SAP S/4HANA Migration: Brownfield vs Greenfield vs Bluefield

The three strategic approaches organizations choose between when moving from ECC to S/4HANA.

Every ECC customer faces the same decision, and the three options trade the same two things against each other: how much change the business absorbs, and how much technical debt survives.

Greenfield is building a new house from scratch with a fresh floor plan; brownfield is renovating your existing house in place, keeping its foundation and layout but upgrading all the wiring and plumbing to modern standards; bluefield is moving into a new house but carefully bringing along only specific furniture and rooms worth keeping from the old one.

Key Concepts

Greenfield — build new. Start from SAP's best-practice configuration, migrate master data and selected open items, and re-implement the processes.

    keeps    nothing technical; the data you choose to bring
    loses    years of accumulated customisation, good and bad
    costs    the most effort and the most organisational change

The argument for it is that twenty years of configuration includes decisions nobody can now explain, and carrying them forward means carrying them forever. The argument against is that it is a re-implementation project, with the business change that implies.

Brownfield — convert in place. SUM with the Database Migration Option upgrades the existing system: configuration, custom code and history come across.

    keeps    everything, including the technical debt
    loses    the opportunity to simplify
    costs    the least business change; a technical project

This is the fastest route and the one most large customers take, because retraining thousands of users on redesigned processes is the expensive part, not the database.

Bluefield — selective. Carry forward some company codes, some configuration and some history, and rebuild the rest. Used when carving out a business unit, or merging several legacy systems into one target.

It needs specialist tooling and is the most complex of the three to run.

What actually decides it. Not preference — the state of the current system:

    heavily customised, processes need redesign   -> greenfield
    reasonable system, business wants continuity  -> brownfield
    carve-out, merger, or partial history needed  -> bluefield

Brownfield is not "no work". Custom code must run on S/4HANA, and much of it will not. The data model changed: MATNR is longer, BSEG and its totals tables are replaced by ACDOCA, customer and vendor masters became business partners.

MATNRBSEGACDOCA

Every custom program touching those is a remediation item, found with ATC checks running the S/4HANA readiness rules.

Run SAP Readiness Check first. It reports which simplification items apply to your system, how much custom code is affected, and which add-ons are compatible.

That converts the decision from an opinion into an estimate — which is the answer an interviewer is listening for. "It depends" is right, and it depends on measurable things you can name.

What the interviewer is probing.1. "Explain the three approaches." Probing: whether you can advise. Stalls: "Brownfield is easier." Moves up: brownfield converts the existing system keeping history and customising; greenfield is a new build with reimagined processes; bluefield moves selected data into a new system.

2. "When is greenfield the right answer?" Probing: judgement. Stalls: "Always." Moves up: when the existing processes and customisation are the problem — otherwise you pay for a reimplementation to arrive where you were.

3. "What is the biggest brownfield risk?" Probing: the practical answer. Stalls: "Downtime." Moves up: custom code and modifications against changed data models, which is why the readiness check and custom code analysis come before any date is promised.

4. "What changes in the data model you must account for?" Probing: the specifics. Stalls: "Nothing." Moves up: the universal journal replacing separate FI and CO tables, business partner replacing customer and vendor masters, and material number length.