Three routes to S/4HANA, and the evidence that should decide between them rather than a vendor preference.
A brownfield migration converts your existing SAP system in place, keeping configuration and history. A greenfield build starts from a clean system and reimplements the processes. Selective data transition sits between them, taking some entities across and rebuilding others. The right answer depends on whether your current configuration is an asset or a liability.
This decision is usually made on who is advising rather than on what the system contains. Implementation partners tend to prefer greenfield because it is cleaner work; internal teams tend to prefer brownfield because it preserves what they built. Both preferences are about the people, not the estate. The honest input is a custom code and configuration assessment.
How much there is, how much still runs, how much reads tables that no longer exist. This is the single largest input.
Whether the current design reflects how the business works now, or how it worked when the system was implemented.
Brownfield keeps it. Greenfield requires a deliberate decision about how much history to carry and in what form.
Greenfield is usually more expensive in change management and less in remediation. Brownfield inverts that.
Whichever route, the reasoning should be documented so it survives a change of sponsor.
Two manufacturers of similar size both run ECC and both need to move. They reach opposite conclusions, correctly.
The first has around 400 custom objects, most of them still in use, and fifteen years of configuration that genuinely reflects how it operates. Its processes are unusual for good commercial reasons. Rebuilding that on standard would mean re-deriving decisions nobody remembers making, and the business case for changing them does not exist. Brownfield is right: convert, remediate what the data model breaks, retire the dead code, keep the rest.
The second has around 1,100 custom objects, perhaps half of them dead, and a configuration shaped by two acquisitions that were never properly harmonised. Its finance team maintains three chart-of-accounts conventions. Carrying that forward means paying to remediate complexity it wants to be rid of. Greenfield is right, and the migration project becomes the vehicle for the harmonisation it has been deferring. The deciding evidence in both cases was the same inventory — which is why that inventory comes before the decision, not after it.
| Term | What it means |
|---|---|
| Brownfield | System conversion in place — configuration, history and custom code carried forward and remediated. |
| Greenfield | A new implementation on standard processes, with selected data migrated in. |
| Bluefield / hybrid | Selective data transition: a new system that takes chosen configuration and data from the old one. |
| Technical debt | Accumulated customisation and workaround. Brownfield inherits it; greenfield is the chance to leave it. |
| Downtime window | The outage required to convert. Often the practical constraint that decides the approach. |
The assessment is three to five weeks and determines a multi-year programme. Skipping it to save time is a false economy.
It is cheaper only if the existing configuration is sound. Converting a bad design costs more over the life of the system.
A clean build means new processes, retraining and resistance. The technical work is often the smaller half.
It offers the appeal of both, and the complexity of both. It needs the clearest scope of the three.
| Metric | What it tells you | What distorts it |
|---|---|---|
| Custom object count and usage | The core input. Objects that exist but are never executed should not drive the decision. | Counting without usage statistics, which overstates brownfield effort. |
| Configuration divergence from standard | How much of your setup is genuinely business-specific versus historical accident. | Treating all divergence as valuable because somebody once chose it. |
| Data volume and archivable share | What you would carry versus what you could leave. Often changes the downtime estimate materially. | Estimating before an archiving assessment. |
| Acceptable downtime | Frequently the binding constraint, and it should be established early rather than discovered. | Taking the IT answer rather than the business one. |
We build and support the systems this page describes. Start with a free working proof of concept — no cost, no commitment.