By Frank Schuler, VP SAP Technical Architecture at Syniti (formerly BackOffice Associates)
In six years, SAP's support for the Business Suite core application ERP will end, and customers will have transitioned to S/4HANA. But data migration can be complex – after all, it is not just a matter of system migration, but the existing data must also be successfully transformed to S/4HANA. And here lies the crux of the issue: according to a Gartner study, half of all data migration projects go beyond budget or schedule, with negative effects on business.
There are two ways to switch to S/4HANA. The first – the 'Greenfield transition' – is to implement S/4HANA from the ground up and then migrate the existing data into the new system. Alternatively, some companies prefer to transfer their existing SAP ERP components, including all customisations and data, to S/4HANA with the so-called Brownfield approach.
When deciding which path is right, software versions, integrations, modifications, and interfaces shouldn't be the only things considered. Of central importance is also the ERP data. On the one hand, because S/4HANA only supports Unicode, the data may need to be converted first. On the other hand, in the new system, only the relevant data should be processed in-memory, it, therefore, has to be decided what information will be stored as well as how and where.
Finally, S/4HANA brings new data quality requirements, because real-time processing eliminates any intermediate steps to cleansethe data: the data quality in the system itself must be right.
No matter which migration strategy a company chooses, an increased focus should always be on upfront data cleansing. This is because complete, accurate and clean data reduces the costs, complexity and risks of the change. Vice versa, if data does not exist and is not converted correctly or formatted after the migration, then the new SAP applications will not work as planned during go-live. And troubleshooting can then be difficult and time-consuming.
In many companies, data migration projects fail because their importance and complexity are not taken seriously enough. Ad hoc planning, limited project transparency, isolated decisions and poor communication are additional factors. Some companies also struggle with the technical side when they try to get data under control through self-programmed custom integrations.
However, data migration is much more than a purely technical task. As with so many other major projects, planning is the key to success.
The first step is to gain an overview of the existing data landscape and the quality of the existing data. Which data is central to the processes in the company and what standards must be met? Where are gaps, inconsistencies, errors?
After that, it's time to formulate clear and realistic quality goals that meet the needs of all business units and processes. What new requirements will S/4HANA place on the data? How can records be transformed and supplemented with missing information? And with which processes will the correctness of the data be authorised and validated in the future?
Once the framework requirements have been clarified, the migration project itself can be subdivided into several steps: data preparation, extraction, profiling (i.e. analysing and purging the source data), design with development of the necessary data schema for the target system and mapping between the source and target system. There is also the consideration of missing data records for new data fields, data transformation with simulation and validation of the loading process, and finally moving all the data into the production system. This structured approach has proven valuable in numerous projects and helps to keep all stakeholders – project managers, system integrators, users, the IT department and data managers – on the right track.
In addition, an iterative approach makes sense, firstly by breaking down the migration data into smaller, more manageable units. Loading simulations should also be considered because they help to detect problems early before the go-live. This form of testing is done by putting load on a system and measuring its response.
Many data cleansing and migration steps can be automated and can't be handled without the right professional tools. When selecting such tools, it's not just about technical integration with data sources and company systems; they should also provide functionalities that support project management, interdisciplinary collaboration and best practice methods. The more complex and dynamic the migration, the more important it is that the system allows a constant overview of the progress of each project section based on defined measures and goals.
But migration is just the beginning. Even after the system changeover, the integrity and relevance of the data must be maintained over the long term. Because digital processes generate tons of new information every day, the value of which can only be unlocked if the data quality is right. Intelligent data monitoring solutions serve as a centralised place for defining all the data rules in the enterprise and can automatically detect and correct data issues before problems arise.
Even with S/4HANA, the reliability of business decisions depends on the reliability of the underlying data.