Data Tag Management

I designed a new tagging system to help enterprise users classify and take action on millions of files. The project focused on solving problems of scale, security, and role-based access. It turned a messy, unstructured system into one that is organized, governed, and easy to use.

Type

Web App

Enterprise UX

Interaction Design

Role

UX Designer

Team

1 Product Manager

3 Software Developers

Duration

4 months

OVERVIEW

Search chaos → Instant clarity

Let’s think about our everyday experience searching for files or directories on Mac or Windows. Most of us have felt frustrated when we couldn’t quickly locate the exact file we needed. Komprise users faced the same problem. For example, to find an SF customer contract file in the old system, they had to rely on filters like file name or modified date. This process required trial-and-error combinations and often wasted valuable time.

To solve this, we introduced a tagging system. Now, users can simply apply tags such as Region: SF, Type: Contract, Year: 2023 and instantly get the precise file they are looking for. This allows them to move much faster into the next step of their workflow, whether it’s analytics or migration.

BUSINESS PROBLEM

The hidden cost of data browsing issues

Data browsing accounted for 24% of all support tickets in 2023. In March alone, tickets spiked to 18, showing just how often users struggled to locate the right files. This wasn’t just inconvenience, it led to wasted time, repeated tasks, and rising support costs. Ultimately, inefficient browsing hurt both productivity and customer satisfaction.

PERSONA

2 core players in data management

Through research, I identified two key personas: IT Admin and Read-only User. They were not just isolated users, but core players working together to explore and manage enterprise data. For example, the Read-only User groups the data they need, and the IT Admin applies policies to those groups, such as backup, migration, or retention.

USER INTERVIEWS

From these interviews, I identified several key insights.

  1. Need for meaning-based grouping

Users wanted to organize data by project name, region, or compliance level, but the system only allowed rigid attribute-based filtering.

  1. Need for efficient filtering

Users had to re-create filter conditions for tens of thousands of files, adding significant time and cognitive load.

  1. Need for collaborative consistency

Different users classified the same data in different ways, leading to errors, increased operational costs, and no ability to reuse or share conditions.

  1. Need for automation

Users wanted to group and store data once, then set rules to automatically migrate, back up, or delete files according to predefined policies.

DESIGN PRINCIPLES

Based on our research, we defined three main design goals for the tagging system.

IA

Designing a Centralized Tag Library with Role-based Access

We created a centralized Tag Library to manage and reuse all tags across the system. To ensure security, Admins can create, edit, and delete tags, while Read-only users are limited to viewing, applying, and un-tagging. This separation minimizes accidental changes and ensures both governance and usability.

DESIGN ITERATIONS

Tag library

The initial design included a heavy left-side panel for creating new tags. However, this layout felt redundant, since users mainly needed a quick way to view tags without adding or editing.
In the final design, I simplified the structure by removing the side panel and consolidating all actions into the table view. Now, tags are listed in a clean table, a shortcut button at the top allows quick creation, and contextual actions like Edit and Delete are aligned on the right.
This iteration process made the interface lighter, easier to scan, and faster to use.

DESIGN ITERATIONS

Create Tags

This screen represents how users create tags, shown as a popup on top of the Tag Library.
In the first iteration, users had to manually add each key and value pair in separate rows, which quickly became tedious when creating multiple tags. In the final design, I optimized the process by allowing users to enter comma-separated values under a single key. These values are displayed as chips, which can be individually removed if needed. This approach reduces repetitive input, makes batch creation much faster, and still gives users full control over editing or deleting each tag.

DESIGN ITERATIONS

Apply tags

We decided not to go with the one-click apply flow. Through user testing, we learned that admins preferred safety over speed, especially for large batch actions. So, we added a confirmation step.

DESIGN ITERATIONS

Un-tag

Because Un-tag is a destructive batch operation, we added a confirmation step as well.
The Confirm dialog summarizes the selected tag keys and values in a table, and the admin can choose Back or Confirm.
In user testing, this extra step significantly reduced accidental mass removals and increased trust in the system.
After completion, a success toast appears and the action is written to the audit log for traceability.
Access is role-based, so read-only users do not have permission to perform Un-tag.

OUTCOME

Data search became faster and clearer

The outcome of this project was clear. The average time required for users to group data dropped from 8 minutes to 3 minutes, a 62% reduction. This was achieved by eliminating repetitive filtering and enabling instant grouping through tagging. In addition, the number of support tickets related to data filtering significantly decreased. From an average of 8 tickets per month at the end of 2024, it went down to just 3 tickets in March 2025, reducing operational overhead and support costs.
In short, this project improved both user efficiency and business performance.

Enter Password