Lesson 5 of 5
Testing period-end order results
9 min
Listen to this lesson
Period-end for production orders is one of the hardest areas to test in a conversion, because the results depend on order states that build up over weeks. A test that settles one order that was created and finished the same afternoon proves only the simplest path. Contoso's plan, owned by Priya with Dan as sign-off, tests a portfolio of order states across two periods.
Design an order portfolio
| Order state at period-end | Expected WIP | Expected variance | Expected balance after settlement |
|---|---|---|---|
| Released, components issued, nothing delivered | Actual costs | None | Equal to WIP |
| Partially delivered (PDLV) | Actual costs minus GR credits | None | Equal to WIP |
| Delivered (DLV) in this period, WIP last period | Cancelled | Balance by category | Zero |
| Technically completed (TECO) without full delivery | Cancelled | Balance, including lot size effect | Zero |
| Delivered with extra scrap | Cancelled | Includes scrap variance | Zero |
| Order without variance key (negative test) | As designed | Not calculated | Follow the design decision; must be visible in the log |
Contoso's order portfolio for period-end testing.
Build the portfolio in the first period and finish part of it in the second, so WIP creation, WIP cancellation and variances are all exercised. For each order, write the expected WIP, variance categories and FI postings before running anything. Use simple, designed deviations, such as 20 extra labor hours at USD 52, so each expected category value is easy to calculate by hand.
- Period 1, days 1-25Create the portfolioRelease orders, issue components, confirm operations, partial receipts
- Period 1, closeFirst closeOverhead, WIP, variances, settlement; check WIP postings
- Period 2, days 1-25Finish selected ordersFinal receipts, TECO, extra scrap on one order
- Period 2, closeSecond closeWIP cancelled, variances by category, balances zero
Reconcile three views
Every test cycle should reconcile three views of the same values. First, the orders: each order's WIP and variance against the expectation. Second, the general ledger: the unfinished goods inventory account balance should equal the sum of WIP on open orders, and the price difference accounts should equal the settled variances. Third, margin analysis: variance categories should appear on the expected accounts with product and profit center.
- OrdersWIP and variances per order match expectations
- WIP accountG/L balance equals sum of WIP on open orders
- Price difference accountsEqual to settled variances, by category if split
- Margin analysisVariance lines carry product and profit center
- Order balancesZero for every DLV or TECO order
Typical defects to look for
- A finished order whose previous WIP is still in the WIP account after settlement: WIP was not recalculated after DLV.
- Variances in "remaining variance" only: the variance variant or target cost version is incomplete.
- No variance on an order with an obvious overrun: missing variance key.
- WIP account in FI not equal to the WIP list: settlement skipped some orders or posting rules point to the wrong accounts.
- Price differences posted without profit center: derivation gap on generated lines.
What would you do?
In the second test close, all orders settled without errors. The WIP account in the G/L shows USD 48,700, but the WIP list for open orders totals USD 52,100. What should Priya do?
- Test a portfolio of order states across two periods, not one order in one afternoon.
- Write expected WIP, variance categories and postings before running the close.
- Reconcile orders, WIP and price difference accounts, and margin analysis.
- Evidence: zero balances on settled finished orders and variance reports that match expectations.
Check your understanding · not graded
Lesson check
1. Why does period-end testing need two periods?
2. What should the unfinished goods inventory account in the G/L equal after settlement?
3. A finished order's WIP from last period is still in the G/L after settlement. What is the likely cause?
4. Order 1001200 used 15 extra rotors at the USD 40 standard. What input quantity variance should the test expect for the rotors?