40% of users immediately changed chart colors post-insert—a clear signal defaults were broken. Championed a modernized palette shipped in 8 weeks that achieved 96% acceptance, eliminated first-touch friction, and became the visual foundation for all AI-powered charting features in FY25.
Liked this project?
Let's talk about what we can build together.
Introduction
Excel charts looked embarrassingly outdated. About 40% of users spent their first 30 seconds after inserting a chart manually changing colors. That's not a preference signal. It's a rejection signal, delivered before anyone looked at the data. I pushed for a deceptively simple fix: modernize the default color palette. It was the bet with the most reach for the least effort of anything on the FY25 charting roadmap. It shipped in eight weeks, hit 96% retention, and became the visual floor every AI feature after it had to render onto.
Results Overview
Kept Modern Colors
Concept to Ship
Blind Preference vs. Legacy
The Problem: First Impressions Were Killing Adoption
CRITICAL INSIGHT
The 'insert moment' was our one chance to make a good first impression. If charts looked bad immediately, users wouldn't stick around to discover the good features.
BUSINESS IMPACT
Users weren't just unhappy. They were taking work out of Excel to finish it somewhere else.
What Users Told Us
"I always have to fix the colors before sharing." - User research
"Something external facing, I'm more inclined to use Canva." - User research
This wasn't aesthetics. It was a credibility problem. If the first thing users saw looked outdated, they assumed:
The Trigger: Usage Was Flat and Low
THE INSIGHT
The biggest drop-off happened at the "create" moment—not awareness, not customization. If we couldn't fix the first impression, nothing else mattered.
With ~400M Excel users, charting usage should have been massive. It wasn't. Leadership made user adoption the #1 priority for FY25 charting investments.
My Approach: Design as Strategic Leverage
With adoption as the north star, I asked: Where can design have the fastest, highest-impact win? When the team was debating priorities, most designers wanted to work on 'bigger' problems like AI recommendations or new chart types. I argued that modern colors should be P0 (highest priority) for four strategic reasons:
| Criterion | Why This Mattered | Risk of NOT Doing It |
| 1. Universal Impact | Every user who inserts a chart sees this. No discovery needed. | Other features only reach users who find them. Bad defaults reach everyone. |
| 2. Low Execution Risk | 6 hex codes. No AI, no new APIs, no complex UX flows. And no dependency on the Legacy/Ivy engine migration that made most other UI work on this team expensive and risky. | AI features could get delayed. This couldn't fail. |
| 3. Foundational | AI insights look better on modern charts. Design recommendations need good defaults to build on. | Shipping AI features on ugly defaults would waste their potential. |
| 4. Measurable | Clear success metric: % who keep vs. revert to old colors | We'd never know if it worked without shipping and measuring. |
The team bought in. Framing this as strategic leverage - not 'just colors' - got approval to make it our first P0 deliverable.
My first pitch was to make the new palette the document theme default - the colors driving formatting across the whole workbook, not a setting scoped to charts. Our PM and the platform design lead pushed back. Platform owns the unified theme experience across every Office app, and asking a team that didn't report to us to reprioritize their roadmap for a change they hadn't requested was never getting approved at that scope.
I argued the bigger version for a while because it felt like the more complete fix. It wasn't wrong. It just wasn't ours to ship. Once I understood whose roadmap I was asking to bend, the smaller version, six hex codes scoped to charts, stopped looking like a compromise and started looking like the correctly sized bet. It's also the version that shipped in eight weeks with no new APIs.
Hero JTBDs
“When I insert a chart in Excel, I want it to look presentation-ready immediately, So that I can share insights with stakeholders without spending time reformatting.”
When we ship chart improvements, we need the default visual foundation to be modern and credible, so that users trust the output, stay in Excel rather than exporting elsewhere, and adopt the advanced features coming behind it.
My Role: Senior Designer IC
Anchor all decisions to the Fluent Design System's pre-validated research. This turned subjective debates ("I like blue") into objective discussions ("Fluent guidelines specify X for data visualization").
Problem: "Modern colors" meant something different to everyone:
Phase 1
A competitive audit covered 30+ apps: AI-powered tools (Napkin AI, Julius), design-focused platforms (Canva, Pitch), BI heavyweights (Tableau, Power BI), and the direct competitor (Google Sheets). It confirmed the adoption mandate was right.
Colors emerged as the clear winner: low effort, universal reach, immediate impact on every single chart.
Phase 2
After initial hit-and-trial across various color schemes went nowhere, I built on the Fluent Design System team's existing work instead of creating new color research from scratch.
What I got -
Phase 3
✅ Gut Reaction + Stakeholder Review
✅ Accessibility & Edge Cases
✅ Initializing Fastfood:
✅ Staged Rollout
Gut Reaction + Stakeholder Review: Blind A/B with 20 designers/PMs—90%+ would present new palette as-is vs. ~40% for legacy. 8 candidates presented to leadership; unanimous first-round approval. Fluent anchoring killed subjective debates.
Accessibility & Edge Cases: Passed WCAG AA, colorblind simulations, grayscale, dark mode. Stress-tested 12+ series, bad projectors, print scenarios—palette held with one minor luminance tweak.
Initializing Fastfood: 5% power users ran it in real work for 2 weeks. Zero reverts, unsolicited praise.
Engineering & Telemetry: Built opt-out toggle (bet users wouldn't use it), instrumented ChartInserted/ColorChanged/Deleted events to measure adoption.
Staged Rollout: Dogfood → Fast Ring → Slow Ring → GA. Zero critical bugs. Internal comms built momentum before launch.
User Journey: Before vs. After
Key Learnings: When to Fight for Small Changes
The Bigger Lesson
Sometimes the most consequential design work is deceptively simple. Six hex codes changed how 400M people perceived the whole product. Picking the colors wasn't the hard part. This was:
Newsletter
Conclusion
Color was the surface. Trust was the actual project. In a product used by 400M people, the insert moment is your one shot at a first impression. Fixing what users see first, before AI, before new features, before anything else, is the highest-leverage design decision you can make.
96% acceptance. 8 weeks. Zero new APIs.
That's what happens when design is grounded in user behavior, validated by data, and pushed through with conviction. The colors were the mechanism. The real outcome was restoring user confidence in Excel's ability to produce work they're proud to share.
microsoft.github.io
microsoft.github.io
developer.microsoft.com
developer.microsoft.com
storybooks.fluentui.dev
storybooks.fluentui.dev
