Skip to content

A process that stays visible.

Seven stages. Each has an objective, artefacts, and a reason to move. We do not skip discovery to look fast, and we do not hide in process to avoid decisions.

A quiet work table in north light.

01

Discover

Understand the work, the constraints, and the decision this product is supposed to change.

Deliverables

  • Problem framing and non-goals
  • User and operational map
  • Risk register (technical, product, organisational)
  • Recommended engagement shape

Client. Access to the people who do the work, the current systems, and the person who can decide.

Technology. None locked. We note constraints (stack, vendors, compliance) so later stages do not pretend they are free.

We only leave discovery when we can say what not to build.

02

Define

Turn the problem into a product slice with a boundary and a success test.

Deliverables

  • Scope for the first release
  • Information architecture / domain model sketch
  • Success measures (qualitative and, if they exist, quantitative)
  • Delivery plan and team shape

Client. Workshops on packaging, permissions, and the first customer or operator who must succeed.

Technology. Stack recommendation, tenancy/identity approach, and buy-vs-build for the obvious vendors.

A written slice. If it cannot be estimated, it is not defined.

03

Design

Make the hard paths visible: empty, error, permission, and the dense screens.

Deliverables

  • Flows for the primary jobs
  • Interface for the unglamorous states
  • Content and component language
  • Prototype of the riskiest interaction

Client. Reviews with the people who will use it, not only the sponsor.

Technology. Design tokens and component boundaries aligned with the intended frontend.

Engineering can implement without inventing product behaviour.

04

Engineer

Build the slice as a system: data, interface, delivery path, and observability.

Deliverables

  • Working software in a reconstructable environment
  • Tests on the expensive paths
  • Instrumentation for the journeys that matter
  • ADRs for the decisions we do not want to re-litigate

Client. Regular review of running software. Decisions in the open when trade-offs appear.

Technology. Schema, APIs, and infrastructure as code. No architecture that only exists in a slide.

The slice can be used by someone other than the people who built it.

05

Validate

Prove the slice against the jobs, the load, and the ways it will fail.

Deliverables

  • Quality evidence on critical paths
  • Accessibility and performance notes
  • Security-minded checks on authz and data
  • A list of known gaps, ranked

Client. UAT with real roles. Sign-off is a conversation about residual risk, not a ceremony.

Technology. Fix, flag, or postpone — each gap gets an owner.

We launch what we understand, not what we hope.

06

Launch

Put the product in the path of real work, with a rollback and a watch.

Deliverables

  • Release plan and communications
  • Runbooks and on-call shape
  • Hypercare window
  • Analytics and error budgets live

Client. The operational owner is named. Launch is not “engineering is done”.

Technology. Feature flags, migrations, and a practiced rollback.

The product is in use. The work becomes evolution.

07

Evolve

Improve the product with evidence, and keep the system fit to change.

Deliverables

  • Backlog informed by usage and operations
  • Quality and cost review
  • Knowledge kept in the repo, not in people
  • A decision to continue, taper, or hand over

Client. A standing cadence. No surprise invoices for “being available”.

Technology. Debt is scheduled or accepted. It is not ignored.

There is no final stage. There is an honest next one.

Start a project

Tell us what you are building.

A short note is enough. We will come back with questions, not a pitch deck.