Whilst we wait for Manuel to come back online I would just like to say that before (baseline) and current metrics are only required to justify what you’ve spent so far on the initiative.
If you’re already in the middle of a platform engineering initiative, and without those BEFORE metrics, it might be good enough to forecast what a future/continued investment could achieve. That requires some canvassing of your dev org to understand the problem space or opportunity for improvement. If you can start to assign lost developer time to certain sub optimal (wasteful) processes eg ticket processing, handovers, discovering information, manual processes, wait time etc, then you could be on your way to connecting ROI to platform eng.