Inventory shrinkage is the difference between the stock your records say you hold and the stock you can actually find, usually valued at cost. For a small business, the useful question is how much of that gap is real loss, because what a count reveals is usually a mix of genuine losses and records that were simply wrong.
Picture a month-end count across a shop, a weekend market stand and a back stockroom. The reports show a net 20 units fewer than expected. After checking, record errors account for seven of them: cash sales at the stand nobody entered, a delivery booked in for more units than arrived, two colours of one product swapped. Six are confirmed losses with a known cause, and seven are still unexplained. Treat all twenty as lost and you overstate the problem; ignore the record errors and they return next month.
This guide starts where the count ends; for the counting itself, see how to run an inventory count. It covers classifying each difference, calculating a shrinkage rate per location, reading retail shrinkage benchmarks and reducing loss with everyday processes.
What inventory shrinkage means, and the terms around it
Inventory shrinkage is the reduction between the inventory shown in your records and the inventory physically on hand, caused by theft, damage, spoilage, unrecorded use or administrative error. It is counted in units, valued at cost and usually expressed as a rate for a defined period.
Retail shrinkage, or simply “shrink”, is the same idea applied to shops; in the UK you will also hear “stock loss”. Accounting and industry definitions commonly include administrative errors, but not every difference found on count day ends up as a loss, so keep these terms apart.
| Term | What it means | How it relates to shrinkage |
|---|---|---|
| Inventory discrepancy | Any difference between the expected and the counted quantity in one location | A starting point to investigate, not yet a loss. |
| Administrative or record error | The stock is fine but the record is wrong: wrong variant, wrong location, a sale not entered, a delivery booked in from the invoice | Included in many shrinkage definitions. Once corrected, no stock is written off, though some errors cost money elsewhere, such as an uncredited short delivery. |
| Confirmed physical loss | Stock that left the business unsold and unintended, with a known cause: broken, expired, damaged in packing, lost in transit | Shrinkage. Valued at cost and written off. |
| Authorised exit | Stock that leaves on purpose without a sale: samples, gifts, testers, staff use | Record and value it. Whether it counts inside your shrinkage measure or as a separate cost is a business policy; apply it consistently. |
| Unexplained loss | Stock that is still missing after every check, with no known cause | Included provisionally in the shrinkage measure. It may be theft or an undiscovered error; the number proves neither. |
| Inventory adjustment | The change you make to the stock record so it matches what you counted | The correction, not a cause. Give each one a reason. |
| Write-off | Removing the cost of lost stock from inventory in your books | How confirmed and unexplained loss reaches the accounts. |
Record error or real loss: why the same gap can be either
A line that reads −3 only tells you the record and the shelf disagree. The units could be in a customer’s bag, in a bin, in a box that never left the supplier, or on the shelf under another size label. Investigate before you adjust.
When the record is wrong
Typical record errors: a sale entered under the wrong variant or location, a delivery booked in before the boxes were counted, a return restocked twice, an event cash sale never entered, two manual edits for one move between locations. Positive differences usually belong here.
Correct the record and the routine behind it. A short delivery is the clearest case: units booked in never arrived, so you correct the record and claim from the supplier. If the shortfall isn’t credited, you have a financial loss, but not shrinkage of stock you physically held. Track it as a supplier cost, separate from your shrinkage rate, and agree the treatment with your accountant.
When the stock is really gone
Physical loss is stock that left unsold and unintended: breakage, expiry, packing damage, units lost in transit, and theft by customers or staff. The adjustment stands, the loss is valued at cost, and what changes is the process that let the stock go.
Samples, gifts, testers and staff use are authorised exits. Record and value them, but whether they sit inside your shrinkage measure or in a separate cost line, such as marketing, is a business policy to apply consistently. Unrecorded, they surface in the next count as unexplained differences.
Known and unknown loss
Within physical loss, known loss has a cause you can name: a dropped jar, an expired batch, units recorded as lost on a transfer. Unknown loss is what remains missing after every explanation has been checked.
In its Crime Report 2026, the British Retail Consortium changed its methodology to report costs as known loss, so the cost of theft covers verified incidents only, not goods missing without a verified theft, with the stated aim of not treating all loss of goods as theft. Its survey measures retail crime rather than a shop’s stock review, so its categories differ from this guide’s, but the principle carries over: unknown loss is a measurement, not an accusation.
How to calculate inventory shrinkage per location

Calculate shrinkage per stock location and defined period, such as the month between two counts; with more than one stock location, a business-wide figure hides where the problem is. After correcting record errors:
- Shrinkage in units = confirmed physical loss + unexplained loss, line by line. Unexplained units are included provisionally: they are missing, but the count doesn’t say why.
- Shrinkage value = shrinkage units × unit cost, per product or variant, then added up. Use cost, not selling price.
- Shrinkage rate (inventory basis) = shrinkage value at cost ÷ book inventory value at cost for the same location × 100. Book inventory is the cost value of everything the record held there on count day, before the count adjustments.
Count-detected loss and loss for the whole period
A count only detects losses that were never recorded. Losses already recorded during the period, such as a breakage adjusted on the day or units recorded as lost when closing a transfer, were deducted from the expected quantity, so they don’t appear as differences.
- Count-detected shrinkage = confirmed + unexplained loss found by this count.
- Period shrinkage = count-detected shrinkage + losses recorded during the period that fall within your shrinkage definition.
Take each loss from one source so it is never counted twice: from your adjustment log if it was recorded before the count, from the count if it surfaced there, even when it matches a log entry that was never applied to the record. Both measures assume the previous count was reconciled.
A worked example with three warehouses
The figures below are illustrative, for a small home-fragrance seller. The counter and a self-service kiosk both sell from the shop warehouse; a weekend market stand has its own stock; a back stockroom ships online and Telegram orders. Signed figures are units against the record. The example covers only losses detected by this count.
Table 1. Count results and investigation, in units
| Warehouse / line | Expected | Counted | Net difference | Record errors | Confirmed loss | Unexplained |
|---|---|---|---|---|---|---|
| Shop: large candle | 40 | 34 | −6 | −2 | −1 | −3 |
| Shop: gift set | 15 | 15 | 0 | 0 | 0 | 0 |
| Market stand: small candle | 30 | 22 | −8 | −5 | −2 | −1 |
| Back stockroom: diffuser, blue | 20 | 23 | +3 | +3 | 0 | 0 |
| Back stockroom: diffuser, white | 18 | 15 | −3 | −3 | 0 | 0 |
| Back stockroom: refill oil | 50 | 44 | −6 | 0 | −3 | −3 |
| All lines | 173 | 153 | −20 | −7 | −6 | −7 |
What the investigation found:
- Shop, large candle: 2 booked in from a short delivery (record error; supplier claim), 1 broken, 3 unexplained.
- Market stand, small candle: 5 cash sales matched to the stand’s takings but never entered (record error), 2 broken in transport, 1 unexplained.
- Back stockroom, diffusers: 3 white sales recorded as blue (record error that nets to zero).
- Back stockroom, refill oil: 3 damaged in packing, 3 unexplained.
The −20 is the net of 23 units short across four lines and 3 units over on one. The record-error column is a net figure too: its −7 combines 10 units short (2 large candles, 5 small candles, 3 white diffusers) and 3 over (blue diffusers). Once those are corrected, 13 units remain in the operational shrinkage measure: 6 confirmed losses and 7 unexplained. The 7 stay in provisionally: missing after checking, cause unknown. If a later check finds them, reverse them in the next period. Illustrative unit costs: $12 large candle, $6 small candle, $15 either diffuser, $4 refill.
Table 2. The same results valued at cost (count-detected losses only)
| Warehouse | Book inventory | Confirmed loss | Unexplained | Shrinkage | Rate |
|---|---|---|---|---|---|
| Shop | $8,000 | $12 | $36 | $48 | 0.60% |
| Market stand | $1,200 | $12 | $6 | $18 | 1.50% |
| Back stockroom | $10,000 | $12 | $12 | $24 | 0.24% |
| All | $19,200 | $36 | $54 | $90 | 0.47% |
All values are at cost. Rate = shrinkage ÷ book inventory × 100. Before investigation, the net difference at cost was $144 ($72 shop, $48 market stand, $24 back stockroom), of which $54 were record errors ($24, $30 and $0).
Three things stand out. The business-wide 0.47% looks calm, but the market stand runs at 1.50%, more than double the shop and over six times the stockroom. Rates on the net differences before investigation would have been 0.75% overall and 4.00% at the stand, inflated by record errors. And the diffuser swap nets to zero but still matters: until corrected, the record undersells blue and oversells white.
Look at money as well as rates: the stand’s 1.50% is $18, while the shop holds the largest unexplained figure, $36 of large candles. The count can’t say what happened to them, or whether the counter or the kiosk was involved, since both sell from the shop warehouse. Only sales records, stock movements or observation can point to a channel.
For the shop’s period figure, add the losses recorded during the month. If staff had adjusted the record for two broken large candles on the day, those units were already out of the expected 40. The shop’s period shrinkage would be $48 + (2 × $12) = $72, or 0.90% of the same $8,000 book inventory.
Inventory-based and sales-based shrinkage rates
The rate above uses an inventory basis. Retail reporting more often expresses shrink as a percentage of sales, which is how the NRF figure below is reported, and a sales-based rate can be built in more than one way:
- Shrinkage at retail ÷ net sales. Both sides at selling prices; a method used in some retail analyses.
- Shrinkage at cost ÷ net sales. A simple internal management ratio. It mixes cost with selling prices, so it reads lower than the retail version and can’t be compared with it.
Published benchmarks apply their own definitions, so check how a figure was built before comparing yours with it.
For the shop’s four large candles in the count-detected measure (one broken, three unexplained), with illustrative net sales of $6,000 for the month and a retail price of $30:
- Inventory basis: $48 at cost ÷ $8,000 book inventory at cost = 0.60%
- Cost to sales: $48 at cost ÷ $6,000 net sales = 0.80%
- Retail to sales: $120 at retail (4 × $30) ÷ $6,000 net sales = 2.00%
Same candles, three numbers, none comparable with the others. Pick one basis, write it next to the figure, and keep it every period.

Measure loss where it happens
Trapyfy holds stock per warehouse and country, so each location has its own stock record to count and compare.
What a normal shrinkage rate looks like for a small business
The most widely quoted benchmark comes from the National Retail Federation. Its National Retail Security Survey 2023 reported an average shrink rate of 1.6% of sales in fiscal 2022, up from 1.4% the year before. It is a survey of large US retailers, reported as a share of sales for one fiscal year.
That makes it context, not a target. A small retailer may measure on another base, over a month rather than a year, across a few locations where one broken box moves the rate. Being below 1.6% doesn’t prove control; being above it doesn’t prove theft.
Your best benchmark is your own history, measured the same way each time. After three or four counts you have a trend. Signals worth investigating:
- The same product appears as unexplained in successive counts.
- One location’s rate rises while the others stay flat.
- Unexplained loss grows as a share of the total, compared with known loss.
- Loss concentrates in a few high-value lines.
- Record errors of the same kind keep repeating, which points to a routine rather than a person.
Where stock goes when you sell through several channels and locations
Channels and locations are not the same thing. A counter, a kiosk, an online store and a Telegram shop can all sell from one warehouse, and when they do, every channel sells from the same warehouse count. Loss is measured where the stock sits; attribute it to a channel only when sales and movement records show where the units went.
| Where it happens | What typically goes wrong | Usually | Control |
|---|---|---|---|
| Counter | Sale recorded against the wrong variant or location; items handed over without a sale | Record error, sometimes loss | Check variant and location at checkout; spot-check fast sellers |
| Self-service kiosk | Products taken from display without a completed purchase | Physical loss | Keep valuable lines within staff sight; count display stock more often |
| Events and pop-ups | Cash sales not entered, samples, transport breakage, stock that never comes back | Both | Give event stock its own warehouse; count it out and back in |
| Online and Telegram orders packed by hand | Wrong variant picked, packing damage, replacements sent without a record | Both | Pick from the order, record every replacement |
| Transfers between locations | Units sent but never received; two manual edits that don’t match | Record error until confirmed lost | Move stock with a transfer document |
| Returns and refunds | Refunded item never restocked, restocked twice, or shelved while damaged | Mostly record error | A restocking decision for every return |
| Supplier deliveries | Booked in from the invoice; short or damaged boxes | Record error; an uncredited shortfall is a supplier cost, not shrinkage of goods received | Count before booking in |
| Samples, staff use and breakage | Taken or binned without a record | Breakage: known loss. Samples and staff use: authorised exits, classified by your own policy | Log it when it happens, with a reason |
Some of these have their own guides: running counter and kiosk sales from shared stock, handling event stock that doesn’t come back, recording units lost in transit, and dealing with returns that never go back on the shelf.
Processes that reduce shrinkage without a security budget

Most small-business shrinkage comes from routines with a gap in them, and closing those gaps costs time rather than equipment.
Count deliveries before you book them in
Check what arrived against the order and the delivery note, book in only what is in the box, and raise claims for short or damaged units the same day.
Give every return a restocking decision
A refund and a restock are separate events. For each return, someone decides and records whether it goes back on sale, back to the supplier or into the damaged pile; otherwise returns get lost on paper or restocked twice.
Move stock with one document, not two edits
Editing one location down and another up can leave a mismatch with no trace. A transfer that records what was sent and what was received turns units that never arrived into a recorded, known loss.
Log samples, staff use and breakage as they happen
Keep a simple log, on paper or in a shared sheet: date, product, variant, quantity, location and a reason from a short fixed list. Each entry turns a future unexplained difference into a recorded movement, with samples and staff use kept separate from breakage.
Limit who can change stock numbers
Decide who may adjust stock, ask for a reason with each adjustment, and review the adjustments alongside the count. The fewer people who can overwrite a figure, the easier every change is to explain.
Count high-risk lines more often
Give valuable, small and fast-moving products a short partial count every week or two. The inventory count guide covers how to schedule it.
Basic theft deterrence for a small shop
Keep valuable items near or behind the counter, keep displays tidy so gaps show, greet customers, and close with a quick check of high-value shelves. None of it needs technology.
Recording shrinkage in your accounts
Confirmed and unexplained losses are written off at cost: inventory falls and the loss is recognised as an expense, often within cost of goods sold. Record errors are corrected, not written off; an uncredited short delivery is a supplier cost; authorised exits are booked according to your own policy.
Keep the dated count report and your classification sheet as evidence, and ask your accountant about the exact entries and any tax treatment.
How Trapyfy helps you see where stock goes
Trapyfy is a commerce platform for sellers who run a Telegram shop, a staffed point of sale and a self-service kiosk from the same warehouse stock. Its inventory control covers the records a shrinkage review depends on:
- Stock per warehouse and country, and per variation for variable products, so blue and white are separate lines with separate quantities.
- Inventory counts per warehouse, full or partial, ending in a discrepancy report with expected, counted and difference for each line. Reports export to Excel or PDF, and completed counts can’t be deleted.
- Warehouse transfers as documents that record what moved, who moved it, when, and the quantities on both sides. Units that never arrive can be recorded as lost when you close the transfer.
- Purchase orders that add the received units to stock when you complete them. Complete one only once the delivery is physically there and counted.
- A stock history on each product’s page, alongside its sales and buyers.
- Cost per country on each product, which you can use to value the differences from a count.
Classifying the differences, valuing them and calculating the rate remain your own steps, using the method above.
A monthly shrinkage review
Run it monthly for each warehouse; the trend tells you more than any single figure.
- Settle open transfers, recording any units that never arrived.
- Check completed purchase orders against what physically arrived, and chase short deliveries.
- Run a partial count of high-risk lines, or use the latest full count.
- Classify every difference: record error, confirmed loss, authorised exit or unexplained.
- Correct the record errors and fix the routine behind each one.
- Value confirmed and unexplained loss at cost, then add losses recorded during the month, each from one source only.
- Calculate the shrinkage rate per warehouse, on the same basis as last month.
- Compare with previous periods and note which products and locations moved.
- Choose one process to change before the next review.
- File the count report and your classification sheet together.

Keep the paper trail behind every unit
Counts with discrepancy reports, transfers that record lost units, and purchase orders that add stock only when you complete them.
Frequently asked questions
What is inventory shrinkage?
Inventory shrinkage is the difference between the inventory your records show and what is physically on hand, caused by theft, damage, spoilage, unrecorded use or administrative error. It is valued at cost and expressed as a rate for a defined period and location.
How do you calculate inventory shrinkage rate?
Correct record errors, multiply the remaining missing units by unit cost, divide by the book inventory value at cost for the same location, and multiply by 100. Sales-based rates divide by net sales instead, with the loss valued at retail or at cost. These aren’t comparable, so name your basis and keep it.
What is a good inventory shrinkage rate?
There is no universal target. The NRF’s National Retail Security Survey 2023 reported 1.6% of sales for large US retailers in fiscal 2022, a different base and scale from most small shops. Track your own trend on a consistent basis instead.
What are the main causes of inventory shrinkage?
Physical loss: theft by customers or staff, breakage, expiry and units lost between locations. Administrative error: wrong variants, unrecorded sales, deliveries booked in from the invoice and returns restocked twice. Unrecorded samples and staff use also create differences, classified according to your own policy.
Is every inventory discrepancy shrinkage?
Not necessarily. Many definitions include administrative errors in shrinkage, but if investigation traces a difference to a record that can be corrected, no physical loss is confirmed and no stock is written off.
What is another term for inventory shrinkage?
Shrink, retail shrinkage, stock loss and inventory loss. “Stock loss” is common in the UK; “shrink” is the usual US retail shorthand.
How do you record inventory shrinkage in accounting?
It is written off at cost: inventory goes down and the loss is recognised as an expense, often within cost of goods sold. Keep the count report as evidence, and confirm the entries and any tax treatment with your accountant.
How can a small retailer reduce shrinkage without expensive security?
Count deliveries before booking them in, give every return a restocking decision, move stock with transfer documents, log samples, staff use and breakage with a reason, limit who can change stock numbers, count high-risk lines more often, and keep valuable items within sight of staff.
Every missing unit with a reason next to it
Shrinkage stops being a vague worry once each difference has a label: corrected, known or unexplained. Record errors get fixed at the source, known losses point to a process, and the unexplained figure becomes a number you can watch fall, location by location and month by month.
If you want your warehouses, counts and transfers in one system, create a Trapyfy account and start with the stock you already hold.
