What is data migration?

Data migration: definition and reasons

A data migration moves data from a source system to a target system. The reason is usually a system change: replacing legacy software, moving to the cloud, switching the store platform, merging two companies’ IT landscapes or moving from Excel to a specialized application. Data model, field names and formats often change in the process.

Technically, this is often called ETL: extract, transform, load. The data is read from the old system, converted into the structure of the new system and loaded there. The real difficulty rarely lies in the copying but in assigning the data correctly in business terms.

How a migration works

It starts with analysis: which data exists, which is still needed, and in what quality? Next come the mapping, that is, assigning each source field to a target field, and the cleanup of duplicates, gaps and format errors. After that, one or more test migrations run, and the business department reviews the result.

At the switchover, the so-called cutover, the old system is frozen, the final data state is transferred and checked by sampling and by comparing totals. For large systems, migration is also done step by step, so that not everything has to happen on one day and a way back remains possible. It is important to decide in advance what result counts as a successful switchover.

• Analyze data volume and data quality in advance
• Deliberately sort out data that is no longer needed
• Document the field mapping in writing
• At least one full test migration, reviewed by the business department
• Check completeness by comparing totals and samples
• Keep a fallback plan and a backed-up old system ready

Real-world example: a retailer changes its inventory system

A specialist retailer replaces its old inventory management with a modern system connected to the store. In the old system, customers appear several times in slightly different spellings, item numbers have been assigned inconsistently over the years, and open orders must continue seamlessly after the change. In addition, there are former customers whose data is only retained but no longer actively needed.

Before the migration, duplicates are merged, item masters cleaned up, and master data and open transactions treated separately. In a test migration, accounting checks customer accounts and balances. Only then is the switchover carried out over a weekend, and on Monday the team works in the new system.

Mistakes, data protection and retention

Common mistakes are underestimated data quality, unclear responsibilities, missing test runs and trying to take everything along, including useless old data. Personal data belongs in the new system only if there is still a purpose or a retention obligation for it; deletion periods should be clarified before the migration.

Tax-relevant records must remain readable for the statutory periods, so the old system or an archive must not be shut down too early. Nano Concepts plans migrations with mapping, test runs and an acceptance protocol and supports the switchover through to stable operation. After the change, a phase of close monitoring makes sense, in which anomalies are fixed quickly.