Industry Insights

Why Perpetual Inventory is a Perpetual Failure in Fresh

Industry Insights

Why Perpetual Inventory is a Perpetual Failure in Fresh

Fresh, healthy food is a huge driver of sales in every grocery store, and customers have high expectations for the quality of the fresh foods they buy. But maintaining a profitable, sustainable, and high-quality fresh department isn’t easy.

Fresh department data is messy, inconsistent, and item-specific and most legacy inventory systems rely on perfect data to maintain accuracy. This means that legacy technology simply doesn’t work in fresh categories; because when inventory, demand planning, and ordering processes rely on perpetual inventory, stores end up with consistently incorrect data that builds on itself. And that creates a perpetual cycle of inefficiency, lost profit, frustration and food waste.

What is a perpetual inventory system?

In grocery retail, perpetual inventory (PI) calculates each item’s assumed in-store item count by taking the store’s shipment quantities and subtracting item-level sales, as well as any scan-outs and weigh-outs. But despite the constant attention PI requires, research from Retail Insight shows that 56% of grocery store perpetual inventory records are inaccurate—and that’s mostly data from center store, where items are shelf-stable with consistent supply and demand..

Produce departments in particular deal with wide variability across SKUs, items, and categories that make it difficult to keep track of what’s been delivered, sold, thrown away, or—quite literally—evaporated. So when fresh departments attempt to use technology that relies on PI, the system underperforms, quality goes down, and customers stop coming back to your stores.

Why do perpetual inventory models fail in fresh?

Data must be perfect for PI to work—an impossible feat in fresh

In fresh, understanding how much food needs to be ordered isn’t just about counting boxes—it’s about understanding the dynamics of each product that’s on the shelf. Because fresh food is highly variable in perishability, packaging, seasonality, size, and weight (to name a few), writing orders with systems that aren’t built for these idiosyncrasies leads to unbalanced inventory and stockouts. And unlike center store items, safety stock isn’t an option for perishables.

Consider the classic romaine lettuce—a staple fresh food that’s actually 96% water! Sitting on the shelf with air circulating around it all day, lettuce begins to evaporate. And unfortunately, perpetual inventory systems don’t understand this dynamic. The system only knows that a certain amount of lettuce came into the store and it automatically assumes the same amount left. This leaves the PI system unable to account for the product loss when calculating quantity or order recommendations. 

While this product loss has a marginal impact on consumers, every molecule of evaporation cuts into margins for the retailer. These overstated inventory numbers are amplified across the entire grocery chain and create perpetually bad data that puts fresh teams in a no-win situation.

Perpetual inventory models can’t adjust to the messy, variable, and often inaccurate data that comes from fresh

Even the most tech-enabled retailers lack a solution that’s agile to the everyday, real-time dynamics of fresh food, and those systems definitely can’t account for factors like misscans or sudden changes in demand.

Let’s take a look at the mango, a favorite fruit of millions, including most of us at Afresh. Not only are there seven possible varieties of mango a customer could find on the shelf, there’s also the variable of organic vs. non-organic. A PI system mistakenly subtracts inventory from the wrong item (i.e. conventional instead of organic) which throws the whole inventory position off for both items. Legacy systems lack the intelligence to identify these data inaccuracies, ultimately impacting order quantities down the line. With at least 22% of consumers purchasing mangos every year, common misscan mistakes can become costly. Read more about this in Fresh Files: The Case of the Missing Mango.

Maintaining accurate inventory data requires a massive labor investment and creates friction

Aside from the huge IT investment it takes to implement a system like this, the manual labor that goes into maintaining PI is enormous and wasteful. Team members waste hours of their day just to ensure the tech they’re required to use is even marginally useful. Here’s some of the work they take on:

Any realistic produce manager knows it’s impossible to track every single banana or bean that counts as shrink, so no matter how hard in-store and IT teams work, the data still gets thrown off. In this era of high turnover, grocers can’t risk alienating employees with cumbersome tech and fruitless tasks.

Store teams can’t trust the data and they don’t like using the tech

One of the most frustrating parts of anybody’s job is being forced to use technology or follow a process that isn’t effective and causes more work in the end. Despite their laborious efforts, fresh department employees still end up with error-ridden orders. Store teams simply can’t trust recommendations, data, or platform reliability for PI-based systems. And lacking a reliable source of truth, order writers eventually abandon the tech and turn to guesswork that’s sometimes just as good—if not better—than the technology itself.

In the image below, a PI-based system has calculated there are -57.33 okra in stock, which are set to be ordered automatically.

Fresh food drives foot traffic, customer loyalty, and unique experiences that we can’t get anywhere but our local grocery store. Without fresh-first technology, stores can’t meet the demand for high-quality fresh food and environmental sustainability that customers are calling for…and that means profitability goes down.

Afresh’s approach solves the decades-long perpetual inventory problem

Rather than relying on a rigid PI calculation that’s tedious to maintain and everyone knows will be wrong, the Fresh Operating System provides an entirely new (and much more effective) solution: We leverage proprietary machine learning to determine an item’s Probabilistic Inventory. This unique approach takes into account an item’s perishability, does not require any scan outs or weigh outs, and allows for an inherent level of uncertainty. When there is an inevitable discrepancy in the data, our system flags it to the store team to provide input only during the normal order writing process.

Here’s how it works in practice: Let's say there were 10 cases of conventional bananas on hand on Monday. But over the course of the day, a large group of customers scanned their organic bananas out as conventional at self checkout. Since Afresh's AI understands to a close degree how many bananas are likely to be sold on any given day—and it knows how much was delivered vs. how much has been sold—the system recognizes that there is too great a discrepancy for it to make a good recommendation. Instead, it will flag the banana inventory for review by a store team member.

In fresh, human insight is essential. The Afresh app keeps employees engaged throughout the order writing process by combining AI with on-the-ground feedback from your store teams—all without requiring excessive admin tasks to maintain the data. This approach results in dramatically more efficient store operations and unparalleled order accuracy that maximizes profits and ensures the freshest food for your customers.

Explore Afresh AI in our whitepaper: Putting AI in Grocery Aisles

This year, 56% of retailers plan to expand fresh offerings and 60% cited fresh produce as the highest growth area for their stores. Without an AI-driven solution in fresh, those stores won’t achieve these goals. We’ve built the first-ever Fresh Operating System to help grocery retailers reach their goals, give customers better product, and future-proof stores from whatever disruptions might occur next.

Want to learn more about the unique challenges our fresh-first approach overcomes? Request a live demo and we’ll walk you through it!

About the author

Related Stories