Lesson 5 of 5
Testing budgets and settlement
9 min
Listen to this lesson
Internal orders look simple, which is why they are often tested lightly. A test that creates an order, posts an invoice and runs settlement without error proves very little. The two questions that matter for sign-off are: does availability control stop spending at the agreed limits, and do the costs land on the right receivers, in the right amounts, leaving the order at zero?
Testing budget limits
Budget tests are boundary tests. Design postings that land just below and just above each tolerance limit, through each path that creates assigned value: a purchase requisition, a purchase order, a direct invoice and a goods issue. Use round numbers so the expected message is obvious. For a USD 10,000 test budget with Contoso's limits, a posting that takes the assigned value to USD 8,900 should give no message, USD 9,100 a warning, USD 10,200 a warning with mail, and USD 10,600 an error.
| Test | Assigned value after posting | Expected result |
|---|---|---|
| T1 Direct invoice | USD 8,900 (89%) | Posted, no message |
| T2 Purchase order | USD 9,100 (91%) | Posted with warning |
| T3 Goods issue of demo stock | USD 10,200 (102%) | Posted with warning; mail received by budget manager |
| T4 Direct invoice | USD 10,600 (106%) | Blocked with error message |
| T5 Budget supplement, then repeat T4 | USD 10,600 against USD 12,000 | Posted without message |
| T6 Exempt cost element | Any | Posted; not checked by availability control |
Boundary tests for a USD 10,000 test order (Contoso tolerance design).
T3 needs evidence outside the posting screen: the mail in the budget manager's inbox. T4 is a negative test; it passes when the posting is refused with the expected message. T5 proves the supplement path works. Each test should record the order number, the document or message number and a screenshot or log of the message text.
- Did the posting go through when the assigned value stayed under the warning limit?Yes → pass · No → check budget, profile and exempt cost elements
- Did the right message appear at each limit?Yes → next question · No → compare tolerance settings with the design
- Did the error block the document, and did a supplement unblock it?Yes → budget control signed off · No → defect on budget profile or tolerance group
Testing settlement outcomes
Settlement tests compare an expected receiver table, prepared before the run, with what the system posted. For Fleet Expo with USD 64,000 of actual cost and a 70/30 rule, the expectation is USD 44,800 to Detroit marketing and USD 19,200 to Chicago marketing, with the order credited USD 64,000. Prepare one such table for each settlement pattern in scope: cost center split, investment order to asset under construction, final settlement to a fixed asset, and settlement to a G/L account.
- ExpectWrite receiver amounts from the rule before the run
- Test runCompare the log with the expected table
- Live runPost settlement in the test client
- OrderBalance zero after full settlement
- ReceiversCost center, asset or G/L shows the exact amounts
Evidence that convinces
- The order's cost report after settlement, showing debits, the settlement credit and a zero balance.
- The receiver side: cost center line items, the asset's value or the G/L line, each matching the expected table.
- The settlement log with no errors, and the document numbers created.
- For investment orders, the asset under construction cleared and the final asset capitalized on the right date.
- For a reversal test, the same values restored on the order and removed from the receivers.
What would you do?
Priya's test log says "Settlement executed, no errors" for the investment order test. Dan asks whether he can sign off. What is missing?
- Test availability control at each tolerance limit, from below and above, through every path that creates assigned value.
- Negative tests pass when the posting is blocked with the expected message.
- Prepare expected receiver amounts before running settlement.
- Evidence: zero order balance after full settlement and exact amounts on each receiver.
Check your understanding · not graded
Lesson check
1. Which result proves a full settlement worked?
2. A posting that pushes the assigned value over the error limit is blocked. How is this negative test rated?
3. When should the expected receiver amounts be written down?
4. Fleet Expo settles USD 64,000 with a 70/30 rule. What should the Detroit marketing receiver show?