A platform team has built substantial capability and product teams report that it slows them down. What is the diagnostic, and what usually causes it?
Show the full answer Hide the answer
The diagnostic
Measure time-to-first-success for a new team, and the proportion of platform work that is self-service versus ticketed.
If a team can get what it needs without the platform team's involvement, the platform scales. If every request is a ticket, the platform team's capacity is the ceiling on everyone else's throughput — which is exactly the bottleneck a platform exists to remove.
What usually causes it
1. The platform is a service desk, not a product. Its interface is a request form rather than an API and a template. This is the structural cause and everything else follows from it.
2. It is more restrictive than the alternative without being faster. Teams accept constraints in exchange for speed. A platform that constrains and does not accelerate is a tax.
3. Missing variation points. A team with a slightly different legitimate need has no way to adapt within the road, so it leaves entirely rather than partially. Requests are frequently a signal of rigidity, not of neediness.
4. Abstractions that leak under pressure. The platform hides complexity until something fails, at which point the team must understand both their system and the platform's — and cannot debug either.
5. Unpublished or unmet SLOs, so teams cannot plan around the platform and escalate instead.
The reorientation
Adoption is the success metric, not delivery. A capability nobody uses has not been delivered.
Concretely: research what teams actually do rather than assuming · make onboarding the primary metric · make migration the platform team's job, since telling teams to migrate competes with their roadmap and loses · publish and meet SLOs · and add explicit variation points.
The categorisation that reveals the fix
Sort the request queue into three: should be self-service (build the automation) · indicates a missing capability (a product gap) · genuinely bespoke (rare and appropriate).
If the first category dominates, it is an automation gap rather than a capacity problem — and adding platform engineers will not fix it.