In this guide
Auditing by IMEI and Serial Number
Serial verification tests identity rather than quantity: every unit selected is matched to a specific recorded serial or IMEI, so the question moves from how many units are here to whether these are the units the records describe. That distinction matters because a quantity match can be perfectly correct while the individual units are wrong. Units sold and not despatched, units substituted with a lower specification, warranty replacements never recorded, and demonstration stock swapped for saleable stock all leave the count intact and the record false. High value justifies the effort because the cost of scanning a unit is fixed while the exposure per unit is large, so the same procedure that would be uneconomic for consumables is straightforward here. What the verification proves is that a named unit was physically present at a named location on a named date, which is the only form of inventory evidence that survives a dispute about which units were actually there.
When Serial-Level Counting Is Necessary
Serial-level work costs several times what a quantity count costs, so it is applied where the return justifies it rather than everywhere. Value per unit is the first test. Where a single unit represents a meaningful amount, the cost of capturing its identity is trivial against the exposure, and the same procedure applied to low-value interchangeable goods would be indefensible. The threshold is a judgement about the population rather than a fixed figure. Warranty and after-sales obligations are the second driver, and they are frequently the stronger one. A business that has to establish which unit went to which customer, when the warranty started, and what has already been claimed against it cannot do any of that from quantities, so serial capture is a requirement of the commercial model rather than an audit preference. Lender or insurer requirement is the third. Where units are financed individually, as in floor-plan arrangements, the lender's security is a list of identified units and only identity verification tests it. An insurer settling a loss will likewise want to know which units were lost, since a claim asserting a number of units without identifying them is considerably harder to substantiate.
Capturing Serials at Scale
The practical question is how to record thousands of identifiers without the exercise becoming unaffordable. Scanning against manual capture is the first decision and it is usually settled by error rate rather than by speed. A scanned serial is captured correctly or not at all; a manually keyed one carries a transcription error rate that is small per unit and substantial across a large population, and every error becomes an exception somebody has to investigate. Where the serials are barcoded, scanning is almost always the right answer even allowing for equipment. Sealed cartons pose the question the exercise turns on. Opening every carton to read the units inside destroys the packaging, takes an enormous amount of time, and may not be commercially acceptable; not opening them means accepting the carton label as evidence of its contents. The workable answer is a sampling rule agreed in advance: a stated proportion of cartons opened and verified, selected across models and locations, with the result extrapolated and the basis disclosed. Time cost per hundred units is what determines the whole plan, and it should be measured on a pilot rather than estimated, because it varies enormously with how the stock is stored.
Duplicates, Gaps and Mismatches
Serial verification produces three classes of exception that a quantity count can never generate, and each points at something different. The same serial in two places is the most serious. A serial number is unique to a unit, so two physical units carrying it means one is counterfeit, one has been relabelled, or a data entry has duplicated a record, and each possibility warrants a different response. Where the duplicate appears across two locations it may also indicate a transfer recorded at one end only. Serials in the system with no physical unit are the population that corresponds to a shortage, but with far more information attached. Because each missing serial is identified, its history is available: when it was received, where it was recorded, whether it was ever picked, and whether it appears in any despatch. Most resolve into unrecorded sales or unrecorded transfers. Units on the floor with no system record run the opposite way. Physical units carrying serials the records have never held usually mean goods received and never booked, warranty replacements issued without a corresponding entry, or units returned and put back without a receipt.
Evidence a Serial Audit Produces
The output is a listing rather than a total, and that is what makes it useful afterwards. Every serial captured is reconciled against the stock record, producing three populations held separately: serials present and recorded, serials recorded but not found, and serials found on the floor that the record does not carry. The third population is the one a quantity count can never produce, and it is where substitutions and unrecorded warranty replacements surface. Exceptions are listed by category rather than in one sequence, because the categories carry very different unit values and a mixed exception list forces the reader to sort it before it means anything. Each exception carries the location it was expected at and the location it was found at where both are known. The listing then supports claims that a quantity count could not. An insurance claim naming specific units lost is settled on a different basis from one asserting a number of units; a warranty position rests on which unit was supplied to whom; and a recall reaches the units affected rather than the whole population.
Deciding How Deep to Go
Full serial capture is not always the right answer, and deciding where to apply it is the practical question. Capture every serial where the unit value is high, where units are individually warranted, where recall exposure exists, or where the stock is financed unit by unit. Sample serials elsewhere, using the sample to test whether the serial records are reliable rather than to enumerate the population, which is a legitimate and much cheaper objective. The categories that justify unit-level work are usually a small proportion of the lines and a large proportion of the value, which is exactly the pattern that makes the decision easy once the analysis is done. Doing it by category rather than by instinct is what prevents effort being spent uniformly across a population that is not uniform. A specialist count is worth commissioning where serials must be reconciled to a lender's funded list, where an insurance claim depends on identifying specific units, or where a previous count found substitutions, and Electronics & White Goods engagements are built around unit-level identity.
