Reducing Packaging Waste

How big of a box is needed for a given product? Is an Amazon box needed at all? To find out, we use machine learning algorithms to arrive at the best possible packaging choices for deliveries. That means identifying which products don’t need packaging at all, and which smaller products are suitable for envelopes or mailers. In cases where the protection of a box is needed, algorithms help us continuously optimize box choices to fit our ever-changing catalogue. Computer-aided engineering also helps us redesign boxes to use less material while making sure customer orders are protected. These improvements reduce volume per shipment, which means less unnecessary packaging and more efficient use of all forms of transportation, reducing carbon in small amounts to create a larger impact across our network.

We also use machine learning to help target products where even small packaging improvements can have significant impacts on reducing waste. We developed a machine learning model to identify liquid products with the highest average rates of customer reported damages. We then subjected those products to extensive additional testing at our Amazon Packaging Lab, where we simulate a package’s journey from the manufacturer to the customer.

Through this process, we’ve teamed up with companies like Rieke—a packaging solutions firm that makes dispensing systems for personal care, food, and healthcare products—to provide data and packaging insights on many of the conventional dispensing systems that are used in millions of products on Amazon. By strengthening materials, creating new locking mechanisms to prevent pumps from twisting open, and sealing potential leak points, Rieke’s team has created dispensers specially designed to withstand the journey to the customer. These new and improved designs include trigger sprayers common on household cleaners and personal care products, pumps for lotions, and pumps for foams.