Key highlights:
I leveraged the HEART framework to prioritize 'Happiness' and 'Task Success'. This led to a complete architectural shift:
The Dashboard-First Approach: I replaced static reports with an interactive, widget-based dashboard that users could customize to their specific 'Campaign' needs.
Based on user stories, I designed a 'Snapshot' save system. If a user like 'Jack' is interrupted while reviewing 1,500+ files, the system now prompts them to save, ensuring zero data loss.
Visual Clarity: I introduced a proprietary design system to replace the 'unconventional gestures' of the old platform with standard, intuitive patterns like column-level filtering and clear hover states.
UI & Visual Identity - Transforming Complexity into "Sexy, Modern, and Painless"
To meet the stakeholder requirement of a "sexy" and "modern" interface, I developed a Dark Mode-first proprietary Design System. This choice was strategic: dark interfaces reduce eye strain for data owners who must review thousands of files in a single session.
1. The Dashboard: A Command Center for Data Health
Instead of forcing users to dig through tables, I designed a High-Density Dashboard that provides an immediate "health check" of the data ecosystem.
-
Key Callout: I used vibrant, high-contrast greens and blues to highlight "ROT Probability," making the most critical data points (High/Medium/Low ROT) stand out against the dark background.
-
Data Visualization: I replaced the reliance on external tools with integrated, custom-coded charts (Donut charts, Bar graphs, and Treemaps) that allow for real-time progress tracking.
2. The Data Review Workspace: Focus & Flow
The core of the application—the Data Review screen—was designed to eliminate "incessant scrolling".
-
Persistent Overview: A top-level summary bar keeps the user oriented with "Total Files" and "Date Requested" always in view.
-
Failsafe Actions: The "Save Changes" and "Complete Review" buttons are placed prominently above the fold, directly addressing the "Jack" user story where users need to pause sessions without losing progress.
-
Streamlined Filtering: I implemented a custom "Automation" sidebar that allows for rapid sorting and filtering by Directory, File Size, or Content Type, drastically reducing the time spent hunting for specific records.
3. Integrated Guidance & Failsafes
To ensure the product remained "painless" to use, I embedded help directly into the UI.
-
Instant Help Overlays: I designed a "help" toggle that highlights interactive elements with red outlines and descriptive tooltips, allowing users to learn the platform in under 3 minutes without a manual.
-
Clear Success Cues: Actionable items like "Request Reassignment" use modal windows to keep the user focused on a single task, preventing errors in complex organizational workflows.
The Impact: From "Mediocre" to "Delightful"
By moving from a legacy heuristic score of 63.3 to this redesigned state, the impact was felt immediately:
-
Usability: Stakeholders defined success as someone saying "That's cool—and it solved my problem". Post-launch feedback confirmed that the "Data Review" process—previously the biggest user headache—became the most praised feature.
-
Aesthetic Trust: The "sexy" and "engaging" layout transformed the perception of the tool from a technical utility to an essential enterprise asset, helping the company differentiate itself from "big company" legacy competitors.
-
Intuitive: Legacy and new users were thrilled with the ease of use and how natural they felt the workflow to be. Many reported the enjoyment of using the application.
-
Confidence: Users reported that they trusted the interactions and the tasks they were performing, with no questions unanswered during use.