Company Inventory: Tracking Movements and Stock Counts Beyond the Total
Design a reliable way to manage physical stock and movements.
SqualiOnline editorial team · 2026-09-07
The total stock figure is the number everyone looks at, and it’s the one that’s no use for deciding anything. It says how much material is in the company, not whether you can accept an order today, not where a part is located, not why three of them were missing yesterday. Operational questions only get answered if every movement leaves a trace. This guide is about designing that trace, before choosing any tool.
Item records: the boring part that decides everything
A systematized inventory stands or falls on four definitions, and it’s worth discussing them at the start rather than later.
- The item: one code for one thing. The typical problem is the opposite, meaning the same part with two codes because two vendors call it something different.
- The unit of measure. If you buy by the case and pick by the piece, the system needs to know the ratio and use just one as the reference. It’s the most common source of unexplained discrepancies.
- The location. Even in a small warehouse, “shelf C, level 2” is worth more than “warehouse.” Without locations, a stock count turns into a treasure hunt.
- The identifier: a barcode, a label, a lot or serial number where you need to trace back to the individual supply batch. The lot number is added only where it’s really needed, because applying it to everything doubles the work for whoever’s recording.
The rule that always holds: whoever records something needs to be able to do it without interpreting. If someone standing in front of the shelf has to decide which code to use, sooner or later the choice will come out different.
Everything that moves is a movement
There’s no such thing as a stock level that updates itself: there are movements that, added together, produce the stock level. It’s a conceptual difference with practical consequences, because it lets you answer the question “why are three missing” instead of just noting that they are.
- Receipt: goods received from a vendor, a customer return, a return from outsourced processing.
- Issue: shipment to a customer, delivery to production, use on a job site, a sample.
- Transfer: from one location to another, or from one site to another. It doesn’t change the total but it changes where the material is, and if it isn’t recorded, you get the classic “it’s there but can’t be found.”
- Adjustment: a correction after a count, breakage, scrap, loss. It’s the only movement that changes the stock level with no external cause, and for that reason it needs a reason and a person attached to it.
Every movement carries with it a date, item, quantity, location, reason code, who carried it out and, when one exists, the document that justifies it. A movement with no reason code is a hole that nobody will be able to explain by year’s end.
Physical, allocated, available
Different numbers that in everyday language all get called “stock.” Confusing them is the most common cause of broken promises to customers.
| Number | What it tells you | What it’s for |
|---|---|---|
| Physical | What’s on the shelves right now | Stock counts and checking for discrepancies |
| Allocated | What’s already promised to orders or job orders | Knowing what can’t be used for anything else |
| Available | Physical minus allocated | Answering whether an order can be accepted |
| Incoming | What’s on order with vendors, with the expected date | Deciding on a future delivery |
Whoever sells needs to see the available figure, whoever counts needs to see the physical figure. If the same screen shows a single number without saying which one it is, every department will interpret it its own way.
Stock counts and discrepancies, with an owner
Discrepancies always exist. A serious system isn’t one that avoids them, but one that makes them visible early and traceable to a specific moment.
- A full annual count finds discrepancies once the cause is no longer traceable.
- Cycle counting — a few items at a time, on rotation, more often on the ones that move the most — finds the same discrepancies while it’s still possible to figure out where they came from.
- Whoever counts shouldn’t be the only one who can make an adjustment. It’s not about distrust: an adjustment made by the same person who got the pick wrong erases the information.
- Every adjustment has a reason picked from a short list — breakage, picking error, receiving error, not found — because it’s that list, read back after six months, that tells you where to step in.
When the discrepancy is recurring on the same item, the problem is almost never theft: more often it’s the unit of measure, a duplicate code, or a routine pick that nobody records because “it’s just for the workshop.”
An illustrative day
A company with a spare-parts inventory. The account is illustrative, but the three situations come up every day.
- At eight, a vendor arrives. The warehouse clerk checks the delivery note against what’s being unloaded: two packages have a different quantity. They record what actually arrived and flag the discrepancy to purchasing. Recording the delivery note instead of the actual contents is the most elegant way to get the stock count wrong for months.
- At eleven, a technician takes two parts for an urgent repair. There’s no order linked to it, but the pick still needs to be recorded, with reason code “service call” and the customer reference. It’s the movement that gets skipped most often, and it’s the main reason the stock level in the system ends up higher than the real one.
- In the afternoon, during a cycle count, three units of one item are missing. Before making an adjustment, the history is checked: over the last two months there were four picks for service calls. The adjustment still gets made, with its own reason code, but the useful conclusion is a different one: that material often leaves without paperwork, and it needs a fast way to be recorded.
Connecting orders and requirements
Inventory becomes a decision-making tool once it’s connected to what’s about to happen.
- An accepted customer order allocates the material. If it doesn’t, two sales reps will sell the same part.
- Minimum stock levels need to be defined per item and tied to the replenishment lead time, not set to the same number for everything: a part that arrives in two days and one that arrives in two months shouldn’t have the same threshold.
- The purchase suggestion stays a suggestion. Whoever buys makes the decision, because the system doesn’t know about the vendor closed for vacation or the customer about to double their order.
When it’s worth stopping to think first
What this guide doesn’t cover
This guide covers inventory from the inside: how it’s defined, how it’s moved, how it’s checked. Syncing with an online store — which system has authority over the stock level, how to avoid selling what isn’t there, with what delays — is a separate problem and is addressed elsewhere, as is organizing purchase requests and approvals upstream of receiving.
Frequently asked questions
Do you need a barcode scanner?
Not from day one. It’s worth it once movements are numerous or typing errors are frequent. But labels and locations need to come first: without those, the scanner has nothing to read and the problem stays exactly the same.
How often should a stock count be done?
Besides the annual count required for compliance, it’s worth running a continuous cycle count on small groups of items, more often on the ones that move or are worth more. It catches discrepancies while the cause can still be reconstructed.
Could the business management system we already have be enough?
Often yes. Check whether it supports locations, movement reason codes, a history you can look up by item, and the distinction between physical and available stock. If those features exist and nobody uses them, the problem is procedures and training, not the tool.
Let’s define the movement flow of your inventory.
If you’d like to talk it through, the service that handles this is Custom software.