I have quarterly sales data for ~80 products spanning 4 years. I also have 16 different external interventions, where one or multiple of them have occured on the products on different dates during the sales period (shcematic example in figures below).
I realize that if one type of intervention occured on all the products at the same time that this could be identified using something like the method presented in this article: http://freerangestats.info/blog/2018/08/14/fuel-prices
The same event in my data can occur for different dates across different products, though. I also have several different events to evaluate.
Is there a way of evaluating one (or several different) intervention(s) across several samples if it occurs at different times?