How to mess up a data migration project
System renewals, deploying new systems to use, and changing from one system to another are often challenging and long-lasting projects in organisations. The more functionalities, uses, and users the system has in the organisation and possibly outside it, the more critical it is that the transition is smooth and successful. An unsuccessful system transition often leads to lots of manual work, expenses, and frustration. For example, HR and payroll system renewals need to be planned and executed with care to avoid undesirable outcomes.
A sure recipe for a disastrous data migration has the following ingredients: A cup of not understanding the data in question, a teaspoon of unrealistic expectations and a pinch of problems in communication.
Not understanding the data is the biggest mistake you can make
The most important part of a data migration is, unsurprisingly, the data to be migrated from one system to another. It is crucial that data is moved correctly and stays intact during the migration. If there are discrepancies between the data in the old system and in the new system, a disaster is bound to happen. The best case scenario is that an enormous amount of manual work is required to correct the data and to compare data between the old and new systems. The worst case scenario is that investigating errors and correcting the data is practically impossible, especially if the old system is no longer in use and original data is not available. To ensure that data is migrated correctly, the data and its limitations must be understood prior to the migration.
A common challenge regarding data is that it is rarely possible to move the data in the exact same format from the old system to the new system. Salary data is a good example of this: For example, teachers may have multiple different salary supplements on top of their base salary, due to experience or extra duties such as recess supervision. It is very likely that such salary supplements have different codes in the old and in the new system. It is also possible that some supplements have been combined in the old system but not in the new system and vice versa. As the example shows, data migrations sometimes require very complex data handling to ensure that data is correctly transferred. To avoid problems with data it is necessary to understand the original data and the format it should be transferred to in the new system.
Unrealistic expectations make for a complicated and expensive project
A sure way to fail in a data migration project is to have unrealistic expectations. To succeed, an organisation must know the related processes and be able to map them clearly, have the capability of renewing and adapting, be able to develop solutions together with multiple different stakeholders, and to be able to make quick decisions. Having unrealistic expectations about what is required for a successful data migration often leads to organisations neglecting to plan adequately and to ensure that the requirements for a successful data migration are met. Challenging data migrations might entail that the old system has no usable APIs, but data must be collected through the user interface. Automation and robotic process automation can be used in cases like this, however, it is important not to have to unrealistic expectations of the data migration if the starting point is challenging.
Problems in communication add fuel to the fire
Lastly, problems in communication are a sure way to fail a data migration. To avoid a data migration disaster, organisations need to ensure that all stakeholders have a clear understanding of the old and new system, of the original data, and of the limitations of the chosen migration method. To ensure this, successful communication is vital. This way it’s also easier to manage stakeholder expectations.
Of course, a million other things can go wrong as well. System renewals and data migrations are almost always large and challenging projects. To avoid costly disasters, keep the points I covered in mind: understand your data, have realistic expectations, and make sure you have good communication practices in place.
