How long, and what drives it
What you'll learn
Don't guess a calendar — size from drivers (live code, modifications, integrations), anchored to the ~12-month / 4M-LOC capstone, and express it as phases with exit criteria you re-estimate each sprint.
- There is no fixed duration — estimate from drivers, not a guessed calendar.
- Anchor: ~12 months for a 4M-LOC estate converting to S/4HANA 2023 (module 13); scale up or down from there.
- Drivers: live-code volume (SCMON), modifications, BDC/GUI dependencies, integration count, share already on released APIs.
There is no single answer to 'how long does Clean Core take.' The honest answer is a range driven by a few measurable factors, not a calendar pulled from thin air. The module 13 capstone anchors a realistic shape: a four-million-line custom codebase converting to S/4HANA 2023 is framed as a twelve-month programme. Smaller estates compress that; estates with heavy modifications and many integrations stretch it.
The drivers that actually move the estimate are concrete: the volume of *live* custom code (from SCMON, not the raw inventory — a large fraction is usually dead), the number of direct modifications and BDC/GUI dependencies, the count of point-to-point integrations that must be re-pointed at released interfaces, and the share of the estate already on released APIs. Estimate from these, and re-estimate every sprint as the burndown reveals the real fix rate.
Sizing is also sequencing. Because you scope by usage first (module 8's loop) and escalate ATC variants deliberately — DEFAULT, then readiness, then cloud — the timeline naturally falls into phases: a measurement-and-baseline phase, a high-value-remediation phase, and a steady-state gate phase. Communicating duration as phases with exit criteria is far more credible than a single end date.
Key points
- There is no fixed duration — estimate from drivers, not a guessed calendar.
- Anchor: ~12 months for a 4M-LOC estate converting to S/4HANA 2023 (module 13); scale up or down from there.
- Drivers: live-code volume (SCMON), modifications, BDC/GUI dependencies, integration count, share already on released APIs.
- Re-estimate each sprint from the actual burndown fix rate.
- Express duration as phases with exit criteria (measure/baseline → remediate → steady-state gate), not one end date.
Source notes: clean-core-curriculum (delivery synthesis)
Ask Claude
Build a prompt from this lesson + your question and open a fresh Claude chat with it pre-filled — handy for adapting a before/after pattern to your own object.