Tracking inventory means keeping a record of what you hold, where it is, and what moves — accurately enough to sell from and buy against. In practice: give every product a code, give every storage place a labelled location, record each movement at the moment it happens, and count a rotating slice of it regularly to check the records are still true.
The method can be a spreadsheet, a barcode system or dedicated software. What determines whether it works is not which of those you pick but whether the recording happens at the point of movement, every time.
What inventory tracking is, and how it differs from inventory management
Tracking is knowing what you have. Quantities, locations, movements. It is a record-keeping discipline, and it answers questions of fact: how many, where, and when did it move.
Inventory management is deciding what to have. How much to hold, when to reorder, what to buy, what to discontinue, what it costs to carry. It is a planning and commercial discipline, and it answers questions of judgement.
The relationship runs one way. Every management decision calculates from the tracking data, so a business with good management and bad tracking is making careful decisions from wrong numbers. That is why tracking comes first, and why “we need better forecasting” is frequently a tracking problem wearing a forecasting costume.
A third term, inventory control, sits between them — the processes that keep the tracking accurate, including counting and adjustment discipline.
Inventory tracking methods: the maturity ladder
Businesses tend to move up this ladder as they grow, and each step solves the problem the one below it creates.
1. Pen and paper. A written stock book or a clipboard by the shelf. Costs nothing, needs no training, works for a handful of lines in one room. It fails as soon as two people are making stock decisions, because neither knows what the other moved, and it produces no history you can analyse.
2. Spreadsheets. The most common starting point for a growing business, and genuinely capable at small scale. A row per product, columns for quantity, location, cost and reorder level, and formulas doing the arithmetic. The ceiling arrives sooner than people expect: one person editing at a time, no audit trail, and silent errors when a formula is dragged one row short. There is a full guide in inventory management spreadsheet.
3. SKU and location discipline. Not really a tool so much as the step that makes everything above it work. Every product gets a unique SKU; every storage place gets a labelled location — aisle, bay, shelf. Businesses skip this and then wonder why their software produces poor results.
4. Barcode scanning. Codes printed on products and locations, read with a scanner or a phone, feeding a system directly. This is the step that changes accuracy fundamentally, because it removes manual transcription from the moment stock moves. Error rates fall by orders of magnitude. Barcode scanning for inventory covers the setup.
5. Dedicated inventory software. Records that update from connected sales channels, multi-location tracking, reorder alerts, audit trails and reporting. Usually adopted alongside scanning rather than instead of it — see inventory management software.
6. RFID. Radio tags read without line of sight, often many at once. It makes counting nearly instant and stock location automatic. The cost per tag confines it to higher-value goods, reusable containers, or operations where counting speed is itself the benefit. What is RFID has the detail.
7. A WMS. At the top of the ladder, software that not only tracks but directs the work — telling operators where to put things and where to pick from. This is warehouse management rather than tracking, and it is the right answer once directing work is the constraint.
A note on the sequence. Steps 3 and 4 deliver more improvement per pound than anything above them, and they are routinely skipped in favour of buying software. Software applied to unlabelled locations and inconsistent processes produces wrong numbers faster than the spreadsheet did.
How to track inventory for free
You can run genuinely adequate tracking without spending anything, provided the range is small.
A spreadsheet, structured properly. One row per SKU. Columns for code, description, location, quantity on hand, quantity on order, reorder level, unit cost and date last counted. Add a conditional format that highlights any row where quantity falls below the reorder level, and you have an alert system.
The essential additions most people omit: a movements sheet recording every receipt and issue with a date, so you have a history rather than only a current figure; and a date counted column, so you can see what has not been verified for months.
Free barcode generation. Any spreadsheet can generate Code 128 barcodes using a free barcode font, and free online generators will produce printable sheets. Labels themselves cost very little.
Your phone as a scanner. Free apps read barcodes and export to a spreadsheet. Slower than a dedicated scanner and perfectly workable for a weekly count or a small receiving operation.
Free tiers of inventory software. Several products offer a free plan capped at a number of orders, SKUs or users. Worth checking before assuming you need to pay, and worth reading the export terms before committing data to one.
What free costs you. Manual entry time, no real audit trail, no automatic sync with sales channels, and a ceiling that arrives when you add a second channel or a second person. It is the right answer at the start and a false economy past a certain point.
How to choose the right method
Match the method to the constraint you actually have.
Number of SKUs. Under about 50, a spreadsheet is fine. Between 50 and a few hundred, the case for scanning and software strengthens. Above that it is difficult to argue against.
Number of locations. One room needs less than one building; one building needs less than several sites. Multi-location is where spreadsheets fail hardest.
Number of sales channels. This is the single strongest trigger. One channel can be tracked manually. Two or more means one figure has to serve both, and manual reconciliation is a losing battle from the first week.
Number of people touching stock. One person can hold a system in their head. Three cannot, and shouldn’t.
Order volume. A handful of orders a day can be recorded by hand. Fifty cannot.
Product characteristics. Anything needing batch numbers, serial numbers or expiry dates pushes you towards proper software regardless of scale, because tracking those by hand is unreliable and, in regulated sectors, insufficient.
Value and consequence. High-value stock justifies more tracking effort. So does anything where running out costs you a customer rather than a sale.
A reasonable rule. Fix your SKUs and locations first, whatever your size — that costs a fortnight and improves everything. Add scanning when manual entry errors start reaching customers. Add software when you add a second sales channel or a second location. Consider RFID only when counting time is a genuine operational constraint.
Best practices for accurate tracking
Record at the point of movement. The single most important habit. A scan when the stock physically moves, not a note typed at a desk twenty minutes later. Records created after the fact record what somebody remembered.
Track by location, not just by total. A product can reconcile perfectly in total whilst being wrong in two locations that offset each other — and those two locations still cause failed picks. Location-level tracking is what makes a record actionable.
Give every location a label and every product a home. Aisle, bay, level, position, in a logical scheme a new starter can follow. Nothing else works reliably without this.
Count continuously rather than annually. Cycle counting a rotating slice — ten locations a day on your top lines — catches errors whilst they are still traceable. An error found at year end happened at some unknown point in the previous twelve months.
Count blind. Never show the counter the expected quantity. People confirm what they are shown, and blind counting is the single largest accuracy gain available at no cost.
Investigate before you adjust. A large share of apparent losses are misplacements. Check adjacent locations, returns, quarantine and the transaction history before writing a number down.
Use reason codes on every adjustment. Mis-pick, receiving error, damage, unrecorded movement, theft. A month of coded adjustments usually points straight at one process that needs fixing. Free text cannot be counted.
Separate goods-in, quarantine and returns physically. Unchecked stock in the pick face is how bad inventory enters the record.
Reconcile what is on order. Stock in transit is yours and is not on the shelf. Reorder decisions that ignore it cause duplicate orders.
Measure record accuracy and publish it. The share of counted locations matching exactly, by location and exactly rather than by units and approximately. Put the figure where the team can see it.
Review the trend, not the point. One month’s accuracy figure is noise. Three months moving one way is a signal.
What to record: the fields that matter
Whatever tool you use, the data model is roughly the same. These are the fields worth having from the start, because retrofitting them across an established range is tedious.
Per product:
| Field | Why it matters |
|---|---|
| SKU | The unique code everything else hangs from |
| Description | Enough to identify it without opening a box |
| Barcode / GTIN | The external code, where the product has one |
| Unit of measure | Whether the quantity means units, cases or pallets — a common source of silent error |
| Location | Where it lives; multiple rows if it lives in more than one place |
| Quantity on hand | The current figure |
| Quantity allocated | Committed to orders not yet picked; on hand minus allocated is what you can actually sell |
| Quantity on order | In transit, so reorder decisions do not duplicate |
| Reorder point | The trigger level |
| Reorder quantity | How much to buy when it fires |
| Unit cost | For valuation and margin |
| Supplier and lead time | So the reorder point can be calculated from reality |
| Date last counted | Shows what has not been verified |
Per movement: date and time, product, quantity, direction (in, out, transfer, adjustment), location from and to, reference (order or purchase order number), and who did it. Movements are what turn a stock figure into something you can investigate, and they are the first thing a spreadsheet-based system tends to omit.
For regulated or dated goods, add batch or lot number, serial number, expiry date and received date. These cannot be reconstructed later, so if there is any chance you will need them, capture them from day one.
The field people most often skip is quantity allocated. Without it, stock committed to orders still looks available, and the system will happily let you sell it twice.
Common tracking challenges
Discipline decays under pressure. Scanning gets skipped during a busy peak, which is precisely when errors are most expensive. The fix is making the scan faster than the workaround, not exhorting people to be careful.
Unlabelled or ad hoc locations. Stock put “wherever it fitted” cannot be tracked by location, and no software solves it.
Similar products stored together. Adjacent bays holding near-identical items is the largest single cause of picking errors. Separate them physically.
Multi-channel divergence. Two systems each holding their own stock figure, drifting apart daily. The only real fix is one master figure that every channel reads from.
Returns entering unchecked. Goods put back on the shelf without inspection or a record. Returns need their own area and their own process.
Kits and bundles not converted. Building kits without deducting the components leaves records believing components are available when they are sealed in boxes.
Units of measure confusion. Ordering in cases and selling in units, with nobody defining which the stock figure is in. Common, and it produces errors that look like theft.
Stock that is present but unsellable. Damaged, expired or quarantined goods counted as available. They need their own status, not deletion from the count.
Nobody owning the number. Where inventory accuracy belongs to everyone, it belongs to nobody. Name a person.
A worked example: setting up tracking from scratch
A business holds about 350 SKUs in one unit, sells on its own site and one marketplace, and tracks everything on a spreadsheet that two people edit. Overselling happens a few times a month.
Week one — codes and locations. Every product gets a unique SKU following one scheme, avoiding leading zeros and the letters O and I. Every storage place gets a printed label: aisle, bay, shelf. This takes four days across two people and it is the largest single piece of work.
Week two — a baseline count. They count every location blind, by location rather than by product, and record what they find. Record accuracy comes out at 71%. Of the discrepancies, most are products found in a location the spreadsheet did not know about — which is the expected result of the pre-labelling era rather than evidence of theft.
Week three — the system. They move to inventory software connected to both sales channels, importing the counted figures as opening stock rather than the spreadsheet’s numbers. Two phone-based scanners are set up, and locations and products are barcoded.
Week four — the routine. Receiving now scans against the purchase order. Picking scans the location and the product. Ten locations a day are counted, starting with the highest-value lines, and adjustments carry a reason code.
After one quarter. Record accuracy on counted locations is 96%. Overselling has stopped, because one figure now serves both channels. The reason codes show that a third of remaining adjustments are coded “found in wrong location”, pointing at putaway rather than picking — so putaway now scans the destination location too.
What it cost. Two days of labelling, a week of setup, two phone scanners, and a software subscription. No consultant, no WMS, no new building.
The sequence is the lesson. Codes and locations first, then a clean baseline, then the system, then the routine. Businesses that start at step three — buying software and importing an unreliable spreadsheet — get a faster, more confident version of the same wrong numbers.
Where to go deeper
- What is inventory control — the processes that keep records true.
- Barcode scanning for inventory — setting up the scanning layer.
- Cycle counting — the verification routine.
- Inventory management spreadsheet — doing it properly for free.
- What is inventory management software — the step up from a spreadsheet.
- What is a SKU — the code every system depends on.
The one-line takeaway
Track inventory by giving every product a code and every storage place a labelled location, recording each movement as it happens, and cycle counting a rotating slice to verify the records. The method matters less than the discipline — and fixing SKUs and locations delivers more than any software you buy before doing it.
Frequently asked questions
How do small businesses track inventory? Most start with a spreadsheet — a row per product with quantity, location, cost and reorder level — and move to barcode scanning and dedicated software as SKU count, channels or people grow. The step that matters most at any size is giving every product a unique code and every storage place a labelled location.
Can I track inventory in Excel? Yes, and it works well at small scale. Build a row per SKU with quantity, location, reorder level and date last counted, add a separate movements sheet for history, and use conditional formatting to highlight low stock. The limits are one editor at a time, no audit trail, no automatic sync with sales channels, and silent formula errors.
What is the best way to track inventory? Barcode scanning feeding inventory software, with locations labelled and cycle counting to verify. That combination removes manual transcription, keeps one figure across every channel, and checks itself continuously. Below a few dozen SKUs in one location with one channel, a disciplined spreadsheet is genuinely adequate.
What is the difference between barcode and RFID for inventory? Barcodes need line of sight and one scan per item, and cost almost nothing to print. RFID reads through packaging and reads many tags at once, so counting is far faster, and the tags cost meaningfully more. Most operations use barcodes throughout and RFID selectively on high-value goods or reusable containers.
How often should I count stock? Continuously rather than annually. Count your top lines by value weekly and a rotating slice of everything else so the whole range is checked over a quarter or so. Keep a full annual count if your accounts require it, and treat it as verification rather than as the primary method.
What is the difference between inventory tracking and inventory management? Tracking is knowing what you hold, where it is and what moved — a record-keeping discipline. Management is deciding what to hold, when to reorder and what to buy — a planning discipline. Management calculates entirely from tracking data, which is why poor tracking makes good management impossible.
How do I track inventory across multiple sales channels? With one master stock figure that every channel reads from, held in inventory software connected to each platform. Manual reconciliation between channels fails quickly, and overselling is the symptom. This is the most common trigger for moving off a spreadsheet.
Why are my inventory records always wrong? Almost always because recording happens away from the movement — typed later, or skipped under pressure. The other common causes are unlabelled locations, similar products stored adjacently, returns going back unchecked, and kits built without deducting their components.
Warehouse Wise is a free, independent library of guides on inventory management, warehousing and fulfilment, written for the people actually doing the work — new warehouse and inventory staff, small and growing retail and ecommerce businesses, and anyone who holds stock and wants to understand it properly. Every guide explains the term, shows a worked example with real numbers where it helps, and stops. We are independent: we do not sell software or equipment, we take no commission on referrals, and our guides are not ranked by who pays us.