// 01

What Is a Golden Path?

A golden path is the opinionated, paved route your platform team builds so that product engineers never have to make the same infrastructure decision twice. Not a restriction. A reduction of cognitive load.

The term comes from Spotify's platform engineering work, but the idea is universal: when you build a platform, you are in the business of making the right choice the default choice. Golden paths are how you do that at scale.

// without golden paths

Every team reinvents the same deployment pipeline.

Every new service takes a week before it can ship anything. Decisions that should be made once are made a hundred times.

// with golden paths

A new service goes from idea to production-ready skeleton in under a day.

Decisions already made. Standards already baked in. The developer picks a template, fills in a name, and gets a fully wired service.

// the gap

Most companies have a platform team. Far fewer have golden paths.

That's the difference between infrastructure and developer experience. Infrastructure keeps the lights on. Golden paths let engineers move.

// 02

The Anatomy of a Golden Path

A golden path is not a single tool or a single pipeline. It's a system of five interlocking components. Remove any one of them and the path becomes a trail — possible to follow, but slow and uncertain.

// FIVE COMPONENTS OF A GOLDEN PATH Service Template Self-Service Env Paved CI/CD Observa- bility Policy Guardrails
// 01 foundation

Service Template

A pre-built, opinionated starter for a new service.

A pre-built, opinionated starter for a new service — containerised, with CI/CD already wired, logging configured, secrets management in place. The developer answers four questions and gets a repository that's production-ready before they write a line of business logic.

// 02 self-service

Self-Service Environment Provisioning

A developer requests an environment and gets one in under 15 minutes.

No ticket. No waiting. No email to a platform team that already has a three-week backlog. Self-service isn't a feature — it's the baseline expectation for any platform worth the name.

// 03 automation

Paved CI/CD Pipeline

One pipeline definition that works for 80% of services.

The pipeline handles security scanning, compliance evidence generation, and deployment. The developer doesn't configure any of it — they just push code. Extension points exist for the 20% of services with legitimate edge cases.

// 04 observability

Integrated Observability

Logs, metrics and traces wired in from day one.

Not bolted on after the first incident. Every service spawned from the golden path gets a pre-built dashboard, alerting rules, and a runbook template. Observability is not optional — it's part of the template.

// 05 compliance

Policy-as-Code Guardrails

Security and compliance checks that run automatically.

Policies are defined once by the platform team and applied everywhere. The developer gets fast feedback at commit time. The auditor gets a machine-generated evidence pack. Nobody waits two weeks for a manual security review.

// 04

How to Build One (Without Starting From Scratch)

You don't need to build the perfect golden path on day one. You need to build a useful one. Here's the five-step process we use with clients.

01
// start here

Pick your highest-frequency service type

Start with the thing engineers create most often. For most teams, that's a REST API or a background worker. One path done well beats five paths done poorly. The goal of the first sprint is a working template that a senior engineer would actually use — not a comprehensive platform.

02
// user research

Interview three engineers who will use it

Ask what decisions they make every time they start a service. Those are your path's scope. Don't ask what tools they want — ask where they lose time. The interviews take two hours. They'll save you six months of building the wrong thing.

03
// build

Build the template, not the docs

A working Cookiecutter or Backstage template beats a 40-page runbook every time. Documentation describes the path. A template is the path. Build the template first, then add just enough documentation to explain the escape hatches.

04
// non-negotiable

Wire observability in before you ship it

If a developer has to add logging themselves, the path isn't paved yet. Observability is the last thing engineers think about and the first thing they miss when it's absent. Make it non-optional by making it invisible — baked into the template, not bolted on after.

05
// measure

Measure adoption, not completion

Track which paths are used in production. If a path has 0% adoption after 60 days, find out why before building the next one. The most common reason: the path solves a problem the platform team had, not a problem the developers have.

// 05

Our Approach

We don't build golden paths and hand them to clients. We build them with the engineers who will maintain them, using the tools the organisation already has. The goal is a path that survives after we leave — not one that depends on us.

Every engagement starts with a Platform Value Map: structured interviews with developers across the organisation to identify where time is actually lost. We build the first golden path together, in a six-week sprint. By the end, the client's platform team owns it, knows how to extend it, and has the data to prioritise the second one.

97% Release overhead reduction
< 1 day To first deployment for new services
60 days To validated golden path in production