Most design relationships begin and end without the designer ever really understanding the problem. A brief gets written. Mockups get delivered. Someone signs off and moves on. The work that ships has the surface of design but none of the depth that comes from having lived inside the business for a while.
This methodology is the opposite of that. Three phases, each named for what actually happens in it. The point isn’t to follow them in lockstep — it’s to never skip the first one.
Signal — study the product before designing it
Before opening Figma, before writing a brief, before agreeing on what to fix first, I use the product. Then I use the parts of the product that internal operators see. Then I sit with the team and walk through what I can’t see from the outside.
The questions are always the same:
- Where do users get confused?
- Which features are people not reaching, and why?
- Which screens have drifted furthest from the rest of the product?
- Which language does the product use that doesn’t match how the business actually talks about itself?
- What’s the gap between what the product is and how it comes across?
What comes out of Signal isn’t a redesign brief. It’s a diagnosis. A sorted list of where the design is costing the business something — in trust, in conversion, in time, in user frustration. Everything that follows is calibrated against that list.
The point of Signal is to make sure we’re fixing the right things in the right order. Most design work goes wrong in the first week, when everyone agrees on the wrong priority.
Form — redesign the parts that matter most
Form is where ideas become decisions. Brand, product, and website get shaped based on what Signal surfaced. The biggest problems get fixed first — the homepage that’s costing trust, the onboarding that’s losing users, the screens where the product feels least like itself.
The cadence is async by default. Walkthroughs over Loom. Iterations over Figma. Decisions that need a real conversation get a call; everything else stays asynchronous so the team keeps shipping its own work without waiting on me.
Some Form work ships quietly — small, continuous improvements that the team rolls in week by week. Other Form work saves into moments of visible transformation: a relaunched site, a new identity, a redesigned core flow. Both modes matter. The first compounds; the second resets the trajectory.
Depth — stay long enough that decisions get faster
After three months, the relationship starts to change shape. After six months, the team has internalised most of the system. After a year, the shift is visible: design decisions that used to require a meeting now happen at the system level. The team ships new screens that look like the rest of the product without me drafting them. Engineers stop asking which colour to use because the system has already decided.
Depth is what most design engagements never reach. It’s the reason this model exists. The work that compounds is the work I do in year three, not the work I did in week one. By that point I know the team, the users, the constraints, and the standard. I make faster, sharper decisions than a new designer ever could on the same problem — not because I’m a better designer, but because I’ve had the time to understand the specific situation.
The framework underneath
Signal/Form/Depth is the operating shape. The framework underneath is older than the studio, and applies whether or not anyone calls it methodology.
Two traditions, treated as one:
- Lou Downe’s 15 principles of good service design — rules calibrated to the reality that services are used by stressed, distracted humans with incomplete knowledge. Easy to find. Clear in purpose. Set the right expectations. Enable the user’s actual outcome. No dead ends. Usable by everyone, equally. Easy to get human help.
- Soviet and NASA engineering doctrine — brutal simplicity, fail-safe defaults, design for the least-trained operator, checklists as cognitive infrastructure. Both traditions arrived at the same conclusion from opposite directions: design for the worst-case user, and the best-case user benefits automatically.
Combined, they sit underneath every decision the studio makes. The product should be findable using the words the user actually uses. Every error should have a forward path. Inclusion is structural, not decorative. Make the right thing the easy thing. None of these are aesthetic positions — they’re reliability positions, applied to design.
Why this is different from how design usually gets done
Different from a freelancer
A freelancer delivers and disappears. The next project starts from scratch with a new freelancer who has no context. None of the depth that makes year three valuable ever accrues. That’s the model most engineering-led teams have already tried, and it’s the one this is built to replace.
Different from an in-house designer
An in-house designer gets buried in requests. They can’t see the forest because they’re standing inside it every day. They execute what’s asked and rarely push back on what should be built differently. An external partner has the distance to challenge the brief; an embedded partner has the context to challenge it well.
Different from an agency
An agency optimises for delivering output. The strongest senior people are on the pitch and the most junior people are on the work. The relationship ends at a deliverable. None of those things are wrong — they’re just the wrong shape for an engineering-led startup that wants design decisions made by someone who understands the business.