Causal Forecasting for Grocery

 Hi,We are currently in the process of setting up Causal forecasting for a Grocery Retailer. This retailer has promotions running every week for almost 1500-2000 sku's out of their total 35K sku's. 

The items (or skus) on promotion are decided in advance and any item within a class or department can be on promotion on any given week.

Would like your suggestions on the best practice for setting up forecasting method. 

1. We have partitioned by Department and if we set forecasting method to AutoES at Dept. level and override specific items to 'Causal', then there will be a huge manual activity of setting up forecasting methods for around 1500 items every week through the views.