Skip to content
Decisions

Judgment calls.

Everyone shows what they made. This is what I decided, and why. The interesting design problem is rarely the interface.

Properti2025

The 20-page data audit

The situation

CEO asked me to set up ClickUp boards to manage a major client renewal. The assumption was that better project management tooling would fix delivery.

The obvious move

Set up the boards, organise the tickets, move on.

What I did instead

Came back the same day with a 20-page data audit that reframed the entire problem. The issue wasn't project management. It was a lack of evidence-based prioritisation across the whole company.

What happened

We ended up doing something completely different and more effective. The audit became the foundation for the product operating model I built from scratch.

Full story: Building a product operating model
Atlassian2023

Returning to IC from management

The situation

Eighteen months as Design Manager at Atlassian. Team was performing, regulated industries motion was established, TIP scores were high.

The obvious move

Stay on the management track. It's the obvious progression.

What I did instead

Asked to return to IC. Not because management wasn't working. Because I missed the daily craft, and I wanted to spend the next two years on the tools rather than next to them.

What happened

Within months of returning to IC, I was promoted to Lead Product Designer. The capstone of the story: returned at the highest level, immediately performed at it.

Full story: Regulated Industries Design team
Properti2025

Not buying CPQ

The situation

Sales team was quoting manually. Spreadsheets, emails, follow-ups. The obvious move was to evaluate CPQ vendors.

The obvious move

Buy PandaDoc or DocuSign. Problem solved in a quarter.

What I did instead

Looked at what was already in the building. Stripe and HubSpot had been core to the business for five years. Both were being underused. Built the quote-to-pay flow myself using what already existed.

What happened

End-to-end system in production. No new vendor, no retraining, no data migration. The CFO didn't have to learn a new tool. The sales team didn't have to retrain.

Full story: Internal sales quoting system
SOOW2018

Hiding the blockchain

The situation

Building a consumer rewards app on Ethereum. The team wanted to lead with the technology because it was defensible and novel.

The obvious move

"We're a blockchain company." Lead with the tech in marketing, onboarding, and product.

What I did instead

Made a deliberate decision to make the blockchain invisible to users. Points just worked. The Ethereum wallet was there but nobody needed to know. The design instinct was: surface the value, hide the plumbing.

What happened

70% of test users said they'd connect their bank account on signup. No barriers introduced by the blockchain patterns. Every project I've seen since that led with "we're a blockchain company" has struggled.

Full story: Redesigning data sharing
Fathom2019

Breaking a seven-year streak

The situation

Fathom hadn't shipped a new product in seven years. Leadership was brilliant but uncertain about modern product development. The culture was risk-averse.

The obvious move

Start building immediately. Ship fast, learn fast. The lean startup playbook.

What I did instead

Slowed down. Ran a vision workshop with the founders first to align on what we were actually trying to do. Established four guiding principles. Then invested heavily in research before writing a single ticket.

What happened

46% month-over-month increase in new customer revenue. $100k MRR in Q1. Patent filed. And the team that hadn't shipped anything new in seven years now had the muscle to do it again.

Full story: Spreadsheet suffering to success
Properti2025

Identifying the PLG gap nobody asked about

The situation

Properti had only ever sold to enterprise with 12-18 month sales cycles. I was hired as Head of Product & Design. Nobody asked me to build a new product.

The obvious move

Focus on the existing enterprise product. Improve onboarding. Reduce churn. The safe bet.

What I did instead

Identified that the lack of a self-serve entry point was a fundamental go-to-market gap, not a feature request. Conceived Agent Grader as the PLG entry point and built it from scratch.

What happened

Zero to public launch in eight months. Scoring engine, SMS OTP, Stripe payments, tens of thousands of agents. A new revenue channel that didn't exist when I walked in.

Full story: Agent Grader
Fathom2020

Forecasting is a communication problem

The situation

Building a financial forecasting product. The team assumed the challenge was making numbers easier to input and calculate.

The obvious move

Build a better spreadsheet. Faster inputs, better formulas, prettier charts.

What I did instead

Research revealed the real problem wasn't numbers. It was communication. Business owners needed to share projections with advisors, partners, and investors, and the tools made that nearly impossible. We reframed the entire product around communication.

What happened

The Business Roadmap feature let people understand financial plans using a mental model they already had. 5,000+ forecasts created with minimal support. Customer feedback: "absolutely world class."

Full story: Spreadsheet suffering to success
Atlassian2024

Defaults are policy

The situation

Designing admin controls for AI features across Atlassian's enterprise products. Thousands of admins would see these settings and most would never change them.

The obvious move

Ship sensible defaults and focus design effort on the controls themselves.

What I did instead

Treated every default as a policy statement. Each one says what kind of customer you want, what risk you're comfortable with, and what behaviour you're nudging toward. We spent disproportionate time on defaults because they mattered more than the visible UI.

What happened

Enterprise AI activation became defensible. When customers asked "why should we turn this on for 50,000 users," the answers were in the product, not just in the sales deck.

Full story: Atlassian Intelligence