Lesson 4 of 5
Reconciliation
11 min
Listen to this lesson
Reconciliation is how finance proves the conversion did not change the books. It compares the before picture from ECC with the same view of the converted system, area by area, and explains every difference. The principle is simple; the discipline is in planning it before Mock 1 so that the same reports, selections and people are used every time.
Three layers of reconciliation
The technical layer belongs to Sam's Basis team and the migration specialist. The migration layer is the reconciliation the finance migration itself produces: it is evidence, and a difference there that nobody explains becomes a finance incident later. The business layer is what Dan signs: reports finance already uses, run with the same selections before and after.
What to compare
| Area | Before (ECC) | After (SAP S/4HANA) | Tolerance |
|---|---|---|---|
| General ledger | Trial balance by company code 1710 and 1810, ledger 0L and 2L, account and period | Same report from the universal journal | Zero |
| Customer and supplier open items | Totals and counts per reconciliation account | Same lists after migration | Zero |
| Asset accounting | Asset history sheet by depreciation area | Asset history sheet after asset migration | Zero, per area |
| Inventory | Stock value by plant 1710, 1720 and 1810 and material | Material ledger values | Zero |
| Controlling | Cost center actuals for the open year | Cost center actuals from the universal journal | Zero |
| Document counts | Documents per company code and fiscal year | Same counts | Zero, or explained |
Contoso's reconciliation set. Tolerances are zero: a converted ledger should not drift by a cent.
Use the same report variant, currency, ledger and period in both systems. Keep the outputs, not just a tick: auditors want to see the two reports and who compared them. SAP's Data Transition Validation tool can extract the before data during ramp-down and compare it after the conversion, which turns some of this into a repeatable job. Where it does not cover a report, a spreadsheet comparison with a reviewer's initials is still evidence.
| Area | Report run before and after | SAP Fiori cross-check after conversion |
|---|---|---|
| General ledger | G/L account balances (for example S_ALR_87012277), per ledger 0L and 2L | Trial Balance (F0996A) |
| Customer open items | FBL5N with the same key date and variant | Manage Customer Line Items (F0711) |
| Supplier open items | FBL1N with the same key date and variant | Manage Supplier Line Items (F0712) |
| Fixed assets | Asset history sheet per depreciation area | Asset History Sheet app |
| Inventory | MB5L stock values by plant, then material ledger values | Inventory valuation apps, as used by Contoso |
| Controlling | Cost center actuals (for example S_ALR_87013611) | Cost center actuals apps, as used by Contoso |
Compare like with like: the same classic report before and after is the reconciliation; the Fiori app is a second look, not the proof.
Handling a difference
- Is the difference also in the migration reconciliation results?Yes → analyze the migration step that produced it · No → check report selections, currency and ledger first
- Was it already in ECC (an old inconsistency)?Yes → document it, agree treatment with Dan, add the fix to preparation · No → next question
- Can it be fixed within the float?Yes → fix, re-run the affected step, reconcile again · No → raise to the go/no-go owner with options
Most differences in early mocks come from three places: an inconsistency already in ECC, a report run with different selections, or a migration step that hit an error and was not repeated. By the dress rehearsal the reconciliation should run clean, and the real weekend should hold no new kind of difference.
Plan the reconciliation like a test
Treat each reconciliation row as a test case with an owner, a report and variant, an expected result (zero difference), a time slot in the runbook and a place to store the evidence. Assign rows to the people who know the numbers: the AR lead for customer open items, the asset accountant for the history sheet, a controller for cost center actuals. Rehearse the timing too. On Contoso's weekend the reconciliation sits between the end of the finance migration and go/no-go 2, so a slow report or a missing person pushes the whole decision back. Running the reports in parallel by company code saved Contoso almost two hours between Mock 1 and Mock 2.
What would you do?
In Mock 2, the trial balance for company code 1710 matches in ledger 0L but differs by 2,140.00 USD in ledger 2L on one account. Luis asks whether the team can proceed with the rest of the mock. What should the migration specialist do?
- Reconcile in three layers: technical, migration results and business reports.
- Compare G/L, open items, assets, inventory, CO actuals and counts with zero tolerance.
- Use identical selections and keep both outputs and the reviewer's name.
- Route every difference: selections, migration step, old ECC problem, float.
Check your understanding · not graded
Lesson check
1. What tolerance should Contoso apply when comparing trial balances before and after conversion?
2. Which SAP tool can extract business data during ramp-down and compare it after conversion?
3. A difference appears in a mock. What should be checked first?
4. Why should reconciliation outputs be kept rather than just ticked off?
5. Contoso’s validation starts Sunday 10:00 and go/no-go 2 is at 14:00. What makes the reconciliation fit?