Home / Approach

Map it, scope it in writing, then build.

Most failed software projects were specified by people who had never watched the work being done. We fix that order.

01

Map

We sit with the people doing the work and write the real process down, step by step, with timings. Not the documented process — the one that actually happens, including the spreadsheet nobody mentions in meetings.

  • You get the map whether or not we go further.
02

Scope

One to two weeks, paid. It ends in a written specification, an architecture, a build plan, and a price for the first phase. The document is yours outright and is useful even if you hand it to another team.

  • No build starts without one.
03

Build

Two-week iterations against a board you can see. Every iteration ends with working software deployed somewhere you can use it, and a review session where scope can change.

  • You talk to the engineer writing the code.
04

Measure

We take the baseline from the mapping phase and re-measure against it. Cycle time, manual steps removed, rework rate. If a number did not move, that goes in the report too.

  • Agreed before the first commit, not after.
05

Operate

Deployment, monitoring, dependency upgrades, and incident response, with a named engineer who already knows the system. A monthly written note on what changed.

  • Monthly retainer, cancel with notice.

What we measure

Baseline first

Cycle time

How long one unit of work takes end to end. The number that says whether the process changed or the work just moved somewhere else.

Manual steps removed

Every hand-off a person no longer performs, counted and costed against the hours it used to take each week.

Rework rate

Work that had to be done twice. Automation is only a gain if it does not quietly create a new class of failure.

How we build it

Engineering
  • BackendJava, .NET, Python, or Node — chosen for the team that will maintain it, not for novelty.
  • DataPostgreSQL, SQL Server, or MongoDB, with versioned migrations reviewed like code.
  • FrontendReact and TypeScript, accessible to WCAG AA, built against a design system rather than page by page.
  • InfrastructureContainerised and declared in Terraform, so the environment can be rebuilt rather than remembered.
  • TestingBehaviour covered by automated tests before a change ships, not after an incident.
  • HandoverRunbooks, architecture notes, and access your team owns from day one.
Stack
  • Java
  • .NET / C#
  • Python
  • Node.js
  • TypeScript
  • React
  • PostgreSQL
  • SQL Server
  • MongoDB
  • Docker
  • Terraform
  • AWS

Start with the map.

The first call is 45 minutes with the person who would build it.

Start a project →