Designing a Scalable BI Platform for Actionable Organizational Insights
A comprehensive redesign of a Business Intelligence platform, making complex data accessible and immediately actionable for every stakeholder.
85%
User Adoption Rate
Achieved 85% adoption across all departments within just three months.
60%
Faster Decisions
4.8/5
User Satisfaction
40%
Less Support Tickets
role
Product Designer
platform
Web Dashboard (BI)
duration
6+ Months · 2026
team
2 Designers · Data Eng
company
Karnival
What is Karnival BI?
The Next Generation of Business Intelligence
We developed Karnival BI (Business Intelligence) as an end-to-end analytics solution engineered to transform raw data into insights that are truly personal, actionable, and universally accessible. We moved beyond traditional BI tools by integrating analysis and feedback into a single, seamless environment. At its core, Karnival BI simplifies data-driven decision-making, ensuring that every stakeholder from the CEO to the store manager has access to insights that are not only accessible but also hyper-relevant to their specific role and context.
Custom Report Generation
Create comprehensive reports, scaling from high-level strategy dashboards for executives down to granular store-level analysis for operations teams.
Direct Voice-of-Customer Integration
Directly embed and integrate user and customer feedback surveys within the live analytics environment, connecting data trends with real-world sentiment.
Dynamic Insight Visualization
Utilize customizable charts, graphs, and advanced data storytelling tools to visualize complex information clearly and communicate insights effectively.
Organizational Scalability
Seamlessly scale the platform's utility across all levels of the business whether analyzing the performance of the entire organization, a specific brand, or an individual retail location.
The Need
Why build our own custom BI?
We recognized that while tools like Power BI and Tableau were powerful, they fundamentally mismatched our unique operational requirements and strategic goals. We needed an intrinsically integrated, scalable, and cost-efficient platform that commercial off-the-shelf (COTS) solutions could not provide.
Lack of Contextual Relevance
Generic, one-size-fits-all reports required excessive customization to deliver the crucial role-specific insights needed by different stakeholders (e.g., Marketing vs. Operations).
Unsustainable Licensing & Support
Relying on costly, per-user third-party licenses for organization-wide adoption was financially unsustainable and created an unacceptable long-term vendor dependency.
Data Silos and Friction
External tools forced users to switch contexts and applications, preventing seamless coexistence of core analytics, feedback surveys, and performance reporting.
Inflexible Granularity & Adoption
COTS solutions struggled to effortlessly scale insights from macro-organization trends down to the minute details of individual store or brand performance.
Low Adoption & Expert Dependency
The complexity of high-end BI tools necessitated reliance on specialized analysts, severely slowing down and preventing democratized decision-making across the company.
Value Proposition
Turning organizational data into personalized insights that matter
The redesign of Karnival BI was centered on a single goal: We wanted to make organizational data truly relevant and actionable for every employee. We achieved this by delivering highly contextualized views tailored to the specific needs of four key stakeholder groups across the business.
Leadership & Brand Performance Dashboards
Tailored dashboards presenting strategic planning forecasts, brand performance, and high-level business intelligence.
- Strategic Planning
- Brand Performance
- Performance Analytics
- Revenue Optimization
Region & Store Operations Dashboards
Actionable interfaces focused on localized territory performance, store operations, inventory metrics, and market trends.
- Regional Analytics
- Store Operations
- Market Trends
- Inventory Management
Personalized Views
Ensuring every user sees only data relevant to their specific role.
Real-time Insights
Delivering live data to support immediate operational decisions.
Actionable Intelligence
Linking insights directly to a clear, recommended next action.
Data Democratization
Making complex analysis intuitive and accessible for everyone.
Pain Points
What Karnival BI was designed to address
Building on the identified limitations of legacy BI systems, we realized we needed more than just an alternative tool — we needed a complete platform redesign engineered around our organizational needs. Karnival BI was built to solve these five core pain points directly:
Discovery & Research
Deep understanding of our users and comparative audit
We initiated the project by focusing on a deep understanding of our users and their current pain points. This involved four parallel tracks: User Interviews (understanding needs), Current State Audit (analyzing existing reports), Competitive Analysis (benchmarking other BI tools), and Technology Review (assessing platform capabilities).
40%
Access reports weekly
35%
Prefer dashboard format
30%
Primary issue is outdated data
60%
Prefer self-service reporting
4 parallel tracks: User Interviews · Current State Audit · Competitive Analysis · Tech Review
Leadership needed strategic, high-level dashboards.
Executive stakeholders highlighted a critical need for performance trends, forecasts, and visual summaries rather than granular spreadsheets.
Prioritize strategic views with easy options to filter and drill down.
60% of users preferred self-service reporting, but struggled with complexity.
While the majority of users wanted to interact with and generate their own reports, 40% still relied on dedicated BI support due to tool complexity.
Simplify report creation with modular design and pre-configured templates.
30% of users cited outdated data as their primary pain point.
Data latency was a major barrier. Manual cleaning and updating of spreadsheets took hours, meaning decisions were often made on old information.
Move to automated, real-time data pipelines with live updates.
Spreadsheet-based reporting suffered from severe version control issues.
Teams were copying and pasting data manually, leading to error-prone calculations and disjointed metrics across departments.
A single, central platform must replace manual static spreadsheets.
Before
Excel-based Reporting
Manual updates prone to errors, version control issues, and lack of scale or interactivity.
COTS (Power BI & Tableau)
High licensing costs, steep learning curve, expert dependency, and poor survey integration.
After
Karnival BI Dashboard
Automated real-time pipelines, central source of truth, and dynamic interactive views.
Karnival BI Custom Solution
Intrinsically integrated, low learning curve (self-service), and custom survey integration.
Design Exploration & User Testing
Google Tabs vs Modularity and testing rounds
We explored two distinct conceptual models: Google Tabs (role-based layered insights with drill-down navigation) and Modularity (pick-and-drop dashboard approach). Through two iterations and rounds of user testing, we synthesized a hybrid approach combining the best of both: intelligent role-based defaults with progressive disclosure of advanced customization features.
High ease of use (8.5/10) but low flexibility and customization (6.8/10) for advanced power users.
Role-based default views with top-bar tab navigation and drill-down metrics.
Novices appreciated it, but it was too rigid for power users needing custom queries.
High customization (9.1/10) but novices felt overwhelmed by options, causing decision paralysis.
A fully flexible modular interface where users build dashboards using a widget library.
Power users loved it, but novice usability dropped due to complexity.
How to serve both novice and advanced users without creating clutter or rigid limitations.
Intelligent role-based defaults (Iteration 1) with progressive disclosure of customization options (Iteration 2).
Novices start with clear defaults while power users easily edit views (Ease of Use: 9.2/10, Customization: 8.9/10).
Research & Discovery
Understanding stakeholder needs, survey habits, and legacy tool auditing.
Ideation & Sketching
Rapid concept sketching and workshops aligning design, engineering, and business stakeholders.
Iteration 1 & Testing
Building the Tab-based navigation model. Testing showed high ease of use but low flexibility.
Iteration 2 & Testing
Building the Modular pick-and-drop model. Testing showed high flexibility but high cognitive load for novices.
Final Synthesis
Creating the Hybrid Approach: Smart Defaults with Progressive Disclosure.
Problem Solving & Features
Addressing organizational data bottlenecks
The final Karnival BI product is the culmination of a Bottom-Up & User-Centric approach. We focused on democratizing data by building flexible intelligence directly into the operational workflows, ensuring every user from the frontline to the C-suite has personalized, actionable, and real-time insights.
Constraint
Self-Service Customisation
Product Decision
Intuitive drag-and-drop widgets
Trade-off
Required custom visual component library development
Outcome
Empowered non-technical users to build reports independently, reducing analyst support request by 40%.
Constraint
Information Overload
Product Decision
Progressive Disclosure & Personalized Views
Trade-off
Hiding advanced metrics under progressive filters
Outcome
Novice users scan dashboard in seconds; strategic managers access deep analytics when needed.
Constraint
Unified Data Ecosystem
Product Decision
Analytics, surveys, and reporting combined
Trade-off
Higher backend integration and API complexity
Outcome
Seamless correlation between sales trends and real-time customer survey feedback.
Constraint
Reactive Decision-Making
Product Decision
Real-Time Dashboards & Alerts
Trade-off
Added data pipeline infrastructure overhead
Outcome
Proactive alert trigger system notifies managers of sales dips or inventory low states in real-time.
Learnings & Future Roadmap
Platform takeaways and next steps
Throughout the project, we gained crucial insights that not only shaped the Karnival BI platform but also redefined our approach to large-scale data product development. Moving forward, the roadmap focuses on predictive modeling and natural language querying.
User Research is Non-Negotiable
- · Deep user research revealed insights that completely changed our design direction multiple times.
- · Close partnership with engineering and product teams was essential to ensure feasibility of data pipelines.
Design for Diverse Users
- · One size doesn't fit all — different roles needed completely different experiences within the same platform.
- · The winning formula was hiding advanced complexity behind intuitive defaults (progressive disclosure).
Future Scope & Roadmap
- · AI-Powered Insights: Integrating predictive analytics to forecast sales trends, inventory needs, and campaign performance.
- · Natural Language Queries: Enabling users to talk to data with plain language (e.g., 'Show me last quarter sales').
- · Real-Time Data Streaming: Moving beyond near real-time to live streaming data pipelines for instant updates.
Want to go deeper?
I can walk through research artifacts, the Figma file, or the PRD in a live conversation.