Lesson 6 of 6
Mock cycles, dress rehearsal, the evidence pack and hypercare
8 min
Listen to this lesson
This course discusses ACCELQ by name. Matt Angerer, who runs The Sign-Off Table, works at ACCELQ. Disclosure
Every check in this course pays off only if it runs the same way in every rehearsal and on the go-live weekend, and if its result ends up in front of the people who sign. This lesson puts the pieces into a cycle plan for Contoso's Wave 2, defines what tightens from one cycle to the next, lists the evidence pack, and sets the conditions for leaving hypercare. Wave 2 is in progress: W2 Mock 1 starts in late October 2026, and the go-live weekend is planned for Friday, February 5, 2027, 18:00 to Sunday, February 7, 10:00, a 40-hour window.
Each cycle is stricter than the last
SAP Business Transformation Center formalizes a useful pattern for its lean selective data transition: a test cycle, in which errors can be bypassed so the team can tune iteratively; a dress-rehearsal cycle, intended to practice the productive cutover with parameters close or equal to it, in which errors block all further steps; and a go-live cycle. Contoso's Wave 2 uses the Migration Cockpit, not BTC, but adopts the same discipline: what may be waived in W2 Mock 1 blocks in the dress rehearsal.
| Cycle | Purpose | Rules and data | Exit criteria |
|---|---|---|---|
| W2 Mock 1 | Every object attempted; problems known | Draft mappings; subset allowed | All nine reconciliation layers run; baseline and rerun regression done; defects logged |
| W2 Mock 2 | Full volume; errors explained | Full extract; mappings signed | All differences explained; volume measured; interfaces passed end to end |
| W2 Mock 3 | Timing inside the 40-hour window | Frozen rules and runbook | Critical objects near zero errors; window met with contingency; untouched check clean |
| W2 dress rehearsal | Practice the weekend | Production rules, production-like timing | No blocking differences; go/no-go criteria all evidenced; fallback rehearsed |
| Go-live | Production load | As rehearsed | Signed reconciliation, regression, interface and access results; decision group's go |
Wave 2 entry and exit criteria tighten from cycle to cycle.
Two rules make the cycles comparable. The same checks run every time: SAP describes reusing LTVF rules across test cycles, dress rehearsals and the production cutover, and the same idea applies to Contoso's saved report variants and regression rows. And the trend is reported, not just the latest result: open differences should fall sharply from W2 Mock 1 to W2 Mock 3, with source-data and mapping causes disappearing first. A flat curve two weeks before the dress rehearsal is a schedule risk, whatever the latest pass rate says.
Running the checks on every cycle
Manual checks drift between cycles; automated ones do not. With ACCELQ as the automation layer, the regression suite and API checks can be packaged as a CI Job and started from a pipeline or through ACCELQ's REST trigger when Luis's runbook marks the load complete, passing run properties such as the cycle name. ACCELQ's external API returns run results with pass, fail and not-run counts, which the evidence pack can store with the cycle. The reconciliation itself stays in SAP's reports and tools and with Dan.
The evidence pack
| Section | Contents | Signed by |
|---|---|---|
| Reconciliation | Nine-layer results for 1910 per cycle, untouched snapshots for 1710 and 1810, intercompany checks | Dan |
| Accepted differences | Each difference with cause, amount, decision and approver | Dan, or the owner of the object |
| Regression | Baseline and rerun reports per cycle, output comparison by company code | Maria (1710), Lucía (1910), Dan (finance outputs) |
| Interfaces | Connectivity, functional, negative, switch-off and volume results; partner confirmations | Theo with Lucía |
| Access | Role tests, Access Control risk analysis, emergency access log | Nadia |
| Volume | Job runtimes against windows at full volume | Sam |
| Cutover | Runbook with planned and actual timestamps, fallback rehearsal result | Luis |
| Decision | Go/no-go criteria, minutes and the decision group's decision | Grace's decision group |
The Wave 2 evidence pack: one section per stream, one signer per section.
Hypercare and its exit
Problems after a consolidation often surface weeks later, at the first period-end, the first payment cycle with a new bank format or the first quarter-end consolidation. EY frames carve-out reporting as Day 1, Month 1 and Quarter 1; use the same frame for hypercare. Contoso keeps Wave 2 hypercare open through the February month-end close and the first quarter-end close for 1910 in April 2027, which also falls before the TSA ends on April 30, so the legacy ERP is still there if a balance needs tracing. Exit criteria: two clean month-end closes for 1910, no open severity 1 or 2 defects, first quarter consolidated in Group Reporting without manual 1910 adjustments, CFDI rejection rate at its agreed level, and a tested path to legacy history after the TSA.
- Day 1 (Feb 8, 2027)First business dayOrders, deliveries and CFDI invoices for 1910; untouched check repeated for 1710 and 1810 on production
- Week 1First payment runMXN payment files accepted by the bank; bank statements matched
- Month 1February closeFirst 1910 month-end close in SAP S/4HANA; intercompany with 1710 reconciled
- Quarter 1Quarter-end close1910 consolidated in Group Reporting without manual adjustments; hypercare exit decision
- April 30, 2027TSA endsLegacy history reachable through the tested read-only path
What would you do?
At the W2 dress rehearsal, all reconciliations tie except open AP in 1910, which differs by MXN 3,214.60 on two suppliers. The team says the same difference appeared in W2 Mock 3 and "was accepted". The accepted-differences log has no entry. What should Priya do before the go/no-go?
- Plan W2 Mock 1 to 3 and a dress rehearsal with time to fix and retest between them.
- Tighten rules each cycle: what is waived in Mock 1 blocks in the dress rehearsal.
- Run the same versioned checks every cycle, triggered by the runbook, and report the trend.
- The evidence pack has one section and one signer per stream; hypercare runs through the first quarter-end close.
Check your understanding · not graded
Lesson check
1. In SAP BTC's cycle types, what happens to errors in a dress-rehearsal cycle?
2. What should the trend in open reconciliation differences look like from W2 Mock 1 to W2 Mock 3?
3. Why does Contoso keep hypercare open through the first quarter-end close?
4. A difference was accepted verbally in W2 Mock 3 but is not in the log. How should the team treat it at the dress rehearsal?