Quicksight dashboards
Quicksight dashboards
Embedded Files
Designing a scalable dashboard framework that standardised reporting across Motion's products while improving usability, accessibility, and collaboration between UX and Data teams.
Cartell is a vehicle-data and automotive-intelligence company, supplying car history checks and analytics to dealerships, insurers and finance providers. Its reporting had grown organically into roughly 306 QuickSight dashboards across three Motion brands(Carzone,CarsIreland, Cartell), built on 343 datasets with little consistency in structure or design.
I led a full redesign built around a reusable three-tier framework - Summary → Detail → Comparison - that gave internal teams one clear path from high-level KPIs to deep analysis. Early validation projected a reduction from 306 dashboards to around 77, and forecasting time per client from 2–3 hours down to 1.5 hours.
Cartell VRM Dashboard Redesign
Cartell VRM Dashboard Redesign
Bringing 306 fragmented dashboards into one consistent, intuitive system
Role: Product Designer · Product: QuickSight VRM Dashboard · Users: Internal B2B teams · Team: Product, Commercial Ops, Data, Engineering
The Problem
The Problem
Users didn't need more dashboards - they needed a clearer, more intuitive way to reach the right information.
Cartell's reporting had been built up over years. As new products, customers and teams appeared, new dashboards were created to solve one need at a time. The result was roughly 306 dashboards on 343 datasets across three Motion brands with no shared structure. Users had to switch between several dashboards to complete a single task, key information was hard to locate, forecasting took 2–3 hours longer than it should, and new employees struggled to learn the system at all.
Research
Research
I used a mixed-methods approach - user interviews, personas, journey mapping, affinity mapping and information architecture -structured loosely around the Double Diamond to move from problem to solution.
The most important decision was to make research a shared input for both design and the Data team. Rather than designing a dashboard and asking engineers to populate it, we collaborated early, so the research shaped what data was exposed, how it was grouped, which filters existed, and how users moved between high-level insights and detailed analysis.
User interview profile - Commercial Operations Manager Cartell
User interview profile - Corporate Account Manager
Interviews with commercial and account-management users surfaced consistent pain points - fragmented dashboards, inconsistent layouts and hard-to-reach insights - which fed directly into the new structure.
Design Approach
Design Approach
Working closely with the Data team, I mapped how users actually moved through the existing dashboards, then defined an interaction model for the new structure - accounting for QuickSight's technical constraints while aligning to how people research and report.
Wireframes
Wireframes
Based on the agreed flow, I created low- and mid-fidelity wireframes across all three sheets - Summary, Detail and Comparison - to test the structure before committing to visual design.
Iteration
Iteration
Continuous feedback from the Data team shaped the design. A few of the changes that mattered most:
Standardised the date filter format to DD/MM/YY
Created a single, predictable filter pattern across every sheet.
Designed the K-Type and N-Type metrics as line graphs
Chosen with the Data team to show mapping-rate trends clearly over time, after weighing alternative chart types.
Introduced a consistent colour palette across dashboards
Replaced ad-hoc styling with one shared visual language, tuned to work within QuickSight's rendering limits.
Theming
Theming
With no formal design system in place, I created a consistent visual language for the dashboards — type scale, colour and reusable patterns — that balanced Cartell's brand identity against the technical limitations of Amazon QuickSight.
High-Fidelity Designs
High-Fidelity Designs
Summary Sheet
Summary Sheet
An executive overview of VRM performance - key KPIs, trends and customer activity at a glance, without needing to dig into detailed data.
Detail Sheet
Detail Sheet
The "why" behind the summary: detailed performance data, period-on-period comparisons and product-level insights for deeper investigation.
Comparison Sheet
Comparison Sheet
A side-by-side view for comparing two datasets - periods, countries or customer segments - to surface differences in performance and activity.
AI Exploration
AI Exploration
Toward the end of the project I began building an AI-assisted design workflow. I used Figma's AI agent to generate an accessible chart-and-graph colour system that extended the existing brand palette, reusable Make Kits to spin up dashboard concepts quickly, and an AI-built KPI dashboard to prototype the QuickSight look and feel before handing over to engineering. This became the starting point for an approach I carried into later dashboard projects.
Impact & Reflection
Impact & Reflection
Impact
Established a reusable dashboard framework (Summary → Detail → Comparison) for future QuickSight work
Improved collaboration between UX and the Data team, so research informed both the experience and the underlying data structure
Introduced consistent layouts, theming and patterns despite the absence of a formal design system
Reduced design iteration time by building reusable components and patterns
Projected from early validation and follow-up with users: 306 dashboards → 77, and forecasting time per client from 2–3 hours → 1.5 hours.
What I learned
Designing enterprise dashboards is a constant balance of user needs, business goals and technical constraints. Close collaboration with the Data team and engineers led to a more implementable, user-centred result than designing in isolation would have. This project also marked the start of my AI-assisted workflow — using Figma Make and reusable kits to explore concepts quickly — and established a scalable approach I've continued to apply since.
Google Sites
Report abuse