Lesson 5 of 5
End-to-end cross-functional testing
12 min
These scenarios cannot be tested by one person in one module. A third-party test needs a sales user, a buyer, an accounts payable clerk and a billing clerk. An intercompany test needs users with access to two company codes and a finance reviewer for each. This lesson shows how to plan and run them so the evidence holds up at sign-off.
Plan the roles and data
| Item | Third-party (Contoso) | Intercompany (Contoso) |
|---|---|---|
| Products | CMC-BRK-500 as a third-party item; purchasing info record with Lakeside | CMC-ALT-100 with stock in plant 1810 |
| Partners | Great Lakes Fleet Services; Lakeside Components | Great Lakes Fleet Services; internal customer for sales org 1710; intercompany supplier in 1710 |
| Prices | Customer price $45; purchase price $30 | Customer price $200; internal price $170; 1810 cost $148 |
| Roles | Sales rep, buyer, AP accountant, billing clerk | Sales rep, shipping specialist, billing clerk, AP accountant in 1710, controllers for 1710 and 1810 |
| Evidence owners | Maria (sales), Dan (finance) | Maria (sales), Dan and the 1810 controller |
Data and roles to prepare before running the scenarios.
Test with real business roles, not a superuser. Many cross-functional failures are authorization gaps, for example a buyer who cannot see account-assigned requisitions, or an AP clerk in 1710 who cannot display the intercompany supplier. A superuser hides all of them.
The scenario set
| # | Scenario | Key expected result |
|---|---|---|
| 1 | Third-party, full quantity, without shipping notification | Requisition, PO, supplier invoice, customer invoice; margin $1,500 on the sales order item |
| 2 | Third-party, two partial supplier invoices | Two customer invoices matching the two supplier-invoiced quantities |
| 3 | Mixed order: one stock item, one third-party item | Stock item delivered from 1710; third-party item procured; both billed correctly |
| 4 | Intercompany, full quantity | Four journal entries; 1810 receivable equals 1710 payable; trading partner on both |
| 5 | Intercompany, partial delivery | IV and customer invoices for the delivered quantity only |
| 6 | Intercompany, missing internal price (negative) | Order incomplete or blocked per configuration; not silently priced at zero |
| 7 | Intercompany period-end cut-off | Goods issue on the last day; IV and customer invoices in the same period |
| 8 | Return of an intercompany sale | Customer credit in 1710 and intercompany credit in 1810, both sides matching |
Cross-functional end-to-end scenarios.
- Hand-off logEach role records the document numbers they create and passes them on
- Document flow, sales sideExpanded flow from the sales order showing all successors and statuses
- PO history (third-party)Requisition, PO, invoice receipt for the account-assigned item
- Journal entries per companyEach posting checked against the expected accounts and amounts
- Intercompany pair checkReceivable in 1810 against payable in 1710, period and trading partner
Running across teams
- Day 1 AMSalesOrders for scenarios 1–8 created; numbers logged
- Day 1 PMPurchasing and shippingPOs from requisitions; intercompany deliveries and goods issue in 1810
- Day 2Payables and billingSupplier invoices; customer and IV invoices; inbound intercompany invoices checked
- Day 3Finance reviewJournal entries and intercompany pairs checked by both controllers
- Day 3 PMTriageDefects logged with the last good document and the failing seam
Writing expected results for two companies
Expected results for cross-company scenarios are best written as a small ledger before the test starts: one row per document, one column per company code, with the accounts and amounts you expect. For scenario 4 that is four rows (goods issue and IV invoice in 1810, customer invoice and inbound supplier invoice in 1710). After execution, the tester fills in the actual document numbers and amounts next to each row. Any empty cell is either a missing document or a missing posting, and any mismatch is a defect with its location already identified. This format also makes the controllers' review fast, because they can compare a single table rather than hunting through screenshots.
What would you do?
Priya proposes running scenario 4 (intercompany) with a single test user that has access to both company codes, to save time. Grace asks whether that is acceptable for the sign-off cycle.
- Plan products, partners, prices, roles and evidence owners before execution.
- Use real business roles in each company; a superuser hides authorization gaps.
- Cover full and partial flows, mixed orders, a negative pricing test, period cut-off and returns.
- Evidence: document flow, PO history, journal entries in every company and the intercompany pair check.
- Triage defects by the last good document and the seam where the chain broke.
Check your understanding · not graded
Lesson check
1. Why should cross-functional tests use real business roles?
2. What additional evidence does a third-party test need beyond the sales document flow?
3. What does the intercompany period-end test prove?
4. Which scenario is the negative test in the set?