Lesson 5 of 5
Testing splitting
7 min
Document splitting is tested by business transaction, not by screen. A supplier invoice that posts without an error message proves almost nothing; the question is whether every segment and profit center balances, whether the split follows the right base, and whether the follow-on documents (payment, clearing, reversal) keep the split. The final evidence is a balanced trial balance for each segment and each profit center.
Build a coverage matrix
List the business transactions Contoso uses and, for each, the document types that map to it. Then add the variations that stress splitting: multiple segments in one document, partial payments, residual items, down payments, cash discounts, reversals, and documents from logistics (goods receipt, goods issue, billing) which are generated by the system and never seen by a finance user.
| Business transaction | Variations to test | Expected split |
|---|---|---|
| Supplier invoice | One segment; two segments; with tax; with unplanned delivery costs | Payable and tax in proportion to expenses |
| Customer invoice (from billing) | Fleet and Aftermarket items on one invoice | Receivable and tax in proportion to revenue |
| Outgoing payment | Full, partial, with cash discount | Bank and payable follow the invoice split |
| Incoming payment | Full, residual item, with difference written off | Residual keeps the original split |
| G/L transfer between profit centers | Same account, different profit centers | Zero-balance lines added |
| Reversal | Reverse any of the above | Exact mirror of the original split |
Contoso's document splitting coverage matrix (extract).
Positive and negative tests
Positive tests post valid documents and check the G/L view. Negative tests try to post documents that should fail: a G/L entry with no profit center when the characteristic is mandatory, a manual posting to the zero-balance clearing account, a document type used for a business transaction it is not allowed for. A negative test passes when the system rejects the posting with the expected message. Without negative tests you only know the rules work on good data.
- PostCreate the document with known segment amounts
- Read G/L viewCheck lines per characteristic
- Follow-onPay, clear, reverse; check the passive split
- ReportTrial balance per segment and profit center
- RecordDocument numbers, views and report extracts as evidence
The evidence that counts
After the test cycle, run the trial balance (for example the Trial Balance app or a financial statement report) with segment and then profit center as a drilldown. Each segment must balance to zero over all accounts. Then check three specific things: the default profit center has no postings you did not plan; the zero-balance clearing account is zero for the company code; and the totals per segment match the amounts you designed into the test documents.
- BuildUnit test rulesConfiguration team tests each business transaction variant
- SITIntegrated transactionsLogistics documents (billing, goods movements) checked for split
- Dry-run closeSegment trial balanceFull month, balanced by segment and profit center
- UATKey users read G/L viewsMaria and the AP team confirm their documents
- Go-liveMonitor default valueDaily report on default profit center postings for the first close
Design the test data so the expected answer is obvious. Use round amounts and fixed ratios, for example every multi-segment supplier invoice split 60:40 between Fleet and Aftermarket, so that anyone reviewing the G/L view can see at a glance whether the split is right. Record the designed amount per segment in the test case before execution. After the cycle, the expected segment totals are simply the sum of the designed amounts, which gives you an independent expectation to compare the trial balance with, rather than reading the answer off the system and calling it expected.
What would you do?
Priya's SIT report shows all splitting test cases "passed" because every document posted without an error. Dan asks if he can sign off. What should Priya add first?
- Test splitting by business transaction, including follow-on payments, clearing and reversals.
- Include logistics-generated documents and negative tests.
- The key evidence is a balanced trial balance per segment and profit center.
- Check the default value and the zero-balance clearing account as part of every test cycle.
Check your understanding · not graded
Lesson check
1. What is the key evidence that document splitting works?
2. What does a passing negative test look like?
3. Why test reversals in a splitting test?