These are some excellent points, especially around adoption. Spotify famously sees 90% adoption of Backstage internally,
but only 10% adoption externally. That means you could end up spending a lot of FTE resources for little payback. The key is to find a few specific, easy to explain use cases that solve real pain and focus on fixing those first to try and bolster cultural buy-in.
Do you have strict requirements to build vs. buy? As Paul mentioned, Backstage is essentially a code repo that you need to customize, build, and maintain. Depending on the size of your engineering team, this can become more expensive than just buying something, like
Cortex.io (transparency - for whom I run Developer Relations) or another commercial SaaS IDP.