Now accepting guest editorial from SAP, Oracle, Workday and ServiceNow practitioners.Pitch a guide
The Sign-Off Table
All guides

Building an SAP Testing CoE · Part 3

Testing an ECC to S/4HANA migration: a phase-by-phase plan

Greenfield, brownfield or selective data transition: what each approach means for testing, where test activities fall across SAP Activate and what S/4HANA changes you must test for.

Matt Angerer
October 8, 2026 · 13 min read

Every S/4HANA migration is sold on its business case and judged on its go-live. Between the two sits the longest and most underestimated workstream in the program: testing. This guide lays out what to test, when, and why the answer depends on the migration approach you choose.

Your approach decides your test scope

Brownfield (system conversion)

You convert your existing ECC system in place, keeping configuration, custom code and history. Testing focuses on proving nothing broke: the same processes, run on the same data, should produce the same results, except where S/4HANA deliberately changes them. A regression baseline captured on ECC before the conversion is the most valuable test asset you can have.

Greenfield (new implementation)

You implement S/4HANA fresh, usually closer to standard, and migrate open items and selected master data. Testing looks like a new implementation: prove the new processes work and the migrated data is complete and correct. Your ECC regression suite is less useful, but your process library is not.

Selective data transition

Somewhere in between: a new or shell system with selected configuration, data and history moved across, often to consolidate or split company codes. Testing has to do both jobs, and data reconciliation becomes a first-class test activity.

Where testing fits in SAP Activate

SAP Activate is the methodology most S/4HANA programs run on. Testing is not a phase in it, but each phase has test work that, if skipped, comes back later at a higher price.

  1. Phase 1

    Discover

    • Readiness Check
    • Test tool and scope assessment
  2. Phase 2

    Prepare

    • Test strategy signed off
    • Baseline regression captured on ECC
  3. Phase 3

    Explore

    • Fit-to-standard
    • Test scenarios per process
  4. Phase 4

    Realize

    • Unit and string tests
    • SIT cycles 1–2
    • Data load validation
    • Mock conversions
  5. Phase 5

    Deploy

    • UAT
    • Performance and roles
    • Dress rehearsal
    • Cutover
  6. Phase 6

    Run

    • Hypercare
    • Regression on every release
Test activities across the six SAP Activate phases. The earlier phases are where most programs under-invest.
  • Discover and Prepare. Run the SAP Readiness Check to see which simplification items, add-ons and custom code will be affected. Sign off the test strategy. On a brownfield, capture your ECC regression baseline now, while ECC is still the system everyone knows.
  • Explore. Fit-to-standard workshops decide your processes. Write test scenarios as the decisions are made, not months later from memory.
  • Realize. Unit and string tests as configuration lands, then system integration test (SIT) cycles, usually two, across end-to-end processes and interfaces. Mock conversions or data loads start here and each one gets reconciled.
  • Deploy. User acceptance testing, performance testing, role and authorization testing, the dress rehearsal and the cutover itself.
  • Run. Hypercare, then regression testing on every support pack, Feature Pack Stack, upgrade or release from then on.

What S/4HANA changes that you must test

SAP publishes a Simplification List describing what changes between ECC and S/4HANA. Your Readiness Check tells you which items apply to you. A few affect almost everyone and deserve dedicated test scenarios:

  • Business Partner. Customers and vendors become Business Partners through Customer-Vendor Integration (CVI). Every process that creates, changes or reads a customer or vendor needs retesting, and the conversion itself needs reconciling.
  • The Universal Journal. Finance and controlling postings land in one table (ACDOCA). Reports, reconciliations and any custom code that read the old FI/CO tables need validating.
  • Material ledger and inventory valuation. The material ledger is mandatory in S/4HANA. Test valuation, period-end and costing closely.
  • Credit management. Classic SD credit management is replaced by SAP Credit Management (FIN-FSCM-CR). Order blocks and releases behave differently.
  • Output management. Forms and outputs may move to the new output management framework. Every invoice, delivery note and purchase order printout is a test case.
  • Custom code. Use the ABAP Test Cockpit with the S/4HANA checks to find affected custom code, then test the business processes that call it.
  • Fiori. Many users move from SAP GUI transactions to Fiori apps. Test both the apps and the business roles and catalogs that expose them.

Cutover: rehearse until it is boring

A cutover plan is itself something to test. Plan at least two or three mock cutovers and a full dress rehearsal, each with a specific goal and its own exit criteria.

Mock 1

Prove it runs

Mock 2

Prove it reconciles

Mock 3

Prove the timing

Dress rehearsal

Prove it on production hardware

Go-live

Sign off

A typical mock cutover ladder. Each rehearsal proves something new; go-live should feel like a repeat.

For every mock, reconcile financial balances, open items, inventory quantities and values, and master data counts between source and target, and record how long each step took. By the dress rehearsal the timings should be predictable and the reconciliation boring.

Go-live readiness evidence

  • All critical end-to-end processes passed in the final SIT and UAT cycles
  • No open severity 1 or 2 defects without an approved workaround
  • Final mock or dress rehearsal reconciled within agreed tolerances
  • Cutover timings fit inside the agreed downtime window
  • Performance tested at peak volumes (month-end close, peak order days)
  • Business roles tested with real user role assignments, not SAP_ALL
  • Regression suite ready for hypercare fixes and the first post-go-live release

Key takeaways

  • Brownfield testing proves nothing broke; greenfield proves the new design works; selective transition does both.
  • Capture an ECC regression baseline before conversion. It is the cheapest test asset you will ever build.
  • Build dedicated scenarios for Business Partner, the Universal Journal, credit management and output.
  • Interfaces are the most common go-live failure point. Test them in every SIT and every mock.

Next: what to automate first for S/4HANA regression.