Platform as a Product
Running an internal platform with the disciplines of a product — voluntary adoption, user research, documentation and a roadmap — rather than as a mandated standard.
The distinguishing test is simple: would teams choose it if they did not have to? A platform that survives only through mandate is not solving the problems its users have, and mandate conceals that signal.
What the product framing requires:
Voluntary adoption wherever possible, which forces the platform to be genuinely easier than the alternatives. The paved road formulation captures it: centralise the default, not the mandate, with an exception path for teams whose needs differ and who accept the ownership that comes with leaving the road.
User research. Talking to product teams about what actually slows them down, rather than building what the platform team finds interesting.
Documentation and self-service onboarding, since the measure that matters is time from arriving to first successful use.
A roadmap and a support model, so consumers can plan and can get help.
Adoption and satisfaction metrics, which are the platform's equivalent of revenue.
The failure modes this prevents: a platform built on assumptions nobody validated; a platform that is a gatekeeper rather than an accelerator; and a platform that cannot demonstrate value and is defunded in the first cost review — which is common precisely because its benefit is diffuse and its cost is a line item.