This is how data migration works!

The introduction of the new enterprise software is underway, the processes are defined and configured. The next important step is the migration of the master data. For most providers, this is an uncomplicated process from a technical point of view - and yet the transfer of the data often causes stress. We explain where the typical problems are and how you can avoid them.

successful team

The smooth import of master data is an important factor for a successful implementation so that you can continue to work seamlessly with the new business software. This is where a well thought-out migration concept helps, which defines exactly which data should be migrated when, how and where.

Contents of a migration concept

  • Technical and organisational requirements
  • Data selection ("what must I do?")
  • Quality / timeliness requirements
  • Specifications for sample data record and mapping
  • Roadmap for the migration process
  • Data archiving of the legacy system

 

Dilemma data quality

Customer data, supplier data, price lists, item details: These unpopular tables and databases conceal a high intangible value. Master data is the basis of day-to-day business and the basis for efficient sales, marketing and controlling. Nevertheless, the quality of customer master data in around 50 percent of all small and medium-sized companies in Germany is low or even very low. Typing errors, duplicate or outdated data records or number shifts lead to delays, errors and poor service. According to studies, the economic damage is 8 to 12 percent of the operating profit!

Example: Joseph Mair, Joseph Maier, Joe Mair - three entries, but the same customer. Which leads, among other things, to Mr Mair receiving each newsletter three times. His contact history is not completely traceable, a change in his telephone number was only corrected in one of the three data records and the price list assigned to him is no longer up-to-date in two data records. This does not only cost time, money, reputation and trust.

The situation is made more difficult by the fact that as a result of digitization, ever larger amounts of data have to be processed - quickly and as error-free as possible. Business software helps you with automated data maintenance, but does not heal any omissions or legacy problems in data management. The introduction of a new enterprise software is therefore the perfect time for an inventory and (if not yet available) the introduction of a professional master data management.

 

Data migration process

Companies, their databases and processes are different. There is therefore no blueprint for the perfect flow of a data migration, but we present you with an exemplary workflow that applies to the majority of our implementation projects.

1. Data analysis and data cleansing

Another great opportunity for new business software is the optimization of your data stock - both quantitatively and qualitatively. Throw off ballast! Usually you export your master data to Excel for this process in the first step and check here which data records you really need in the new system and whether each data record is correct and up-to-date. Pay attention to duplicates, incorrect spellings and correct links between master data such as company / contact person or company / price list.

2. Mapping

While the cleansing of the data was "only" meticulous diligence, the mapping between old and new data structures usually requires technical know-how. Each field of the existing data set must be assigned to the counterpart of the new data structure.

Since the structure of the data records is often changed in the course of the redefinition of many processes, a 1:1 assignment is often not possible during mapping, but existing data records must be split or merged. If necessary, use the experience of your software provider! It is also important that the data types of the mapped fields match ("continuous text", "numeric", "time", "date" etc.).

3. Test and import

Once the cleanup and mapping are complete, the master data can be imported into the new system. The import is usually run on a test system with a small amount of data records and used for training purposes at the same time. If everything goes smoothly, the entire data set can be transferred to the live system.

A frequent question from our customers when planning the migration concept is: "Can the data be imported automatically or do we have to enter everything manually? As already mentioned in our last article on implementation: The export-import process of master data via CSV files is a standard that is available in almost every business software. But we do not always recommend automatic import. Especially with smaller data sets, it can be more useful to import the data records manually, for example if the mapping would be very complex due to a fundamentally new data structure.

 

And what happens after the migration?

Now you have checked your data records for the migration in a hurry, removed all legacy data, identified and corrected errors - project completed? No, because the uncontrolled growth starts all over again and the next data chaos is pre-programmed. We therefore recommend continuous master data management in order to keep your data at a high level. This often means that the data maintenance processes have to be fundamentally changed. Your company software supports you, for example, with intuitive masks and standardized workflows for error-free data entry ("first time right").

Ideally, there are also fixed rules and responsibilities for dealing with your data treasure and a company-wide awareness of the value of your data. Concrete and verifiable key figures are helpful (e.g. "maximum 5 percent mail returns") - only what is measurable can be controlled and thus improved in the long term.

Only the strategic interaction of people, processes and technologies ensures data quality in the long term. We will be happy to advise you on how to establish a professional master data management system.

You will also find a brief visualization of the implementation process in our infographics: