Lesson 4 of 5
Dependencies and risks
11 min
Listen to this lesson
Most schedule slips in a conversion start as a dependency nobody owned: a stream waiting for another stream's output without a date, or two streams needing the same people. The integrated plan makes these visible; managing them is weekly work. This lesson covers the dependencies that matter most in a brownfield program and how to manage them with risks and leading indicators.
The dependencies that bite
- Business partner synchronization (CVI)Must finish in ECC before conversion; feeds finance and sales testing
- Add-on and business function compatibilityMaintenance Planner blocks the plan until resolved
- SI-Check consistencyReturn code 8 and 12 errors must be fixed before SUM's last check
- Custom code transportsMust be complete before CMQ, or testing finds known findings
- Finance preparation and CustomizingNeeded before reconciliation can be trusted in CMQ and mocks
- Roles and Fiori contentNeeded before testers can work in their real roles
- Interfaces and partnersPartner test windows needed for integration testing
Some dependencies are external. The RISE contract sets when SAP provides the target environment and which technical steps SAP performs, so SAP becomes a workstream partner with dates in the plan. Add-on vendors must release conversion-compatible versions. Integration partners such as banks and carriers need test slots. Each external dependency needs an internal owner who chases it and a fallback if it slips.
| External date | What it means for the plan |
|---|---|
| SAP ECC 6.0: mainstream maintenance ends 2027, extended maintenance 2030 | The conversion must land with room for hypercare and follow-on waves well before support pressure rises |
| SAP Solution Manager 7.2: mainstream maintenance ends end of 2027 | ChaRM on ED1 to EP1 has a horizon; the run organization moves to SAP Cloud ALM after go-live |
| SAP Process Orchestration 7.5: mainstream maintenance ends end of 2027, extended to end of 2030 | Interfaces leave PO after stabilization, in planned integration waves |
| SAP ERP, private edition, transition option (subscriptions to the end of 2035 under SAP Note 3591251) | An alternative the steering group may ask about; it requires SAP ERP, private edition on SAP HANA before the end of 2030 and is not Contoso’s path |
Product dates that shape a brownfield plan. Verify current dates and conditions with SAP before quoting them to a board.
Don’t stack transformations. Contoso’s about 60 interfaces run on SAP PO 7.5 (XP1). Theo’s team keeps them on PO through the conversion weekend and moves them to SAP Integration Suite after hypercare in integration waves: a pilot of four low-risk interfaces at about three months, procure to pay (including the Lakeside order flow) at four to five months, and order to cash (including Great Lakes Fleet Services) at six to seven months. Changing middleware on the same weekend as the conversion would double the number of things that can fail and blur the cause of every defect.
A dependency register
| Dependency | From → to | Needed by | Status | Fallback if late |
|---|---|---|---|---|
| CVI synchronization for 400 one-off customers | Data team → finance and sales testing | Before CMQ conversion | At risk: 60 failures on postal codes | Archive inactive one-off customers instead of converting them |
| Tax add-on compatible release | Vendor → Maintenance Planner | Before CMD plan | Committed date received | Uninstall with a replacement process |
| Readiness ATC clean in scope | Kenji → Priya | G2 | On track: burn-down 82% | Reduce scope with owners' agreement |
| Business roles for test users | Security → Priya | First CMQ test day | Late by one week | Test with broad roles first; repeat role tests later |
| Bank test environment | Bank → integration lead | CMQ integration cycle | Not yet scheduled | Escalate through Dan's treasury contact |
Extract from Contoso's dependency register. Every row has both owners and a fallback.
Shared people are dependencies too
The scarcest resource in a conversion is rarely developers. It is the people who know the business: Dan's controllers, Maria's key users, Ana's AP team. They prepare data, test, reconcile and sign off, while running the business. Plot their load across the cycles and the business calendar. Contoso moved its first CMQ regression cycle out of the year-end close weeks after seeing that the same six people were booked twice.
Risks and leading indicators
A risk register lists what might happen; leading indicators show whether it is starting to. For each major risk, pick a measure that moves before the risk lands, and review it weekly. Lagging measures such as missed milestones tell you too late.
| Risk | Leading indicator | Trigger for action |
|---|---|---|
| Custom code not ready for CMQ | Readiness ATC burn-down against plan | Two weeks behind the planned curve |
| Finance cannot reconcile | Open differences per mock and their causes | Any unexplained difference after Mock 1 |
| Downtime does not fit | Measured duration and contingency per rehearsal | Contingency below the agreed target after Mock 2 |
| Testing quality | Pass rate, defect arrival and closure rates | Arrivals above closures for two weeks |
| Data blocks the conversion | SI-Check errors and CVI failures over time | Any return code 8 or 12 error four weeks before a conversion |
| Business overload | Key user hours booked versus available | Above 100% in any week |
What would you do?
Six weeks before the CMQ conversion, CVI synchronization still fails for 60 of Contoso's roughly 400 one-off aftermarket customers, all with missing postal codes. Alongside its other cleansing work, the data team says the fixes will take five weeks; the sales team says most of these customers have not bought in years. What should Grace's dependency review decide?
- Most slips start as unowned hand-offs; give every dependency two owners, a date and a fallback.
- External parties (SAP under RISE, add-on vendors, banks) are dependencies in the plan.
- Key business people are the scarcest resource; plot their load against cycles and calendar.
- Manage risks with leading indicators and agreed triggers, reviewed weekly.
Check your understanding · not graded
Lesson check
1. What should every row in a dependency register include?
2. Which is a leading indicator for custom code readiness?
3. Why are key business users often the critical resource?
4. How did Grace protect the conversion from Wave 2 pressure?
5. Why does Contoso keep its interfaces on SAP PO 7.5 through go-live and move them to SAP Integration Suite afterwards?