Phantom inventory, and what makes it so scary

May 7, 2019  |  4 min read

ScaryWhen you know a product is out-of-stock and costing lost sales, you can do something about it. As the saying goes, knowledge is power. But what if you don’t know that a product isn’t available for consumers who want to purchase it?

That’s exactly what phantom inventory is: when a product is not reported as out-of-stock because the stores’ systems register that they have on-hand inventory, but in reality the product is not available. In other words, the recorded inventory is like a phantom — it doesn’t really exist on the shelf.

As a result, brands and retailers often don’t even know lost sales are happening. And that’s scary.

A multi-billion dollar problem

You might be questioning whether phantom inventory is really a problem. After all, inventory management systems are pretty sophisticated, and accurate inventory reporting is important not only for daily business operations, but also financial accounting.

However, MIT researchers found that unobserved stock-outs cause lost sales that are almost five times greater than what was previously assumed. Automatic replenishment does not kick in because the system thinks there’s still inventory available, and if the problem is not flagged, lost sales can continue on for weeks until a manual store inventory audit is performed. Phantom inventory also undermines the performance of any product promotions.

What’s more, phantom inventory’s impact can easily extend beyond immediate lost sales. The researchers concluded most measurements of known stock-out levels give a misleading impression of how a store or product is actually performing. The resulting poor sales performance could send teams on a wild goose chase trying to identify the root cause of the issue, and more dangerously, lead them to make an unnecessary change to product mix, distribution, or promotion strategy.

In addition, if these inaccurate demand signals are used to develop forecasts and plans, the problem will be propagated into the future. Just as reported out-of-stocks need to be accounted for in demand forecasting, so should phantom inventory that artificially depresses consumer demand. But as we’ve pointed out, it’s a lot harder to do so if you don’t know it’s happening.

Detecting phantom inventory

Phantom inventory can be caused by a wide variety of reasons, including both human and system errors. A separate MIT research study on the root causes of stock-outs during promotions found that when a product was not on the shelf, the cause was a reported out-of-stock only 18.5% of the time. That means over 80% of the time, the system shows product is available!

Common causes of phantom inventory include:

  • Theft (by employees or shoppers)
  • Product damage that doesn’t get recorded in the inventory management system
  • Scanning errors or issues at checkout
  • Products in the wrong place (moved by customers or stocked incorrectly)
  • Products “lost” in the backroom
  • Receiving errors

Because of the variety of causes of phantom inventory, it is very difficult to detect it by looking for the cause itself. A more effective approach is to look for symptoms that it’s occurring. The same way a doctor diagnoses a patient with the flu by looking for presenting symptoms (fever, aches, chills, etc.) vs. testing for the actual flu virus, you can diagnose phantom inventory by looking for exceptional zero product sales situations.

We define an exceptional situation as when a product is expected to be selling-through, based on historical sell-through and forecasted demand, but actual sales are zero. Because it’s unlikely sales would be zero if the product were available on the shelf, it’s a good indicator that there’s phantom inventory preventing customers from buying.

Translate phantom inventory into lost sales

Identifying phantom inventory at scale is immensely challenging because it requires forecasting sell-through at a SKU/store level to determine if sales should be zero or not, and comparing that against the latest sell-through data on a frequent basis.

For any company, there are typically an enormous number of SKU/store combinations, and very few even forecast at that level today. To do the comparison, sell-through data also needs to be collected at that same level, and ideally on a continual daily basis to catch the problem as soon as it starts happening. Finally, whenever we talk about these volumes of data, there needs to be an efficient way to sort through it all, highlight the exceptions, and prioritize the top ones that require the team’s attention.

Exception-based alerts, surfaced in a news feed or notifications like what you’re used to on Facebook, can solve the problem of highlighting all the instances of phantom inventory occurring throughout your supply chain. But even then, you need a way to prioritize them all.

That’s why translating phantom inventory into Lost Sales $ is key. When the cost of phantom inventory is put into dollar terms, it makes it easy for anyone to put it into context, prioritize it amongst all the issues fighting for their time, and drive action.  

For example, Lost Sales is a very powerful metric to drive rich conversations with retail buyers and provide proof of performance to justify larger buys. If you can point out the problem, along with an accurate representation of what should have been selling based on history, you can easily justify why larger buy orders should be made. The alternative is a buyer relying on history to build their forecast, with that history including days of zero sales of those SKUs in those locations. Retailers’ systems are not flagging phantom inventory problems, let alone adjusting for the volumes that should have been sold during the period it occurred.

Brands need to take control and lead that conversation with buyers. Look to an advanced retail data and analytics solutions to take care of the heavy lifting and dynamically identify the top phantom inventory opportunities, so you can focus on selling.

Posted by Alloy