Explorazor

In 2023, vPhrase Analytics, a Data Analytics & NLG Technologies company based in Mumbai launched Explorazor, a market research and data analysis platform tailored for large FMCG (Fast-Moving Consumer Goods) enterprises. The software integrates large datasets, enabling users to uncover actionable insights quickly.

Here’s the story of how we redesigned the way brand managers and sales teams analyse, explore and visualise their data.

Clients:

My Role

When I joined the vPhrase product team in August 2022, they already had an initial prototype for Explorazor. Despite its powerful capabilities, the initial version of the platform faced challenges in usability and user experience, limiting its effectiveness. I worked alongside a product manager and a junior UX designer for 6 months to revamp Explorazor’s UI/UX with the following objectives:

Improve usability

For diverse user profiles, from data-savvy analysts to general managers with minimal technical expertise.

Enhance the visual appeal

Of the platform to instil user confidence and reflect the enterprise-grade nature of the software.

Develop a robust design system

For consistency, maintainability, scalability and smooth collaboration with the tech team.

Throughout the product lifecycle, we undertook many sprints to try and test various features that align the platform with the fast-paced needs of its enterprise users in an agile environment in before deploying them.

The Challenge

Initial feedback from users highlighted several critical challenges:

Overwhelming Dashboard: Users reported the dashboard presented too much information, with little hierarchy leading to cognitive overload.

Inconsistent UI Design: The lack of a cohesive design language caused visual inconsistencies, confusing interactions, and a lack of professional polish.

Complex Navigation: Important features and datasets were buried under multiple layers of unintuitive menus, making them hard to locate.

Limited Workflow Support: The platform lacked intuitive workflows that integrated with users' tasks, slowing down insight generation.

Steep Learning Curve: New users found onboarding difficult due to insufficient guidance and poorly structured workflows.

Feature Experimentation: As the product was in its nascent stage, new features were tested frequently, requiring flexible designs that could evolve quickly without creating a fragmented user experience.

Research and Discovery

1. User Interviews

In order to understand user needs, pain points and their current workflows, close to 50 interviews were conducted with Brand Managers and Insights Teams of various FMCG and pharmaceutical companies. The insights revealed that:

The biggest challenge was dealing with disconnected datasets. Brand and sales users had to switch between multiple platforms to gather and analyse data, which created a fragmented workflow. This made it overwhelming for users, especially non-technical business users, to navigate large amounts of data and find meaningful insights.

PowerBI was widely favoured for data exploration, despite its shortcomings in workflows and dataset centralisation. Its popularity stemmed from users' long-term familiarity with the tool, creating a sense of attachment that made it a strong indirect competitor.

Users had diverse levels of technical expertise. Analysts wanted advanced tools, while managers preferred quick, high-level summaries. But both needed to be able to drill down into the data to understand the reasons and causes behind a certain insight.

Time was a critical factor. Enterprise teams needed to derive insights quickly, often under tight deadlines.

Collaboration was essential. Users needed features that enabled seamless sharing of data insights with team members.

2. User Personas

Through detailed user interviews with end-users from brand and sales teams, we created three primary personas:

Data Analysts

Advanced users comfortable with complex datasets, requiring customisable tools.

Degree in Statistics, Economics, Mathematics, Computer Science, or Business Analytics.

Uses SQL, Python/R, Excel, Power BI, Tableau, Google BigQuery, Snowflake, internal BI tools.

Tasks Performed:

Extracts and prepares data from various sources.
Builds and maintains dashboards.
Performs advanced analyses (e.g., sales trends, pricing impact, market share).
Supports brand and sales teams with data-driven insights.

Brand Managers

Mid-level users needing quick summaries and insights for presentations.

Background in Marketing, Business Administration, or Management, often with an MBA.

Excel, PowerPoint, Power BI, Nielsen/IRI tools, CRM platforms, internal planning tools.

Tasks Performed:

Plans and monitors brand campaigns.
Analyses brand performance (e.g., awareness, market share, distribution).
Uses insights from data teams to inform strategy.
Collaborates with sales, media, and product teams to align execution with goals.

Sales/ Brand Leads

Executive-level users looking for high-level insights with minimal effort.

Degree in Business, Marketing, or Commerce; often MBA-level education.

CRM systems (like Salesforce), Excel, Power BI, internal sales platforms, syndicated retail data tools.

Tasks Performed:

Reviews sales performance across geographies or channels.
Makes strategic decisions on promotions, pricing, and product placement.
Relies on data to optimize sales force actions and retail strategies.
Works closely with both analysts and brand managers to align sales efforts with broader brand objectives.

3. Heuristic Evaluation

I performed a heuristic evaluation of the existing Explorazor interface based on Jakob Nielsen's usability principles, identifying key usability gaps:

Unintuitive Navigation, e.g., Important features were buried under multiple layers, making them hard to locate., e.g., lack of loading indicators during data processing.

Low consistency and standards, e.g., inconsistent use of fonts, colours, and UI components.

Cluttered Dashboard, e.g., overwhelming dashboard due to excessive information density.

Poor visibility of system status, e.g., lack of loading indicators during data processing.

4. Competitive Analysis

I also studied competitors like ThoughtSpot (direct competitor) and Power BI (indirect competitor) to identify:

Data visualisation best practices

Efficient enterprise workflows

UI Design and design industry standards

The Redesign Process

1. User-Centred Design Approach

I adopted a user-centred design process, involving stakeholders and end-users at every stage to validate concepts and ensure alignment with their needs.

Empathy Mapping

To understand user pain points, I created empathy maps, identifying what users:

Prioritisation Matrix

To address user pain points, I created a prioritisation matrix that evaluated each issue based on two dimensions: business impact and implementation feasibility. This approach ensured that high-impact, easily implementable solutions were tackled first, optimising resources and results. On evaluation, we fixated on - and - as the most

2. Conceptualisation

I began by creating the information architecture and low-fidelity concepts for the main user flows. Once the product manager, developers, and stakeholders were on board with the direction, we ran usability tests to make sure the designs worked well in practice. After gaining confidence from the results, we moved ahead and turned the wireframes into high-fidelity designs for development.

2. Designing the Dashboard

Key Changes:

Modular Widgets: Users can now customise their dashboard by adding, removing, or resizing widgets to prioritise relevant insights.

Focus on Key Metrics: I grouped data into easily digestible sections such as "Top KPIs," "Recent Trends," and "Critical Alerts."

Visual Hierarchy: Typography, spacing, and colour coding were optimised to draw attention to critical information first.

Result:

Reduced user confusion by presenting relevant data upfront.

Improved user productivity by enabling task-specific configurations.

Solution

Introducing Explorazor , a data exploration tool that allows users to connect multiple datasets and effortlessly analyse them with a "Google - like" search interface.