Redesigning the Core Dashboard of SaaS Product

Problem

The dashboard had 11 years of visual debt — 30+ screens built by different people with no shared system. Users were dropping off at the core workflow screen at a 60% rate. Three different teams had built UI in completely different ways and nobody owned the problem.

My role

I led the full redesign — user research, information architecture, visual design, and frontend implementation in Next.js. Worked directly with the CTO and two engineers throughout.

Outcome

• User drop-off reduced from 60% to 22% within 6 weeks of launch

How i solved it

Conducted eight user interviews focused on user workflows rather than aesthetic preferences. The key insight:

We didn't have a design problem. We had a structure problem wearing a design problem's clothes.

Information architecture was restructured around the user's job-to-be-done rather than internal team structure. Visual implementation used feature flags enabling incremental deployment across sections while measuring drop-off metrics at each phase.

What shipped

A completely rebuilt dashboard — cleaner information hierarchy, 40% fewer clicks to reach core actions, and a consistent component system the engineering team now builds on top of.

Incremental deployment meant users experienced gradual UI familiarization, resulting in smooth adoption without a disruptive cutover.

What worked

Shipping behind feature flags and measuring impact at every step gave the whole team confidence. Settling the information architecture before touching visuals made every design decision faster and cheaper downstream. Pairing with an engineer from day one meant nothing got lost between design and build.

What I'd Do Differently

Run usability tests throughout the process instead of only at the end — a couple of issues surfaced later than they should have. Document decisions more thoroughly for whoever maintains this after me. Push for the design-system conversation earlier — some inconsistency still lingers.