Inventory management is one of the pillars of retail. Stock that is too low leads to stockouts and lost sales. Stock that is too high ties up cash, takes up space and increases the risk of unsold goods.
Managing your inventory well therefore doesn't mean stocking as much as possible. The goal is to have the right quantity of products at the right time, while limiting the gaps between recorded stock and the stock actually available.
To get there, several metrics and methods let you track movements, anticipate needs and prepare restocking.
Understanding the stock cycle
A store's stock is constantly changing. It goes up when goods are received and goes down with sales, transfers, supplier returns, losses, breakage or other recorded outflows.
The principle seems simple:
opening stock + inflows - outflows = theoretical stock
But in practice, gaps can appear between this theoretical stock and the quantities actually on hand. Good management therefore rests on two complementary dimensions:
- recording movements correctly;
- regularly checking that the data matches reality.
Which metrics should you track to manage your inventory?
Not all metrics answer the same question.
Available stock
This is the quantity considered available for sale. Depending on the management system used, it may account for physical stock, reserved quantities, customer orders or other movements already recorded.
Inventory turnover
It measures how quickly stock is renewed over a given period. High turnover can signal a product that sells quickly; low turnover can reveal a slow-moving, seasonal or overstocked item. Calculation details in Inventory turnover: how to calculate it.
Stock coverage
It estimates how long the available stock can cover sales if the observed pace continues. In particular, it helps spot items with little lead time before a possible stockout.
Stock value
It represents the financial value of the goods held. For simple tracking, it can be estimated from the quantities in stock and their cost. In accounting, however, its valuation depends on the method the company uses.
Dead stock
An item that stays in stock for a long time without recording a sale ties up cash and space. See Dead stock and tied-up inventory.
Actual stock vs. theoretical stock: why do gaps appear?
Theoretical stock is the quantities calculated by the system from recorded movements. Actual stock is the quantities physically on hand. The two can differ for many reasons:
- breakage;
- theft or unknown shrinkage;
- data entry error;
- unrecorded movement;
- error during a goods receipt;
- misplaced product;
- counting error.
A physical count then lets you compare the stock actually present with the theoretical stock and correct the inventory discrepancies identified. The more regular these checks are, the more reliable the data used for restocking and analysis remains.
Reorder threshold and reorder point
Waiting until a product is out of stock before reordering it is generally too late. Between the moment the order is placed and the moment the goods arrive, sales continue. Minimum stock is the quantity needed to cover sales during that lead time.
In a simple approach:
Minimum stock = average sales × restocking lead time
But this calculation doesn't protect against the unexpected: stronger-than-expected sales, supplier delays or incomplete deliveries. Safety stock can therefore be added. The reorder point, or reorder threshold, is then the level at which a new order should be prepared.
In a simple approach:
Reorder point = minimum stock + safety stock
The threshold must be adapted to each item or group of comparable products. An item that sells ten times a day with a supplier delivering within a week isn't managed like a product that sells twice a month and is available within 48 hours.
Stock counts to make inventory reliable
Even with an accurate management system, system stock can gradually drift away from reality. A stock count measures that gap.
The annual inventory
It consists of counting a large part or all of the stock on a given date. This method gives an overall snapshot, but generally requires significant organization.
Cycle counting
Cycle counting, also called rolling inventory, consists of regularly checking part of the stock rather than waiting for a full inventory. Items can be split according to their importance, value, turnover or risk of discrepancy. The most sensitive products can thus be counted more often.
This method lets you gradually detect anomalies and keep theoretical stock closer to reality.
How do you prepare restocking?
The reorder point tells you when to restock. It doesn't determine on its own how much to order. The restocking quantity can depend on several factors:
- sales pace;
- available stock;
- orders already in progress;
- supplier lead time;
- desired safety stock;
- sales forecasts;
- seasonality;
- supplier packaging;
- minimum order quantity;
- target stock.
Let's take an example. An item is approaching its reorder point. This indicates that restocking should be considered. But if a delivery is already on its way, it would be wrong to calculate the new order solely from the stock currently available. Likewise, an item usually sold individually may need to be ordered by the case of 12.
Restocking must therefore combine stock data with real supply constraints. This complete cycle — from calculation to order — is detailed in our article on supplier orders.
Taking supplier lead times into account
Two products that sell at the same pace don't necessarily need the same stock level. If the first supplier delivers within two days and the second within ten days, the risk of a stockout isn't the same. The supplier lead time directly affects when the order must be triggered.
Ideally, you should use the lead time actually observed, not just the announced theoretical one. A supplier announced at three days who regularly delivers in five should be factored into the calculation with that operational reality.
Avoiding overstock
Inventory management isn't only about avoiding stockouts. Ordering too early or in too large a quantity can cause the opposite effect: overstock. Excess stock can:
- tie up cash;
- clutter the stockroom;
- increase the risk of depreciation;
- generate more dead stock;
- reduce the ability to buy better-performing items.
You therefore need to seek a balance between availability and a useful stock level. Good restocking isn't the kind that fills the stockroom to the maximum. It's the kind that covers needs with a stock level suited to the business. See Overstock: how to identify and reduce it.
Identifying products to watch
Not all items need the same attention. Several signals can help bring up the products to watch:
- stock close to the reorder point;
- slowing turnover;
- no sales for an unusual period;
- high stock relative to the sales pace;
- repeated gap between theoretical and actual stock;
- increasing supplier lead time;
- sharply accelerating sales.
Taken in isolation, each of these metrics gives only partial information. It's combining them that gives a better understanding of an item's situation. See also Anticipating a stockout.
How can AI help manage inventory?
When sales, stock and supply data are accessible, artificial intelligence can make analyzing them easier. A merchant can, for example, ask:
“Which products are likely to run out this week?”
The analysis can cross-reference the remaining stock, the recent sales pace and, when available, the supplier lead time.
They can then ask:
“Of these products, which already have a supplier order in progress?”
Then:
“Prepare a restocking proposal for the remaining items.”
In this kind of scenario, the agent can chain several steps: detecting a threshold crossed or a stockout risk, analyzing stock and sales, proposing restocking, then preparing a supplier order. The value isn't simply in displaying the stock level, but in using several pieces of data to prepare a decision. For the general principle behind this agent, see AI agent for retail: definition, uses and examples.
AI can also spot stock that no longer moves
The reasoning also works the other way around. Instead of looking for products likely to run out, the merchant can ask:
“Which products still have a lot of stock but hardly sell anymore?”
The analysis can then bring together:
- quantity available;
- last sale date;
- recent sales;
- turnover;
- stock coverage.
This brings up items that are potentially dead or overstocked. It is then up to the merchant to decide on the right action: keep the product, reduce upcoming orders, transfer the stock or consider a promotion.
From analysis to action with Gillia
Gillia lets you query your store's data in natural language and chain several steps from a single request. For example:
“Which products risk a stockout before the next delivery?”
Then:
“Prepare the necessary restocking.”
Gillia can use the available information to bring up the items concerned and prepare the corresponding operation. The merchant thus keeps continuity between analyzing stock and the action to take.
When an operation requires approval, it can be checked before it is carried out. Requests that are useful day to day can also become routines to regularly monitor the same metrics, as discussed in Task automation for retail.
Good inventory management starts with reliable data
Artificial intelligence doesn't automatically fix incorrect stock. If movements aren't recorded, if receipts are wrong or if physical gaps are never checked, the analyses will also rest on incorrect data. The priority therefore remains:
- RecordEvery stock movement — inflow, sale, transfer, loss — must be entered correctly.
- CheckA regular count verifies that theoretical stock matches actual stock.
- AnalyzeMetrics (turnover, coverage, value, dead stock) give a clear picture of the situation.
- AnticipateThe reorder threshold and reorder point let you prepare restocking before the stockout.
- RestockThe supplier order is prepared and then approved, with the help of AI when useful.
AI can make analysis and certain operations easier, but the quality of management always depends on the reliability of the data used. With Gillia's inventory management, the goal is simple: connect this whole cycle so that the merchant gets the right answer without searching by hand across several screens, and can act on it directly.