Optimizing Data Visualization for Scale

Optimizing Data Visualization for Scale

Optimizing Data Visualization for Scale Image

Role

Lead Product Designer

Mentor

Optimizing Data Visualization for Scale

Optimizing Data Visualization for Scale

My Role

Lead Product Designer, Design Mentor

Team

Project Manager, Development Team, 2 Junior Designers

Problem

Caseware couldn’t render 30,000+ account groups took a few mins to load, costing an estimated $10–50M in lost enterprise deals annually.

Solution

Refactored 3 core interfaces — data visualization, configuration workflows, and risk identification — introducing a scalable List view, bulk management, and an optimized risk review experience.

Impact

  • 10× data scalability

  • 60% faster setup

  • 3 enterprise clients secured

Uncovering Technical and System Complexity

Part 1 : How would we enhance the visuals to support scalability?

Our starting point was the existing graph visualization used for trial balance reports. The introduction of fund reporting mandated a solution to scale this component, as it required handling significantly larger, multi-level organizational hierarchies in a user-friendly way.

Current Issues
  • Interface was missing zoom in/out functionality required to manage and interpret extensive data

  • Inconsistency with menu options and user messages

  • No search capability

  • Interface needed visual update for improved usability


Existing Interface

Data Driven Synthesis and Intentional Design

The solution began with a deep synthesis phase, translating qualitative user insights into data modeling and visualization requirements essential for large-scale data efficiency.

Defining Critical Data Density

User interviews with 5 users rapidly established that reviewers prioritize a handful of data points for rapid triage before diving into detail. 

These requirements became the constraints for the new card design, ensuring density without visual noise:

  • Firm Name: Primary identifier.

  • Ownership %: Key financial relationship metric.

  • Year-End Date: Critical for audit scheduling and jurisdictional compliance.

  • Engagement Status: A binary signal (Caseware Linked / External Data), vital for managing data import workflow risks.


Card view explorations


Optimizing Layout for Scale

Early testing showed the standard vertical layout wouldn't work for the massive amount of data in fund reporting. We chose a horizontal layout because it aligns with the user's existing mental model and significantly improved how many cards could be viewed on the screen, which was critical for managing large datasets.


Chart view explorations

The Grid vs Graph Dilemma

Rendering 30,000+ hierarchical nodes in the Graph View proved computationally difficult, generating unacceptable load times of over 10 seconds and creating potential blocker for enterprise adoption.

The decision to introduce the Grid View wasn't a feature addition; it was a strategic design intervention to solve a critical platform performance crisis at scale.

This technical compromise allowed us to meet the < 2 seconds load time requirement for the largest client datasets, successfully transforming a technical constraint into a scalable, high-performance solution for the reviewer workflow.


Final Designs


Accounting firms hierarchical graph view


Accounting firms hierarchical grid view


Part 2 : Data Configuration Enhancements for Scalability

This interface was another critical component that focused on the configuration of categories and dimensions for large-scale reports.

This is not just updating a UI. It is a data configuration challenge where the interface must support complex, many-to-many relationships across potentially tens of thousands of account groups.

In a nutshell:

  • There can be many categories

  • Each category can have multiple dimensions with multiple levels

  • Categories and dimensions must be assigned to multiple account groups

Problem: Efficiency Debt at Scale

The existing workflow was built for small datasets as a source of "efficiency debt". For large enterprise clients, manually assigning categories and managing complex hierarchies became a bottleneck, resulting in:

Repetitive Effort: Assigning these configurations required property edits for each accounting group. 

Cognitive Load: The lack of bulk controls made managing the complexity of multiple categories, each with multi-level dimensions, tedious and error-prone.

Adoption Barrier: This inefficiency made the Cloud setup process significantly slower than legacy solutions.


Existing interface for setting up dimensions


Strategic Solutions

My design strategy was to focus on centralized bulk management instead of individual property edits of accounting groups. The solutions were directly targeted at high-volume pain points:


Inline Group Assignment

  • Decoupled data assignment from the main trial balance properties screen, allowing users to assign hundreds of account groups directly from the dimension setup interface.

  • Impact: Dramatically cutting setup time for large datasets

Centralized Default Setting

  • Introduced "Default" column with radio button for one-click action on every dimension

  • Impact: Reduced UI clutter

Drag and Drop Interaction

  • Updated the interaction model from slow, click-based movers to a modern drag-and-drop hierarchy management, acknowledging that dimension order is critical.

  • Impact: Significantly accelerated workflow speed for power users.

By tackling the efficiency of bulk group assignment, we inherently designed for solving the blocker for creating dimensions more than 2 levels.


Dialog box to setup categories and dimensions before user feedback


Dialog box to setup categories and dimensions

Wrapping Up

This redesign transformed a scalability bottleneck into a high-efficiency configuration experience, directly supporting the project's goal of enhancing user efficiency while working with large datasets.



Part 3 : Risk identification at enterprise scale

The final stage focused on ensuring that the scalability project did not compromise on data integrity - a mandatory step of Risk Identification within the auditor's workflow.

Challenge: Information density vs limited real estate

Risk identification is critical for auditing, but the UI had to support a single risk being associated with an exponentially larger number of accounting groups due to the new fund hierarchy structure. Combining this with limited real estate because of standardized designed components and ensuring usability, increased the complexity.

Solution:

My approach was to optimize the existing design system patterns to accommodate the new data scale, focusing on rapid triage and review process:

Collapsed View Optimization: I restructured the risk card layout to divide its content into two distinct, scannable sections. This strategic segmentation improved the readability and information hierarchy of the risk summary, allowing reviewers to do quick assessments.

Impact: Improved overall velocity of the risk review workflow


Collapsed view optimization


Expanded View Optimization: For the detailed view, I prioritized pattern consistency. Ensuring that users can now associate a risk to a large number of accounting groups without having to edit individual properties.

Impact: Reduced cognitive load during configuring and investigating phases.


Expanded view optimization


With this work, the auditor's mandatory process of identifying and assessing risk made the review process efficient, clear and improved usability.

Measurable Impact

The project launched successfully in 2024, and the subsequent performance data confirms that the design solutions not only addressed the immediate pain points but delivered on the core strategic platform goals for scalability and efficiency.

System scalability: improved 10X in data handling capacity 

The final designs rendered a dataset of 30,000 account groups with target performance of < 2 seconds load time.

Efficiency: User productivity improved by 60% 

The redesign of the dimensional configuration workflow resulted in a 60% reduction in setup time for assigning accounting groups. By minimizing clicks and converting an hours-long task into a minutes-long process, the design dramatically enhanced user efficiency. This was a key driver for Caseware Cloud adoption by three large fund reporting clients.

Learnings

The importance of gaining a holistic understanding of the problem

  • This project affected multiple interfaces across the system, each with unique user requirements.

  • By persistently seeking answers and collaborating with the right people, I was able to connect the dots and develop tailored solutions for these challenges.

Bridging design and engineering

  • After conceptualizing the design solution, I shifted to an engineering perspective to evaluate its integration with existing systems.

  • This approach allowed me to define the project's scope effectively and propose practical solutions that aligned with both business goals and development feasibility.

Impact from my design leadership

  • Guiding junior designers on this project required time and patience to bridge the gap between business knowledge and design skills.

  • Seeing the new designers gain confidence by the end of the project was a fulfilling moment and a testament to the power of strong design leadership.

Let's connect

I'm not just here to design products; I'm here to connect with people.

As a product designer, I'm on an exciting journey to blend creativity with technology to craft memorable user experiences.

linkedin

Email

nanajkarshweta@gmail.com

location

Toronto, Canada

Designed with purpose and plenty of snack breaks🍪

Copyright @ Shweta Nanajkar 2026. All Rights Reserved.

Designed with purpose and plenty of snack breaks🍪

Copyright @ Shweta Nanajkar 2026. All Rights Reserved.

Designed with purpose and plenty of snack breaks🍪

Copyright @ Shweta Nanajkar 2026. All Rights Reserved.