Syntha Lab RU

Loss prevention: shrinkage, counting, fraud

A report of goods and money lost: how much is lost by records and in fact, and from which causes (errors, suppliers, staff, customers, returns). It shows whether protective measures pay off and is used to decide where to tighten control and how much to spend on tags, cameras and checks.

Production, logistics, channels Brand and retail Verified: 2026-11-14

Open interactively →

Why this document

Losses are rarely theft in one go: more often they are receiving errors, mis-sorts and returns, and anti-theft measures are powerless against them. So the retail director, security and the finance lead read the report: they separate causes and compare the cost of measures with what these recover. Without a breakdown the brand installs expensive gates where the loss happens at receiving. Counting and causes by group therefore come first, then the protection budget, otherwise money goes to secondary issues.

At a glance

What is this document and why is it needed?
A report of goods and money lost: how much is lost by records and in fact, and from which causes (errors, suppliers, staff, customers, returns). It shows whether protective measures pay off and is used to decide where to tighten control and how much to spend on tags, cameras and checks.
Who owns the document and how often is it rebuilt?
Run by security or inventory control with finance and operations; computed after every count (at least yearly, ideally cyclic); the result goes to the commercial and finance directors.
What mistakes are the most common?
Not naming the base of the percentage → losses are compared incorrectly → always state: to sales at cost or retail revenue. Writing off discrepancies without checks → errors and theft are hidden → check against documents and CCTV before write-off. One yearly count → causes are not visible → cycle counting and a count after a change of responsible person.

For whom

Brand
A brand loses in its own stores and warehouse and in transit (damage, short deliveries at partners); a separate risk is returns and fraud on online orders and consignment. For wholesale customers losses are handled by reports and claims.
Retail
For a retailer losses are a direct report line affecting margin: customer and employee theft, receiving errors and mis-picks, fictitious returns, damage; measured by counts and compared with sales and cost.

What it consists of

1. Stock balance

  • Opening stock + receipts + transfers (in − out) − sales at cost − returns to suppliers = book stock; compared with counted stock.

2. Causes of loss

  • Administrative (record errors, mis-picks, wrong prices), suppliers (short delivery, mixed picks), internal (employee theft, fictitious receipts and returns), external (customer theft), natural loss and spoilage.

3. Counts and cycle counting

  • Retailer: Full counts on a schedule and priority cycle counting; discrepancies are checked against documents before write-off; a recount of disputed locations.

4. Protective measures

  • Retailer: Anti-theft tags and gates (EAS), CCTV, warehouse access limits, keys and seals, segregation of duties (who receives, who writes off), control of returns and discounts at the till.

5. Fraud and anomalies

  • Till anomalies: frequent returns without receipts, manual discounts, canceled receipts after payment, negative stock, sales by one employee to the cards of acquaintances; signals are analyzed by rules and investigated.

6. Metrics

  • Losses as % of sales (or at cost), by store and category, trend; share of losses by cause; cost of protective measures and their contribution to reduction.

What each side does

  • Retailer: store and warehouse losses are a direct report line: computes losses to cost of goods sold by store, runs cycle counts, investigates cash anomalies and measures the effect of protection measures by pilot.
  • Brand: loses in its own warehouses and in transit and for wholesale customers papers losses with acts and claims; separately counts returns and fraud in online orders and on consignment.

Brand and retailer: what differs in practice

  • Brand: splits losses at the factory, warehouse and commission customers by link in the chain to know who is at fault and whom to claim from.
  • Brand: records short delivery and mis-picks by acts at receipt from the factory and carrier within the claim term.
  • Brand: compares losses in its own stores with wholesale channels only on one base (cost of sales).
  • Brand: keeps marked goods by codes: losing a code on a shortage is a separate problem with the marking system.
  • Brand: assigns accountable persons at the warehouse by order and changes them only with a stock count.
  • Brand: accounts for losses of samples, showroom and photo shoots as a separate item and does not mix them with theft.
  • Retailer: decides on stores with a high shrinkage percentage: audit, training, cameras, tags or a review of staff.
  • Retailer: runs stock counts on a schedule tied to risk: stores with high losses more often.
  • Retailer: compares customer theft with the cost of error: tags pay back while theft losses exceed the yearly cost of protection.

How to read the result and what decisions follow

  • The structure of causes matters more than the overall percentage: record errors are cured by process and discipline, theft by physical protection, suppliers by claims; one measure does not close everything.
  • Compare the payback of measures with the lease and season horizon: protection that pays back longer than the lease makes no sense.
  • Decision: the shrinkage threshold at which a store gets an audit; the protection budget; the stock-count schedule and claims plan.

Control checks

  • The sum of losses by cause equals total losses and the cause shares add up to 100%.
  • The book stock agrees with the ledger on the count date and all transfers, returns and write-offs before the count are posted.

Where the data comes from

Data sourceFields
ERPstock, movements, write-offs, sales at cost
Countactuals by location, discrepancies, recounts
Till and CCTVreturns, discounts, cancellations, tag alarms
Securityincidents, investigations, amounts

How to build it step by step

  1. Step 1: Set the basis: the Lab standard base, losses at cost (book stock minus actual) ÷ revenue excl. VAT; keep the “losses to cost of sales” reference as a separate line and do not mix it with the main one.
  2. Step 2: Run a count or cycle count; record the book stock on the same date.
  3. Step 3: Build the stock balance and compute the loss; reconcile with documents of the period (receipts, issues, returns).
  4. Step 4: Check disputed items against documents and CCTV before write-off; correct record errors.
  5. Step 5: Break losses down by cause (records, suppliers, employees, customers) from the checks; flag the unexplained remainder.
  6. Retailer: Step 6: Compare stores on loss share and find outliers; check processes there.
  7. Brand: Step 7: Assess protective measures: cost, loss reduction, payback.
  8. Retailer: Step 8: Record the write-off per accounting rules and update the loss plan for the next period.

Formulas

MeasureFormulasNote
Retailer: ShrinkageShrinkage = book stock − counted stockPositive is a shortage; negative is an overage, which is also a signal of record errors.
Retailer: Shrinkage as % of salesShrinkage, % = shrinkage at cost ÷ revenue excl. VAT × 100 (the single Lab KPI basis); against cost of goods sold it is an accounting reference onlyAlways state the base; foreign benchmarks are only a hint.
Retailer: Protection paybackMonths = cost of measure ÷ (monthly loss from the cause × reduction share)Take the reduction share from a pilot in several stores, not from supplier promises.
Retailer: Record accuracyIRA, % = locations matching records ÷ locations checked × 100Low accuracy masks real losses: the records get “adjusted”.
Retailer: Base of the loss percentage (single rule)Losses, % = Losses at cost ÷ Revenue excl. VAT × 100 (accounting reference: ÷ cost of goods sold)The main base in Lab is the cost of goods sold; a percentage of revenue excl. VAT is acceptable as a second figure if the base is named. Lab gives no benchmark for the loss level: there are no verified data and the shrink-acc document is aligned with this.

Formulas in notation

Shrinkage
L = S_book − S_count
S_book = S_open + Receipts + Transfers − Sales − Supplier returns. Example: RUB 170,000.
Shrinkage percentage
p = L / COGS × 100%
The percentage base: cost of goods sold. Example: 2.39%. There is no norm benchmark: compare with your own history.
Shrinkage by cause
L_j = s_j × L
s_j share of cause j. Customer theft: RUB 47,600 per month.
Payback of protection measures
T = C / (L_t × ε)
C tag cost, L_t monthly theft loss, ε effect. Example: 28.8 months.

Worked example

Store stock balance for a period and causes of loss

ItemAmount, RUB
Opening stock at cost18,400,000
Receipts6,200,000
Transfers: in − out−150,000
Sales at cost−7,100,000
Returns to suppliers−200,000
Book closing stock17,150,000
Counted closing stock16,980,000
Shrinkage (book − count)170,000
— Record errors and mis-picks, 38%64,600
— Suppliers (short delivery, mixed picks), 12%20,400
— Employees, 22%37,400
— Customers (theft), 28%47,600

Book stock is RUB 17,150,000, the count is RUB 170,000 lower. Shrinkage is 2.39% of sales at cost and 1.28% of revenue (RUB 13,300,000): always name the base. Customer-caused loss is RUB 47,600 per period; if anti-theft tags cost RUB 480,000 a year and cut this part by 35%, the yearly effect is RUB 199,920 and payback 28.8 months. Cause shares in the example are notional: in reality checks determine them and some remains unexplained.

Figures

From book stock to counted stock
Waterfall: Book stock 17.2m, Shrinkage −170k, Counted stock 17m05m10m15m20mBook stockBook stockBook stock: 17.2m17.2mShrinkageShrinkageShrinkage: −170k−170kCounted stockCounted stockCounted stock: 17m17mRUB
  • positive
  • negative
  • total
Data as a table
StepChangeLevel
Book stock—17.2m
Shrinkage−170k17m
Counted stock—17m

From book stock to counted stock. Shrinkage is the difference between book stock (opening plus receipts minus sales and returns) and the count at the end. Shrinkage of RUB 170,000 is 2.39% of cost of sales and 1.28% of revenue; the percentage base is always the same. Illustrative figures.

Causes of shrinkage in roubles
Bar chart: Record errors and mis-picks 64.6k, Suppliers (short delivery, mixed picks) 20.4k, Employees 37.4k, Customers (theft) 47.6k020k40k60k80kRecord errors and mis-picksRecord errors andmis-picksLoss, RUB: 64.6k64.6kSuppliers (short delivery, mixed picks)Suppliers (shortdelivery, mixed…Loss, RUB: 20.4k20.4kEmployeesEmployeesLoss, RUB: 37.4k37.4kCustomers (theft)Customers (theft)Loss, RUB: 47.6k47.6kRUB
Data as a table
MeasureLoss, RUB
Record errors and mis-picks64.6k
Suppliers (short delivery, mixed picks)20.4k
Employees37.4k
Customers (theft)47.6k

Causes of shrinkage in roubles. Cause shares come from the investigation; record errors and theft together give the main part but are cured by different measures.

Payback of anti-theft tags by month
Line chart: Cumulative result, RUB: −480k → 120k−500k−400k−300k−200k−100k0100k200k0102030payback, monthsCumulative result, RUB, 0: −480kCumulative result, RUB, 6: −380kCumulative result, RUB, 12: −280kCumulative result, RUB, 18: −180kCumulative result, RUB, 24: −80.2kCumulative result, RUB, 30: 19.8kCumulative result, RUB, 36: 120kRUB
Data as a table
MeasureCumulative result, RUB
0−480k
6−380k
12−280k
18−180k
24−80.2k
3019.8k
36120k

Payback of anti-theft tags by month. Cumulative saving from reduced customer theft minus tag cost; the zero crossing is the payback time. Tags RUB 480,000, a 35% cut in theft gives RUB 199,920 per year; payback 28.8 months. The assumption is illustrative.

Specifics by category

Apparel
Apparel: theft in fitting rooms and on display, damage, size-record errors; tags on all units and fitting-room control help.
Footwear
Footwear: pairs get separated (one of a pair), size errors in records, fitting without a pair; pair control and one-foot display limits.
Accessories
Accessories: the highest theft risk because of small size and value; locked cases, tags and more frequent counts than apparel.
Kids
Kids: more mis-pick errors from similar sizes and sets; theft is lower but damage and returns are higher.
Men
Menswear: stable sizes and fewer mis-picks, but expensive items (outerwear, suits) attract theft; tags and issue control.
Women
Womenswear: more styles and small items (lingerie, accessories), a higher share of errors and theft of small items; frequent counts of high-risk zones.

Common mistakes

  • Not naming the base of the percentage → losses are compared incorrectly → always state: to sales at cost or retail revenue.
  • Writing off discrepancies without checks → errors and theft are hidden → check against documents and CCTV before write-off.
  • One yearly count → causes are not visible → cycle counting and a count after a change of responsible person.
  • Fraud searched for manually → anomalies are missed → till rules (returns without receipt, manual discounts) and an exception report.
  • Assessing protective measures on supplier promises → the assessment is wrong → a pilot and control stores.

Russian specifics

  • Counts and write-offs: accounting rules for inventories and the procedure for shortages and overages, as well as the liability of responsible persons, are set by law and internal documents; confirm details with the accountant; not verified in Lab.
  • Natural loss and the inventory impairment reserve: expenses are recognized for tax under rules that depend on the tax regime; USN expenses are defined by Art. 346.16 of the Tax Code. Rule: Tax Code Art. 346.16: Tax Code, Art. 346.16 (expenses on the simplified regime).
  • CCTV and data of employees and customers: personal data requirements (152-FZ), notice of filming. Rule: Law 152-FZ of 27.07.2006: Federal Law of 27.07.2006 No. 152-FZ on personal data.
  • Customer returns and fictitious returns: consumer rights (Arts. 25, 26.1 of the Consumer Protection Law) and till receipts (54-FZ). Rule: Law 2300-1 of 07.02.1992: Law of 07.02.1992 No. 2300-1 on consumer rights protection.

Open interactively →