Process

How I actually work.

Roughly the same six steps run through everything on this site. It's not rigid — a quick project and a six-month one go through the same steps at very different depths. Here's the honest version, minus the process-diagram theatre.

01 / 06

Discover

I talk to users, watch people do the thing they're stuck on, read support tickets, and dig through whatever data exists. I try not to open Figma until the problem feels real instead of theoretical.

Activities

Stakeholder interviews
User interviews & shadowing
Analytics deep-dives
Competitive teardowns
Support ticket mining

Deliverables

Research findings
Personas
Journey maps
Problem framing

“Almost every project I've done, the brief and the real problem were two different things.”

02 / 06

Define

This is where the pile of notes turns into a few clear points. I write down what we're solving and what 'better' means here, so that later, when everyone disagrees, we have something to point at.

Activities

Affinity mapping
Insight synthesis
Design principles
Success metrics definition
Scope negotiation

Deliverables

Insight report
Design principles
Metric targets
Opportunity map

“If we can't agree on the problem, no amount of pretty UI is going to save us.”

03 / 06

Ideate

I sketch the flows and the information architecture first — often on paper. I like having a few genuinely different directions on the table so the choice is a real one, not just my first instinct dressed up.

Activities

IA restructuring
User flows
Crazy 8s & sketching
Concept evaluation
Team workshops

Deliverables

Information architecture
User flows
Concept directions
Chosen direction rationale

“My first idea is rarely my best one. It's just the loudest.”

04 / 06

Prototype

Wireframes with real data first, then high-fidelity once the bones are right. How polished it gets depends on the question — if I'm testing whether a layout makes sense, grey boxes are fine; if I'm testing trust, it needs to look real.

Activities

Wireframing with real data
Design system application
High-fidelity UI
Interactive prototyping
Motion design

Deliverables

Clickable prototypes
UI specifications
Design system contributions
Motion specs

“Greeked 'lorem ipsum' hides all the problems real content would show you.”

05 / 06

Validate

Usually a handful of moderated sessions, sometimes an unmoderated study or a quick A/B test if the stakes are high. I set up a couple of real tasks and time them, so 'this is better' has a number behind it.

Activities

Moderated usability testing
Benchmark task suites
Unmoderated studies
A/B test design
Accessibility audits

Deliverables

Test reports
Benchmark comparisons
Prioritised fixes
Go / iterate decision

“People are kind in interviews and honest with their cursor. I trust the cursor.”

06 / 06

Ship

Clean handoff with the boring states spelled out, checking the build against the design, and agreeing up front on what we'll measure. Then the part I care about most: seeing the numbers after launch and fixing what's off.

Activities

Annotated handoff
Design QA on staging
Instrumentation planning
Launch support
Post-launch iteration

Deliverables

Handoff documentation
State matrices
QA reports
Impact readouts

“A design isn't done when it looks good in Figma. It's done when it works in someone's hands.”