Skip to content

Our Core Principles

These eight principles are the standard we hold ourselves to on every engagement, whatever the discipline. Our Engineering, Product Delivery and QA practices each own some of them and make sure they show up in client delivery.

01 Be informed

We understand the problem before we commit to a solution.

Clients often arrive with a solution already in mind, and the written requirements are only one view of what they need. Time spent understanding the goal, the people affected and what's in the way costs far less than building the wrong thing will.

02 Customer outcomes define success

We measure success by the difference we make to customers, not by the volume of features we ship.

A service that ships on time but changes nothing for the people using it hasn't succeeded. We agree the difference we're aiming for early, so every scope decision can be tested against it.

03 Quality built in from the start

We agree how we'll know something works before we build it.

Every engagement agrees who owns the test approach, and the whole team works to it. We agree measurable acceptance criteria and non-functional requirements at the start, so everyone builds and tests to the same target.

04 Build secure, resilient systems

Security and resilience are decisions we make at the start and keep making for the life of the service.

Security added late is expensive, disruptive and rarely complete, and much of what we build holds public data and public trust. We design for the threats and load we expect, then keep the service secure and resilient for as long as it runs.

05 Deliver in safe, reversible steps

Every release has a way back, agreed before it goes out.

We deliver in small, frequent changes, so each one is understood and tested as we go and every release is one we and the client can trust. Each release has an agreed, rehearsed way back, so if something goes wrong we recover quickly.

06 Anything repeated should be scripted

We automate the work we repeat, wherever it pays back the effort.

Pipelines, infrastructure, tests, security checks and recurring delivery tasks all run better when they're scripted, because they run the same way every time and anyone on the team can run them. That makes services cheaper to run, safer to change and easier to hand over.

07 Use AI deliberately

We use AI where it helps, and it's held to the same quality, security and data standards as anything else.

We use AI where it makes the work better, within the client's AI policy and our own data rules, and a person reviews anything it helps produce before it reaches the client. The same goes for AI we build into client services: it meets the client's policy and is tested and secured like every other part of the service.

08 Leave clients able to run what we build

Whoever runs the service next, whether that's the client, CDS or another supplier, can run it easily, with the support they need.

A service only its builders can run is a risk to the client, however well it was built. We document as we go and plan handover from day one, so support can move to the client, to us or to another supplier without a cliff edge.

Who owns each principle

Every principle applies to every engagement. Ownership means a practice makes sure its principles are reflected in how we deliver, working with the delivery and technical leads on each account.

Principle Owning practice
01 Be informed Product Delivery
02 Customer outcomes define success Product Delivery
03 Quality built in from the start QA
04 Build secure, resilient systems Engineering
05 Deliver in safe, reversible steps Engineering
06 Anything repeated should be scripted Engineering
07 Use AI deliberately Engineering
08 Leave clients able to run what we build Product Delivery, with Service Management

Find out more about each practice in the Practices section.