I explained that I would track metrics like assignment time, transit time, and order-to-delivery ratios. I suggested breaking the data down by micro-zones to identify specific bottlenecks and using predictive modeling to better forecast demand and position riders before peak hours hit.