PROCESS

We understand before we build

The same six phases run on every engagement, whatever its size. Each one has a stated outcome, which means each one is something you can hold us to.

01Discover02Design03Build04Test05Deploy06MaintainMaintenance is not the end of the workIt re-enters it

PHASE BY PHASE

What happens, and what you are left holding

01

Discovery and understanding

Nothing is built until the problem is agreed

We ask what usually gets skipped: what is actually broken, who it affects, and what it costs to leave it that way. Sometimes the honest answer is that software is not the fix, and we would rather say that now than bill for it later.

What happens

Interviews with the people who use the thing
Mapping the process as it actually runs
Scope, written down
What success looks like, agreed in advance

What you get

A written, signed-off understanding of what is being built and why, before anyone opens an editor.

02

Architecture and design

The system is designed before it is built

How the pieces fit, where the data lives, how the parts talk to each other, and what happens when one of them fails. Decided deliberately rather than discovered halfway through.

What happens

System architecture and its documentation
Data model design
The contracts between components
Security and data protection review

What you get

Architecture documents that guide the build and are still useful two years later, when nobody remembers the reasoning.

03

Build

Reviewed, tracked, and never a surprise

Work goes out in increments you can see rather than in one delivery at the end. Everything is reviewed by someone who did not write it, and documentation is written next to the work, not after it.

What happens

Delivery in increments, with a working system at each one
Review by someone other than the author
Regular updates, whether or not the news is good
Documentation written alongside the work

What you get

Working software, delivered incrementally, in a codebase the next engineer can read without an introduction.

04

Test

Nothing reaches production untested

Automated checks that run on every change, plus a person walking the journeys a customer will actually take. A release that fails its checks does not go out, and that is a structural rule rather than a judgement call.

What happens

Automated suites across the system
Journey testing, end to end
Quality gates in the release process
A person reviewing what automation cannot

What you get

A tested, stable release, and a record of what was covered.

05

Deploy

A real pipeline, and a way back

Releases go through a rehearsal environment first, run on an automated pipeline, and land with monitoring already watching them. Every deployment has a documented route back if it needs one.

What happens

An automated release pipeline
A staging environment that mirrors production
Deployment without taking the system down
Monitoring and alerting, wired to real people

What you get

A live deployment that is being watched, and a rollback path someone has actually tested.

06Loops back into 03

Maintenance

Software is not finished when it ships

It needs patching, adjusting, and occasionally rethinking as the business changes around it. We stay on as the engineering team, so the people answering for the system are the people who built it.

What happens

Ongoing engineering capacity, monthly
Fixes, adjustments, and new work
Security patching
Changes fed back into the design

What you get

A system that stays healthy and keeps up with the business, with the same people who built it still answerable for it.

ENGINEERING PRINCIPLES

Six positions that decide what the process refuses to skip

A process is only worth anything if it holds when a deadline is tight. These are the parts that do not get dropped when it is.

01

We understand before we build

The most expensive mistake in software is building the wrong thing correctly. Discovery is not a line item that can be dropped to save money, because it is the thing that saves the money.

02

Documentation is a deliverable

Architecture, decisions, and operating instructions are produced as part of the work and belong to you. Code without them is a system only one person can maintain.

03

Every project ships with tests

Automated coverage is part of the definition of done, not an upgrade. A release nobody tested is a bug report with a delay on it.

04

Designed for cloud from the first decision

We do not start somewhere cheap and migrate later. Migrating later is how a small system becomes an expensive rewrite.

05

The same standard on small work

A two-week engagement gets the same discipline as an eight-month one. Cutting corners on small work is how the habit forms.

06

We do not build and disappear

We offer ongoing retainers because software needs ongoing care. A system nobody is maintaining is a system on borrowed time.