Inventory management is one of the pillars of retail. Too little stock leads to stockouts and lost sales. Too much ties up cash, takes up space and increases the risk of unsold goods.
Good inventory management doesn't mean stocking as much as possible. It means having the right quantity of each product at the right time, while keeping recorded stock as close as possible to what's actually on hand.
Several metrics and methods help you get there: they let you track stock movements, anticipate demand and plan restocking.
Understanding the inventory cycle
A store's stock changes constantly. It rises when goods are received and falls with sales, transfers, returns to suppliers, losses, breakage and other recorded outflows.
The principle seems simple:
Opening stock + inflows - outflows = theoretical stock
In practice, though, discrepancies can appear between theoretical stock and the quantities actually on hand. Good inventory management therefore rests on two complementary habits:
- recording every stock movement correctly;
- regularly checking that the data matches reality.
Which inventory metrics should you track?
Each metric answers a different question.
Available stock
This is the quantity considered available for sale. Depending on your inventory system, it may take into account physical stock, reserved quantities, customer orders or other movements already recorded.
Inventory turnover
Inventory turnover measures how quickly stock sells and is replenished over a given period. High turnover can signal a fast-selling product; low turnover can point to a slow-moving, seasonal or overstocked item. For the full calculation, see Inventory turnover: how to calculate and improve it.
Stock coverage
Stock coverage estimates how long available stock will last if sales continue at the current pace. It's especially useful for spotting items that are close to running out.
Inventory value
This is the financial value of the goods you hold. For simple tracking, you can estimate it from quantities on hand and their cost. For accounting purposes, however, valuation depends on the method the business uses.
Dead stock
An item that sits in stock for a long time without a single sale ties up cash and space. See Dead stock: how to identify unsold items.
Actual vs. theoretical stock: why do discrepancies appear?
Theoretical stock is the quantity your system calculates from recorded movements. Actual stock is what's physically on hand. The two can differ for many reasons:
- breakage;
- theft or unexplained shrinkage;
- data entry errors;
- unrecorded movements;
- errors when receiving goods;
- misplaced products;
- counting errors.
A physical count lets you compare the stock actually on hand with theoretical stock and correct any inventory discrepancies you find. The more regular these checks, the more reliable the data you use for restocking and analysis.
Minimum stock, safety stock and reorder point
If you wait until a product runs out to reorder it, it's usually too late: sales keep coming in between the moment you place the order and the moment the goods arrive. Minimum stock is the quantity needed to cover sales during that lead time.
Put simply:
Minimum stock = average daily sales × restocking lead time (in days)
This calculation doesn't protect you against the unexpected, though: higher-than-expected sales, supplier delays or incomplete deliveries. That's why you can add safety stock. The reorder point (or reorder threshold) is then the stock level at which you should prepare a new order.
In its simplest form:
Reorder point = minimum stock + safety stock
Set the reorder point for each item or group of similar products. An item that sells ten times a day from a supplier that delivers within a week isn't managed the same way as a product that sells twice a month and is available within 48 hours.
Physical inventory counts: keeping stock data reliable
Even with an accurate inventory system, recorded stock can gradually drift away from reality. A physical count measures that gap.
Annual physical inventory
This means counting all or most of your stock on a given date. It gives you a complete snapshot, but it usually takes a lot of organization.
Cycle counting
Cycle counting, also called rolling inventory, means regularly counting a portion of your stock instead of waiting for a full inventory. Items can be grouped by importance, value, turnover or risk of discrepancy, so the most sensitive products get counted more often.
This approach helps you catch anomalies as you go and keep theoretical stock closer to reality.
How do you plan restocking?
The reorder point tells you when to restock, but not how much to order. The order quantity can depend on several factors:
- sales velocity;
- available stock;
- orders already in progress;
- supplier lead time;
- target safety stock;
- sales forecasts;
- seasonality;
- supplier pack sizes;
- minimum order quantity (MOQ);
- target stock level.
Here's an example. An item is approaching its reorder point, which signals that it's time to think about restocking. But if a delivery is already on its way, basing the new order on current stock alone would be a mistake. Likewise, an item sold individually may have to be ordered in cases of 12.
Restocking therefore has to combine stock data with real-world supply constraints. We cover the full cycle, from calculation to purchase order, in our article on supplier orders.
Factoring in supplier lead times
Two products that sell at the same pace don't necessarily need the same stock level. If one supplier delivers within two days and the other within ten, the stockout risk isn't the same. Supplier lead time directly determines when you need to place the order.
Ideally, use the lead time you actually observe, not just the one the supplier quotes. If a supplier promises three days but regularly delivers in five, your calculation should reflect those five days.
Avoiding overstock
Inventory management isn't only about avoiding stockouts. Ordering too early or too much creates the opposite problem: overstock. Excess inventory can:
- tie up cash;
- clutter the stockroom;
- increase the risk of products losing value;
- add to your dead stock;
- limit your ability to buy better-selling items.
The goal is to balance product availability with a reasonable stock level. Good restocking doesn't fill the stockroom to the brim. It covers demand with a stock level that suits the business. See Overstock: how to identify and reduce it.
Spotting the products that need attention
Not every item needs the same level of attention. Several signals can help flag the products to watch:
- stock close to the reorder point;
- slowing turnover;
- no sales for an unusually long period;
- high stock relative to sales velocity;
- recurring gaps between theoretical and actual stock;
- lengthening supplier lead times;
- sharply accelerating sales.
On its own, each signal tells only part of the story. Combining them gives you a clearer picture of each item's situation. See also How to prevent stockouts.
How can AI help with inventory management?
When sales, stock and purchasing data are accessible, AI makes them easier to analyze. A store owner can ask, for example:
“Which products are likely to run out this week?”
The analysis can combine remaining stock, recent sales velocity and, when available, supplier lead time.
They can then follow up with:
“Which of these products already have a supplier order in progress?”
Then:
“Prepare a restocking proposal for the rest.”
In this kind of scenario, the AI agent can chain several steps: detecting a crossed threshold or a stockout risk, analyzing stock and sales, suggesting a restock, then preparing a supplier order. The value isn't just in displaying stock levels; it's in combining several data points to prepare a decision. For how this kind of agent works in general, see AI agent for retail: definition, use cases and examples.
AI can also spot slow-moving and dead stock
It works the other way around, too. Instead of looking for products likely to run out, the store owner can ask:
“Which products still have plenty of stock but barely sell anymore?”
The analysis can then combine:
- quantity on hand;
- last sale date;
- recent sales;
- turnover;
- stock coverage.
This flags items that may be dead stock or overstocked. It's then up to the store owner to decide what to do: keep the product, cut back upcoming orders, transfer the stock or consider a promotion.
From analysis to action with Gillia
Gillia, an AI agent for independent businesses, lets you query your store's data in plain language and chain several steps from a single request. For example:
“Which products are at risk of running out before the next delivery?”
Then:
“Prepare the restock order.”
Gillia can use the available information to flag the items concerned and prepare the corresponding operation. That way, the store owner moves straight from stock analysis to action.
When an operation requires approval, you can review it before it runs. Metrics you track every day can be pinned to the dashboard, and recurring checks can be handed off to a routine, as explained in Task automation for retail.
Good inventory management starts with reliable data
AI doesn't automatically fix inaccurate stock data. If movements aren't recorded, receipts are entered wrong or physical discrepancies are never checked, the analysis will be built on bad data too. The priorities stay the same:
- RecordEvery stock movement (receipt, sale, transfer, loss) must be entered correctly.
- CheckA regular physical count confirms that theoretical stock matches actual stock.
- AnalyzeMetrics (turnover, coverage, value, dead stock) give you a clear picture of the situation.
- AnticipateMinimum stock and the reorder point let you plan restocking before a stockout happens.
- RestockThe supplier order is prepared, then approved, with help from AI where useful.
AI can make analysis and some operations easier, but inventory management is only ever as good as the data behind it. With Gillia's inventory management, the goal is simple: connect this whole cycle so store owners get the right answer without digging through multiple screens, and can act on it right away.