The signal is not team size. It is the queue.
Developers wait on a platform team to get environments, namespaces, databases, credentials or pipelines, and that queue is now the constraint on delivery. Many teams, many services, more than one cloud or more than one Git provider, and a platform or infrastructure team that already exists and is underwater.
- The platform team has become a ticket queue and everyone knows it.
- Onboarding a new team means a human walking them through undocumented steps.
- Several teams have built parallel half-platforms that now need reconciling.
- A cloud migration or a merger left two ways of doing everything.
- Leadership has been sold on platform engineering and wants it.
If none of that is your situation, the honest answer is that a platform is the expensive way to get what you actually want. SDLC Golden Paths is the lighter answer, and it solves this for a fraction of the cost.
The most valuable outcome is often the finding that you do not need one
We look at where the delivery time actually goes, what developers are really waiting on, what already exists that can be reused, and whether the self-service bottleneck is severe enough to justify a platform.
Where the delivery time goes
Measured against what developers are actually waiting on, rather than where it is assumed to go.
What already exists
Half-platforms, internal tooling and conventions that can be reused rather than replaced.
Whether the bottleneck justifies a platform
The question the assessment is really answering.
A written recommendation
Which may be a recommendation against. We would rather sell the smaller engagement than the wrong one.
What gets built
Self-service provisioning
For the things developers currently wait on — environments, namespaces, databases, credentials, pipelines.
A service catalogue and developer portal
The way in, and the record of what exists.
Golden paths as the content
The paths are what runs on the platform. A portal without paths behind it is a catalogue of links.
Policy as code
So controls are enforced in the pipeline rather than in review.
Observability and cost visibility
Wired in from the start rather than added once the bill arrives.
We will not build you a bespoke platform
A few of our engagements grew from golden paths into bespoke platforms, and we steer away from repeating that. A bespoke platform becomes a maintenance liability for whoever inherits it, and that is usually the client. Standardising on a foundation we already maintain means you are not the only organisation in the world running your platform.
Your cloud accounts
A platform always runs in your own cloud accounts, and you own what is built.
Your cloud bill
You pay your provider directly. No resale margin on cloud spend — you keep your own commitments, discounts and negotiated rates, and you can see exactly what the platform costs to run.
Not included in the fee
Cloud spend is yours and sits outside what we charge. Better said here than discovered later.
What you are choosing between
Buy a product
Backstage, Port, Humanitec. A product is a container, and someone still has to decide and build what goes in it. That someone is the work.
Build in-house
Common, and frequently the right answer at this size. The failure mode is not capability — it is that the platform team never gets uninterrupted time.
A large systems integrator
Which will staff it with more people for longer.
Do nothing
And keep absorbing the queue.
Most consultancies will happily build a custom platform and bill for every hour of it. Declining costs us billable hours, which is exactly why it is worth stating: the assessment can conclude that golden paths solve the problem for a fraction of the cost, and we would rather sell the smaller engagement than the wrong one.
The questions we get asked
This is going to take a year and cost a fortune.
The most common fear, and usually a correct one. The answer is a first phase with a defined end and something working at the end of it, decided by the assessment rather than assumed before it.
We tried this and nobody used it.
Adoption is the failure mode for platforms just as it is for golden paths. Netflix’s own experience is the useful reference: a portal that only offers visibility is not compelling enough to pull developers off their existing workflow. What drives adoption is covering the whole workflow, not the first step of it.
Isn’t this just Backstage?
Backstage is a portal. A portal without paths behind it is a catalogue of links.
Our developers do not want another abstraction.
Frequently true, and often the signal that SDLC Golden Paths is the page you should be reading.
Will we depend on you afterwards?
It runs in your cloud accounts, you pay your own cloud bill, and you own what is built.
Book the assessment
Framed honestly: it is an assessment that might talk you out of the project. Its output is a written recommendation, and that recommendation may be against building a platform at all.
- Where your delivery time actually goes, and what developers are really waiting on
- What already exists that can be reused
- Whether the self-service bottleneck is severe enough to justify a platform
- A written recommendation — which may be a recommendation against