The Environmental Impact of Online Retailers’ Distribution Centers

Using experimental and observational data of an online retail platform, we investigate causal impact of platforms’ shipping strategies on the environmental emissions and propose new ways for the platform to positively impact environment using its AI-based features. We causally estimate that distribution centers lead to 16-30% decrease in CO2 emissions through more efficient item consolidation policies. Environmental savings are highest in metropolitan areas where existence of distribution centers decrease average emissions by 11% more compared to the non-urban areas. Next, we show that the platform could further decrease its CO2 emissions via accounting for regional demand when choosing product assortment in distribution centers. Finally, we develop a ranking algorithm that exhibits items closer to consumers’ location higher in the product ranking pages without hurting consumer experience. We show that our proposed ranking algorithm saves both platforms’ shipping costs and environmental costs without hurting consumer welfare.