Engine-metric self-declarations are the most consequential paperwork most SAP customers fill in each year, and they are routinely treated as administrative housekeeping. The submission template arrives by email in late summer, the analyst pulls the volumes from the standard transactions, and the response goes back to the SAP licence team. The numbers in that response set the licence baseline for the next contract year, drive the maintenance bill, and become the data of record in any subsequent audit. Errors in the submission propagate through every commercial interaction with SAP until the next measurement window opens.
This guide walks through the standard self-declaration workflow, identifies the six places where customers most commonly overstate the position, and explains the reconciliation steps that drop the declared number toward the defensible minimum without crossing the line into non-compliance.
What the self-declaration actually does
The annual self-declaration is the customer's own statement of the consumption of each licensed engine metric for the measurement period. Where USMM and the License Administration Workbench (LAW) handle named users and certain technical metrics automatically, engine metrics frequently require manual declaration because the underlying measurement is bespoke to the engine in question — order line counts, payroll records, document volumes, kilowatt-hours of HANA compute, hectolitres of beverage produced, and so on.
The declared number is the legal starting point for the next year's licence position. If the declared number exceeds the contracted entitlement, SAP raises an uplift invoice at list price minus the customer's negotiated discount. If the declared number is below the entitlement, the licence position is reaffirmed and no commercial action follows. Either way, the declared number is what the auditor will hold the customer to in any subsequent dispute. See our order-to-cash engine explainer for the contractual definition of one of the most contested engine families.
Trap one — counting documents that are not chargeable
Every engine has a contractually defined unit of measurement. For the FI engine the unit is the financial document; for OTC it is the sales order line; for the procurement engine it is the purchase order line. The contract definition almost always carves out specific document types from the chargeable count — internal transfers, intercompany postings, technical reversal documents, automated correction postings.
The default extracts from SAP standard transactions include the carve-out volumes. A customer who submits the raw extract figure overstates the chargeable count by anywhere from eight to forty per cent depending on the engine and the business model. Reconciliation against the contractual definition is the single largest source of declared-number reduction in our practice.
Trap two — including non-production system volumes
Engine metrics are chargeable for production system volumes only. Sandbox, development, quality assurance, training, and disaster-recovery system volumes are not chargeable, but the standard transaction extracts run against the consolidated landscape will pick them up if the system filter is not applied carefully.
The trap is particularly acute for landscape-consolidated extracts that aggregate across SIDs. Customers running a hub-and-spoke architecture, or operating BW alongside the source ECC system, can easily double-count by pulling volumes from both the source and the consolidated reporting layer. The remediation is to extract only from each production SID and to document the exclusion.
Trap three — failure to apply the contractual measurement window
Each engine has a measurement window defined by the contract — typically a calendar year or a fiscal year. The default extracts return the rolling twelve months from the date of extract, which is rarely the contractual window. A customer who extracts on 15 September and submits the rolling-twelve number is declaring against the wrong period.
The implication is bidirectional. In a year of growing volumes, the rolling-twelve overstates the position relative to the prior fiscal year. In a year of declining volumes, it understates. Either way the declared number is not the contractually defined number, and the auditor will challenge it.
Trap four — applying the wrong unit of measurement
The HR engine is the canonical example. The contract specifies an FTE count, but the standard extract returns active employee head-count. The two numbers are close but not identical: contractors, part-time workers, multiple-assignment employees, and employees on extended leave each have a contractual treatment that diverges from a simple head-count read. A customer who declares head-count when the contract requires FTE has overstated the position by between three and twelve per cent. See our HR FTE engine analysis for the precise treatment.
Trap five — failure to net off legacy carve-outs
SAP contracts entered into before specific commercial events frequently contain carve-outs for legacy business lines, divested entities, or specific use cases. A 2014 master agreement may exclude volumes generated by a business unit that was divested in 2017; an extension signed at the time of an acquisition may include volumes that were ring-fenced during integration.
These carve-outs are recorded in the contract paper but rarely surface in the workflow that drives the annual declaration. We routinely identify carve-outs that have been ignored for three to five consecutive declarations, with cumulative back-overstatement in the seven figures.
Trap six — incorrect aggregation across engines
Some engines aggregate. The Solution Manager engine combines multiple subordinate metrics; the BW engine charges on data volume across all sources; the CRM engine has separate counts for accounts, contacts, and activities that combine into a composite measure. Customers who declare each underlying metric separately and add them up are normally over-declaring because the aggregation formula in the contract is more sophisticated than simple addition.
The fix is to read the contractual aggregation formula, apply it once to the underlying numbers, and submit the formula output rather than the sum.
The reconciliation workflow we use in practice
A defensible declaration follows a four-step reconciliation that takes ten to fifteen working days for a typical mid-market estate.
Step one is contract definition extraction. For every engine on the licence, pull the contractual definition of the chargeable unit, including all carve-outs, exclusions, and aggregation formulae. The output is a one-page reference per engine that the analyst running the declaration can reconcile against.
Step two is system-by-system extraction. Pull the chargeable volumes from each production SID separately, with the contractual measurement window applied. Document each exclusion explicitly so the reconciliation is auditable.
Step three is reconciliation against the prior-year declaration. Compare the current-year extract to the prior year, by engine, and explain any variance greater than five per cent. Most variances surface either a genuine volume change, an extraction error, or a missed carve-out. All three are useful to identify.
Step four is the submission file with footnoting. The declaration goes back to SAP with explicit footnotes naming the contractual carve-outs that have been applied. The footnotes are the audit defence: they make the reconciliation visible to SAP and stop the auditor reverse-engineering the declared number against the raw extract.
Where engine declarations and S/4HANA conversion intersect
Customers who are mid-conversion from ECC to S/4HANA face a specific complication. The engine metrics convert on a path defined by the conversion paperwork, and the declaration in the conversion year often needs to reflect partial-year ECC volumes and partial-year S/4HANA volumes. The conversion mathematics are detailed in our analysis of how engine metrics convert under S/4HANA and in the conversion economics chapter of the S/4HANA migration economics white paper.
For the related case study on engine-metric overstatement and reclassification, see the consumer goods engine-metric reconciliation matter.