How an independent check uncovered financial misstatement and damage of up to EUR 890,000

It all began with a figure that would not explain itself. Cost of goods sold at one of the production sites was rising faster than at the others, with comparable output volume and comparable product mix. The difference was not dramatic - a few percentage points - but it persisted over several quarters.

Local reporting offered an explanation, and it sounded plausible: rising purchase prices, logistics, exchange rate fluctuations, one-off deliveries for urgent orders. Each of these factors exists, each is capable of moving cost of goods sold. The problem was that each of them explained the deviation individually, but none explained its persistence.

The second discrepancy concerned inventory. Stocktaking showed mismatches between recorded and actual balances for certain items. The discrepancies fell within a range usually written off to natural wastage and accounting error, but their distribution was uneven: some groups of items diverged regularly, others never.

Head office of a company with turnover of more than EUR 60 million approaches such matters without panic. First explanations were requested, then reconciliations, then additional analysis. Every request received a response, and every response was formally correct.

This, in the end, is what prompted the engagement. Not the existence of the discrepancies, but the fact that all of them were explained - and none of them went away. The task was framed in neutral terms: to establish whether the deviations were explained by operational and market causes, or had another origin.

The internal audit worked correctly and within its methodology. The problem is that its methodology answers a different question. An audit checks compliance with procedures; what was needed here was to check whether the procedures themselves matched reality.

Once the discrepancies had been localized, the next question was through whom the relevant transactions passed. Counterparty checks were carried out using officially disclosed information in four jurisdictions, and only in relation to organizations, not the private lives of individuals

The features assessed were the ones standard for this type of review: whether the scale of activity matched the stated volume of deliveries, whether the resources needed for performance were present, the organization's history and profile, overlaps in corporate service arrangements, and the sequence of corporate events over time

For some counterparties, the picture proved internally consistent: active organizations matching the profile and volume. These lines of inquiry were closed and not carried further

For others, features requiring explanation were identified: a mismatch between the stated volume of performance and the observed scale of the organization, overlapping corporate elements between formally independent counterparties, and connections between individual participants and persons associated with the client company

The public wording here is essential, and we maintain it in the report as well. We have not stated, and do not state, that any organization is fictitious or that any employee acted unlawfully. Indications of affiliation and inconsistency requiring legal assessment have been established. No organization name and no individual's name appears, or can appear, in the public version

The flow diagram was built on what the company is entitled to see about its own operations: contracts, invoices, delivery notes, payment orders on its own accounts, warehouse and production records. The review did not, and could not, address the banking secrecy of counterparties, their accounts, or anyone's personal finances

Within this perimeter, the sequence was reconstructed: which position was ordered, on what basis it was paid for, what and in what volume was accepted, what was consumed in production, and how this was reflected in the accounts. Each step was tied to a point in time.

The chronology proved more informative than the amounts. Certain events recurred with a stable periodicity and clustered around specific stages of the operating cycle. The coincidence in time between corporate changes at particular counterparties and changes in the nature of operations became one of the most significant observations.

What could not be established was recorded separately. Part of the scope lay outside what was accessible to the company: transactions between third parties, their mutual settlements, and the subsequent movement of funds. These areas are marked in the diagram as unestablished - and it was precisely there that, on our recommendation, the client's lawyers were to work through procedure rather than through a private check.

The result was a diagram with clearly marked gaps. A diagram without gaps raises more doubt in such matters than one with them: it means that assumptions have been presented as facts.

The estimate of possible damage ranged from EUR 620,000 to EUR 890,000. A range, not a figure - and this is not caution for its own sake.

The lower bound is built on positions where the discrepancy is documented and confirmed on several independent grounds at once. The upper bound includes positions where the discrepancy is established, but its full extent depends on the interpretation of disputed transactions and on the period taken into account. The difference between the bounds is precisely the measure of uncertainty, and it is shown honestly.

Here it is necessary to state the status of this figure directly. This is an estimate of possible damage, not a sum determined by a court and not a sum that has been or will be recovered. The verification created a basis for procedure; the outcome of that procedure depends on the court, on the evaluation of evidence, and on a range of circumstances beyond the analytical team's control.

Each item in the material was given a status: documented, assessed with an indicated level of confidence, or not established. This distinction is essential for further work: part of the material is suitable for procedure directly, part - only for understanding the picture and framing questions.

And one more limitation, repeated several times in the report. Established indicators of financial statement distortion and indicators of affiliation do not constitute a finding of fraud. Such a determination is made by a court, and until its decision the correct wording is precisely indicators.

The most useful part of such projects is not what was established, but what the company changes afterward. Below are the measures the client implemented as a result; they are general and do not disclose details of the matter.

First. Regular cross-layer reconciliation. Matching of accounting data against the physical scope and cash flow, carried out systematically rather than in response to suspicion. Most discrepancies are visible precisely between the layers.

Second. Verification of counterparties beyond the register. Whether the scale of the organization corresponds to the claimed volume of supply, whether it has the resources to perform, and the stability of its activity profile. Registration in the register confirms existence, not the capacity to perform.

Third. Separation of authority. The selection of a supplier, the agreement of terms, acceptance, and confirmation of payment must not converge in one set of hands or within one narrow circle. This is a basic measure, and it is precisely the one most often relaxed for the sake of speed.

Fourth. Attention to systemic minor details. Deviations below the materiality threshold must be analyzed for recurrence. A stable minor pattern carries more weight than a single large deviation.

Fifth. A predetermined course of action in case of suspicion. Who decides on a check, what data is recorded first, how to prevent premature disclosure. Improvisation at such a moment is most often what destroys the possibility of gathering material.

And last. Nineteen days and the range given are the circumstances of this matter, not a norm. With different quality of records, different geography and a different volume of available data, the same work may take substantially longer or yield substantially less.

Indicators of a task that is resolved by an independent check rather than an internal audit. This is not a diagnosis and not a promise of a result.