01
STARTED WITH A UX AUDIT, NOT A FIGMA FILE
11 customer calls. 14 documented heuristic violations, ranked by support-ticket cost. Fixed the top three first. Deferred the rest to phase two with a paper trail.
W.03 — BENCHMARK · B2B SAAS · FINTECH
SUBTRACTION WITH A PAPER TRAIL.
I redesigned a feature-dense fintech project tool. 40% drop in support tickets. 2.5x longer sessions. The first redesign in this company's history that actually shipped.
THE CHALLENGE
EVERY FEATURE REQUEST HAD BECOME A SIDEBAR ITEM.
The product had grown organically over four years. Every feature request had become a sidebar item. Power users tolerated it. New customers churned in week one. Support was eating 60% of engineering time.
The solution: I didn't open Figma for two weeks. I audited. Eleven customer calls. Fourteen heuristic violations, ranked by support-ticket cost. The audit was the deliverable that unlocked the redesign.
THE APPROACH
THE FIRST DELIVERABLE WASN'T A SCREEN. IT WAS A DECK.
I owned the redesign end to end. Four engineers, one product lead, ten weeks.
The first deliverable wasn't a screen. It was a deck. Every issue in the product, ranked by how often it showed up in support tickets, with a fix proposed and an estimate beside it.
Once the team agreed on what was getting cut, the redesign took itself the rest of the way. Subtraction with a paper trail.
KEY DECISIONS
FOUR CALLS THAT SHIPPED THE REDESIGN.
02
SIDEBAR FROM 24 ITEMS DOWN TO 6
Reports, dashboards, and exports collapsed into a single workspace. Search took the place of navigation for the long tail.
03
FILTERS MOVED FROM MODAL TO INLINE
Filtering happens in the table header now, not in a popover that closed and lost the user's place.
04
SEVEN "DOWNLOAD CSV" BUTTONS COLLAPSED INTO ONE
Seven different 'Download CSV' buttons collapsed into one canonical export dialog with proper format options. Engineering agreed. Legal agreed. Nobody missed the others.
THE SHIPPED WORK
THE CLIENT'S WORDS
"We'd been told for two years the product needed a redesign. Two designers tried and stalled. Shahriar walked in, audited it for two weeks, and we had a defensible plan with engineering buy-in. The 10 weeks after that were the first redesign in this company's history that actually shipped on schedule."
HOW I WORKED
FIXED SCOPE. TEN WEEKS.
Model — Fixed scope, 10 weeks.
Communication — Two weekly syncs. Async via Slack and Loom.
Tools — Figma, FigJam, Linear, Loom, Notion.
UX audit deck — 11 interviews, 14 issues, ranked
IA rewrite
Design system, 32 components
38 final screens
Engineering handoff
Post-launch QA support
WHAT THIS MEANS FOR SIMILAR PROJECTS
A feature-dense product rarely needs more design — it needs subtraction with a paper trail. The audit deck was the deliverable that bought permission to redesign. Every team I've seen stall on a redesign was missing this step.
WHAT COMES NEXT
- Phase two: the report builder. It's the last 'feature pile' surface that didn't fit in this scope.
- Wire the audit into the support tool. Ticket counts should drive the next prioritization automatically, not manually.
- Build a public changelog. Power users want to see what changed; right now they find out by accident.