Platform engineering: when did your internal dev portal actually pay off?

We're 8 months into building an internal developer platform (IDP) with Backstage. Current adoption: 3 of 14 teams have migrated their service catalogs and CI/CD pipelines. The rest are still on the legacy Confluence + Jenkins + manual infra request flow. The pitch to leadership was: reduce onboarding time from 2 weeks to 2 days, standardize deployment patterns, give PMs visibility into service ownership. The reality so far: - Onboarding: down to ~5 days (still not 2 — credential provisioning and access reviews are the bottleneck, not the platform) - Deployment standardization: working well for new services, but migrating brownfield services requires touching 15+ config files per repo - Visibility: PMs love the catalog, but the 'cost per service' dashboard shows stale data because FinOps hasn't tagged 60% of AWS resources What made the difference for teams that got their IDP to critical mass? Was it a mandate, a killer feature (e.g., 'one-click ephemeral environments'), or something else? Specifically interested in: - The adoption inflection point: what % of teams needed to be onboarded before network effects kicked in? - The 'carrot vs. stick' question: did you force migration or make the platform objectively better than the legacy path? - Metrics that actually mattered to leadership (not vanity metrics like 'number of templates')