Phantom inventory is one of the most complex problems to deal with for retailers because, as the name implies, it is challenging to discover. There is no single root cause or an identical manifestation between stores to identify it. Sometimes it presents itself as product availability errors, as the system indicates that the store has stock, even though no physical units are available. Other times, a stockout is alerted when in fact, there is none.

While retailers try to balance assortment variety in their stores with optimal inventory management, the discrepancy between physical and recorded stock can lead to significant complications. Considering that every time a customer faces a stockout, there is a 40% chance that the store will perceive a direct loss, it should come as no surprise that studies by the Consumer Brands Association (formerly known as the Grocery Manufacturers Association or GMA) estimate a $50 billion loss from out-of-stock issues in North America alone between 2017 and 2018.

Retailers should also not forget that lost sales are only one of the consequences of phantom inventory, affecting inventory management, customer experience, e-commerce fulfillment, and decision-making. So, are retailers forced to take the impact without acting? How can retailers address this problem?

The answer is simple: data. Phantom inventory is hard to see, so the best way to identify it is to have data and analytics solutions that leverage artificial intelligence (A.I.) and machine learning (M.L.) development to see its signs or symptoms. Algorithms are fed with sales projections, inventory level, POS sales information, and more and can determine if any product presents an anomalous behavior. An anomalous sale would be, for example, having sales of a product registered as out of stock; or not having sales of alcohol gel in the middle of the pandemic if there is stock available. Thus, the probabilistic analysis of the models will be able to raise alarms when there are signals that suggest anomalies.

The critical point for effectively managing phantom inventory comes after identifying abnormalities. The retailer must ask itself what to do with this information. What action to take now that they know that some product’s sales or turnover are not behaving as expected. We are talking about converting the data into action, into concrete activities that can effectively solve the problem on the sales floor.

In our experience, the best way to get the information down to the stores is in the form of product-specific tasks. Frogmi® offers a task management solution at SKU level, which can automatically trigger tasks against A.I. alerts for the store to check the product’s status. For example, a task can be sent to audit the product’s condition, answering some basic questions: Does it have enough stock? Are the products in good condition (undamaged, not expired)? Is it displayed correctly, following the planogram guidelines? Does it have a price tag? Is the price correct? And, if a problem is found, a new task is automatically triggered to provide the solution.

Effective phantom inventory management goes beyond traditional inventory management. It is a verification of all product’s commercial variables that will help retailers ensure shelf availability and validate that the protocols and standards defined by the company are being met. We are actively reducing the gap between what is planned and executed in every store for each item. In this way, customers will be able to find the products they are looking for, when they need them, and in the best condition, improving the shopping experience and ensuring sales.