Redesign vs lift-and-shift

The most common and expensive mistake in APO replacements is to rebuild APO inside IBP: the same planning books as views, the same macros as calculations, the same overnight batch chain as jobs. It feels safe because planners see familiar numbers, but it carries over the complexity and workarounds of an on-premise system built over fifteen years into a cloud product designed differently. It also doubles test scope, because every rebuilt macro needs its own proof.

Anti-patterns to recognize

Classifying APO content

Contoso classified each APO object into four buckets: retain (same business rule, rebuilt in the successor), redesign (the need remains, the way changes), replace (an IBP or S/4HANA standard function covers it) and retire (nobody uses it). Of 140 DP macros, 22 were retained, 31 redesigned, 46 replaced by standard key figures and 41 retired (illustrative course data). The retained and redesigned 53 got test cases; the replaced ones were covered by the standard key figure tests; the retired ones got a sign-off from Sunita that they were not needed.

Workshops that work

1. Start each workshop with the decision, not the screen: "Which decision does Jonas take on Tuesday mornings, with which data?"
2. Show SAP's standard flow in IBP with Contoso data loaded, before discussing APO.
3. List gaps against the decision, not against APO.
4. For each gap, decide retain, redesign, replace or retire, and name the owner.
5. Write the acceptance test for the decision while the planners are in the room.

How redesign changes the test plan

Lift-and-shift produces a test plan that checks the new system against the old one, object by object. Redesign produces a test plan built from planner decisions and their outcomes. The difference shows in the number and kind of test cases. Contoso's first draft, written by object, had 412 test cases, most of them "macro X gives the same result as in APO". The redesigned plan had 96: one per planner decision, one per retained or redesigned rule, the standard key figure tests, the integration tests and the parallel-run comparisons.

The last row matters most for quality. After APO is switched off there is no old system to compare with, but IBP keeps changing with every release. A decision-based test set becomes the regression suite Contoso runs before each IBP release reaches production.

Signs a project is sliding into lift-and-shift

• The design documents are organized by APO planning book, not by planning process.
• Change requests ask for "the same as APO" without naming the decision behind it.
• The custom key figure count rises every sprint while the sample planning area's standard key figures go unused.
• Planners are not in the design workshops; APO support staff speak for them.
• Test cases compare objects with APO instead of outcomes with agreed criteria.

Priya reports these signs to Grace's steering group monthly as a simple count (custom key figures, open "same as APO" requests, workshops without planners). When the count rises, the design authority reviews the latest requests before build.