SAP ECC's mainstream maintenance ends on December 31, 2027, and about half of ECC-era customers had not moved by mid-2026 [1]. The usual decision is to wait: start the analytics, planning and archiving work after the migration. That costs two years and often means building things twice. This guide shows how to decouple three kinds of work from the migration so they deliver now and survive the cutover: reading ECC data in place, delivering planning datasets from ECC and later from S/4HANA, and moving history into an archive that stays readable. It ends with a 12-point readiness check.
If you wait. Analytics, planning integration and archiving start after go-live. Go-live slips; the work starts later still. Two years pass in which the business runs on spreadsheets and the planners work around the data.
If you build on ECC only. Reports, loads and archives are written against ECC tables. At cutover, MATDOC replaces the material document tables, ACDOCA replaces the ledger tables, PRCD_ELEMENTS replaces KONV, and everything built on the old names is rebuilt.
The way out. Build once on a layer that reads both. Lima's catalog maps ECC and S/4HANA objects to the same business terms; pipelines and questions are written against the terms. When the cutover comes, the connector changes and the mapping updates; the pipelines, the reports and the archive stay.
The three kinds of work below each deliver value now and keep it afterwards.
The Data Fabric reads ECC through released interfaces alongside Oracle, the warehouse, the WMS and 3PL feeds, and answers questions in seconds with the access policy applied. A control tower with exceptions, KPIs and two governed agents reached production in 22 days on the showcase estate.
At cutover, the SAP connector points at S/4HANA and the catalog's mapping updates. The questions and the dashboards do not change.
Kinaxis, o9, Blue Yonder, SAP IBP and OMP need a delivered dataset on the planning calendar. Lima Backbone builds it from ECC today through table reads via a released function module, BAPIs and change pointers, and returns the plan to ECC. After cutover, the same pipelines read S/4HANA through released CDS views and table reads, with the mapping updated in the catalog.
The first planning dataset is live in four weeks at a fixed price. Because the ODP RFC path is closed from December 2026, an extraction built on released interfaces now also avoids a second conversion later.
Years of ECC history do not all need to move into S/4HANA. Lima extracts the tables in scope in partitions with a count and control total per table, lands them in a lakehouse or object storage in Iceberg or Delta tables, and keeps them readable through the Fabric for auditors and analysts. When the ECC system is retired, its history is still a question away, and any table can be restored into a target on demand.
A check that works on its own, whether or not Lima is involved.
Only if what is built reads ECC table names directly. Built against the catalog's business terms, the reports, loads and archives survive the cutover with a mapping update.
The planners are working around the data now, and the ODP path closes in December 2026 either way. A first dataset from ECC in four weeks, on released interfaces, carries over.
SAP's data services are sanctioned roads with their own licences; they land data in SAP's cloud first. The guide's point is to decouple the work from the migration; the tool choice follows from the estate.
A control tower in 22 days, or a first planning dataset from ECC in four weeks. Either way, built once.