Lesson 6 of 6
Testing master data
9 min
Listen to this lesson
Production master data is tested the same way you would test a recipe: cook it once and check the result. The two cheapest and most revealing tests are a probe production order and a standard cost estimate. Both read the product views, the production version, the BOM, the routing and the work centers. If both come out as expected, the master data is very likely right.
- Write the expected resultComponents, quantities, operations, work centers, dates and cost per alternator
- Create a probe orderProduction order for 100 CMC-ALT-100 in plant 1710, CMQ client 100
- CompareComponents, operations, dates and planned costs against the expectation
- Run the cost estimateStandard cost estimate for the alternator; read itemization and log
- Record evidenceOrder number, estimate, screenshots or exports, differences and decisions
The probe production order
Create a production order in the test client (transaction CO01 or the Fiori app) with a realistic quantity, and do not change anything the system proposes. Then compare:
- Production version selected: the one you expected for that quantity and date.
- Components: every BOM item with quantity scaled correctly (100 rotors, 100 stator cores, 100 housings, 100 fastener kits), allocated to the right operation, with the right storage location and backflush flag.
- Operations: 0010, 0020, 0030 with the right work centers and control keys, and durations that match setup plus per-piece times.
- Dates: scheduled start and finish consistent with the lead times and the factory calendar.
- Planned costs: components and activities at plausible values; no zero-cost operations.
The standard cost estimate
Run a standard cost estimate for the alternator (transaction CK11N or the Fiori costing apps, with the costing variant your project uses). Read the itemization: each material line should match a BOM item, each activity line should match a routing operation and activity type, with a price. Read the costing log: messages about missing prices, missing versions or invalid BOMs are findings, not noise. For Contoso, the expected result is about $150 per alternator in plant 1710; a result of $90 or $300 is a master data defect until proven otherwise.
| Symptom | Likely master data cause | Where to fix |
|---|---|---|
| No components on the order | BOM usage, plant or validity; no production version | BOM, production version |
| No operations on the order | Routing not in the version, routing status or lot size | Routing, production version |
| Operation with zero cost | Missing activity price or activity type | Controlling planning, work center |
| Start date in the past | Lead times too long, calendar wrong | MRP 2 / Work Scheduling, routing |
| Components needed too early | No component allocation | Routing component allocation |
| Cost far from $150 | BOM quantity, base quantity, standard values or prices | BOM, routing, controlling |
Negative and mass checks
Add a few negative probes: a quantity outside every version's lot size range should fail clearly; an order dated before a component's validity should explode the old component. For mass coverage, use reports: multi-level BOM explosion, where-used lists, production version consistency checks and a costing run for the whole product range. In SAP S/4HANA Cloud Public Edition, the same checks run through Fiori apps only.
What would you do?
The cost estimate for CMC-ALT-100 comes out at $150.20, but the costing log shows a warning that the test operation has no activity price. What should you do?
- A probe production order and a standard cost estimate test all production master data at once.
- Write expected components, operations, dates and cost before you run the probe.
- Read the cost estimate's itemization and log, not just the total.
- Add negative probes and mass checks; sign off with evidence and an owned list of deviations.
Check your understanding · not graded
Lesson check
1. Why write the expected result before creating the probe order?
2. Which part of the cost estimate shows each component and activity line?
3. A probe order has no operations. Which master data do you check?
4. What is a good negative probe for production versions?