Artificial Intelligence in Reverse Logistics Management
There is no doubt that logistics is a key link in the supply chain, just as there is no doubt that it has a direct impact on customer satisfaction and, therefore, on customer loyalty and portfolio retention. Whether an item reaches its destination within the estimated time frame and in optimal condition undoubtedly makes the difference between a company with a future and one without.
In recent times, but especially over the past two years with the arrival of the pandemic, we have witnessed an explosion and consolidation of online markets, which has multiplied digital transactions. Nowadays, something unthinkable a decade ago, all kinds of products are bought and sold online, regardless of type, with a particularly significant rise in fresh goods, which require even greater speed in the transaction process.
Purchasing channels have multiplied, inventories have grown, and there is generally a high variability of product references, which has forced companies to seek solutions—solutions they have found through their alliance with Artificial Intelligence and Data Analytics. Demand forecasting systems, fleet management, load optimization, and route design are already a daily reality for businesses looking to streamline their processes while gaining in quality, profitability, and customer service, in an area of operations where a great deal of money is at stake.
If, as we are seeing, purchases have grown exponentially in recent years, becoming even more widespread in recent months due to the new reality brought on by COVID-19, it stands to reason that returns have followed the same trend. The higher the percentage of online sales, the greater the need to manage complex reverse logistics.
What is reverse logistics?
We are entering a period that, while it serves as a boost for companies closing out significant revenue figures in the final months of the year, also forces them to deal with complex and challenging returns. Black Friday and the Christmas season move thousands of products and billions of dollars worldwide, and in the end, not all customers will be satisfied. This marks the beginning of a new type of logistics management—this time in the opposite direction, that is, reverse logistics.
The return of a product from the consumer to the warehouse—for repair if applicable, reinsertion into stock, or disposal in certain cases—involves a cost far higher than that of shipping. Specifically, a return, in terms of transportation costs alone, is three times more expensive than standard outbound shipping (warehouse to customer). And these costs cannot be measured in monetary terms alone, but also in time, since they require inspecting, repairing, or restocking the merchandise, as well as completing the administrative and financial procedures for receiving the product and processing the refund.
As a result, returns are extremely costly operations, since not only is the product not sold, but costs are also incurred throughout the process (transportation, repackaging, lost sales opportunities, etc.). This is so much the case that large companies like Amazon often choose not to collect certain unwanted items and instead refund the customer directly, avoiding the costs of returning the item to the company.
This new policy, which cannot be applied to all products or implemented across the board—in order to prevent abuse—is based on a clear principle: cost reduction. But how is it determined whether or not this method should be used? Let’s take a closer look.
Artificial Intelligence and Reverse Logistics
As already mentioned, logistics management has increasingly relied on Artificial Intelligence and Big Data to make operations more efficient, and reverse logistics follows the same trend. In this case, AI simplifies decision-making by analyzing, evaluating, and determining whether or not it is worthwhile to carry out each product return operation.
Several parameters come into play here, such as shipping cost, package volume, product price, pickup location, likelihood of future resale, warehouse stock, and more. All of these data points are interrelated and define the product itself, but they are not the only factors considered. Artificial Intelligence also analyzes the buyer, looking at their purchase history with the company, buying behavior, or loyalty, for example. These parameters are equally important in determining whether the customer is encouraged to keep the purchased product and receive a refund, or whether the item is requested back to process the return.
As a result, taking into account the high costs of returns, if the product price is low, resale opportunities are limited, and there is a high likelihood of future sales with the customer—to give a clear example—the most sensible course of action is usually to minimize losses by not requesting the return of the item, allowing the consumer to keep the product in question.
However, as previously noted, this is not a practice that can be applied universally, in order to prevent fraudulent behavior. The ongoing pandemic, along with customers’ limited willingness to take items to be returned to drop-off points such as stores or post offices, has driven greater adoption of the refund-without-return policy. This approach will need to be studied in detail to assess its effectiveness, advantages, disadvantages, and the fraud rates it may present.
In any case, Artificial Intelligence currently makes it easier to make these kinds of decisions, aiming for profitable transactions with customers who are—or are expected to be—profitable as well. Its evaluation of the data is entirely objective, not based on subjective criteria, and therefore it points to the operations that will truly represent the lesser evil for the organization, since a return always results in some kind of loss.