Twenty-four years building for national broadcasters, federal departments and global consultancies — long enough to have seen how it goes wrong. You get that read on every decision: what to do, why, and the trade-off nobody else mentioned.
A documented slice of that work is dated and attributed in the career archive—not all of it, but enough to show the range.
A rebuild, a platform migration, a grant with a deadline attached. The decisions that determine whether it works are all made before anyone writes code, and they are the expensive ones to get wrong. Discovery is where they get caught.
Discovery & Strategy AfterThe developer stopped answering. The site is compromised, or slow, or nobody left knows how it was built. Inherited work is most of what arrives here — and none of it is a lost cause until someone has actually looked.
Support & RescueBoth doors lead to the same standard. Anything built here, or taken over here, is built to WCAG 2.1 AA — audited, keyboard-tested, and gated in the build rather than checked at the end. That is why accessibility leads the list below instead of sitting at the bottom of it as an add-on. How that practice works →
Built on a WCAG 2.1 AA audit and remediation practice developed for enterprise clients at EY. Audits that name the specific failure and the specific fix — not a score out of a hundred.
The questions that decide whether a project is worth doing come before the first line of code, not after it. You leave with a written recommendation you could hand to any developer — including one you hire somewhere else.
Interfaces and the systems behind them. Built to WCAG 2.1 AA as a matter of course, because a product that excludes people isn't finished.
Launch is not the end of the engagement, and work somebody else walked away from is still welcome. Compromised sites, orphaned codebases, portfolios in transition. Stabilising and rebuilding are different bills, and you see both before choosing.
The best work happens when strategy and execution are inseparable. Every line of code should serve a purpose. Every design decision should solve a problem.
Discovery isn't a checkbox. It's where transformation begins. Understanding your business, your users and your goals comes first. Deciding what to build for them comes second.
This is a collaboration, not a vendor arrangement—from kickoff through launch and long after. That extends to systems someone else built and walked away from: inheriting a project is a normal way to begin here, not a favour granted on request.
AI tooling is part of how the work moves faster here, and saying so plainly beats pretending otherwise. It compresses the work; it doesn't supply the judgement about what to build, what to refuse, or what will still stand up in five years. It still takes a pilot to fly the plane—and twenty-four years of flying is what you're actually hiring.
Quick fixes don't cut it. Systems are architected to scale and code is written to stay maintainable—and WCAG 2.1 AA is in place from the first commit, because retrofitting accessibility costs far more than designing for it.
53% Faster, measured against a 30-day baseline
Support & Rescue
A security and performance audit across a twenty-year hosted portfolio found vulnerabilities in every property. Thirty days after remediation: spam signups from 200 a day to zero, bot support requests gone, performance up 53% — no downtime taken to get there.
Read the case study
PDF → Web AODA compliant by design, and maintained in-house without a developer
Discovery & Development
The Prison Visitation Guide was a static PDF. It came back as a searchable web application with custom data modelling and a tailored CMS — accessibility designed in from the start, not retrofitted.
Read the case study
17K Daily bot attacks filtered before they reach the origin server
Support & Rescue
An emergency response to a compromised site under sustained attack. Backdoors closed at source, infrastructure hardened, staff logins restored — and the saves that had been failing at server timeout stopped losing data.
Read the case study
3 months Of build time saved by going no-code
Strategy & Development
A full launch package for a coffee shop startup: brand identity, Square POS across retail and restaurant, an online storefront, two marketing grants secured, and a 30-day launch campaign.
Read the case study
Upstream Accessibility moved into the design stage, not the pre-launch scramble
Discovery & Strategy
An ongoing audit practice for enterprise clients: code-level findings evidenced in the markup, translated into guidance the delivery teams could act on, with AODA training pushed back to the designers.
Read the case study
8 years A continuous partnership, through the agency's transition
Support & Rescue
Ten to twelve legacy WordPress sites carried through an agency's change of direction — documentation, maintenance, direct client contact, and an infrastructure audit that halved the hosting bill.
Read the case studyA documented selection—76 projects, 80 technologies, 24 years—from early WordPress builds to the applications shipping today. Not everything built over those years is in here. What is, is dated, attributed and specific enough to check.
The Variables team became an extension of our agency, managing our legacy portfolio seamlessly during our business transition. Their technical expertise and client communication skills gave us confidence to evolve our focus without compromising service quality.
Have a project in mind, or something that's gone wrong and needs a second pair of eyes? Book a 20-minute call and get an honest answer about whether this is the right place for it.
Or send a message — a reply comes back within one business day.