For starters, how about not using a proprietary service from one of the biggest tech oligopolies which has strong economic incentives to misuse any data that passes through it? If government needs accurate tracking numbers to base its decisions on, it should roll its own tracking for those numbers.

My proposal was simply that data-driven decisions should be made based on accurate data.

Nepotism and bias aren't the alternative here. If you require that decisions are driven by data and you don't have access to accurate data, you should default to doing nothing. If anything, nepotism and bias sneak in when decisions are made despite the fact that no accurate data is avaliable.