Data Migration is not Copy / Paste
Data migration is the simplest part of a project and easily executed. At least, that is what someone about to fail an ERP implementation would say.
Sorry to the seasoned professionals whose hearts just skipped a beat. You take the existing data in your legacy systems, extract it, then upload it into your new system. The whole thing should take a week, max.
Until you have run an ERP project, or at least read through the hundreds of lines a project schedule dedicates specifically to data migration, you will not grasp how complex it actually is. Most people do not actually fit my hyperbolic opening example, but they still underestimate it.
This is a real problem, because data migration takes large amounts of time and input from the business stakeholders who own each data domain. If they do not understand the complexity, you will struggle to get the time commitment from them. You need that commitment to succeed. Your data migration manager, if you have one, already knows this. But that does not mean you are off the hook. Driving stakeholder engagement is a multi-person job, so arm yourself with the knowledge to help when it is time.
Not a Lift and Shift
This is not downloading from your legacy system and uploading to the new one.
For each in-scope data domain, and you will have many, there is a meticulous list of tasks: pulling the data, cleaning it, normalizing it, then mapping and converting it to a new structure that likely looks nothing like your current one.
No One Knows Both Systems
Heavy coordination across multiple groups is needed just to execute the steps.
Data migration requires deep knowledge of both the current system's data and the new system's structure and needs. Because IT work tends toward specialization, it is unlikely one resource knows both systems well.
Decisions, Not Just Extraction
You can't, and probably shouldn't, bring it all.
Beyond the functional needs the data must support, your stakeholders have to decide what data even comes over. How many years back do you go? What's no longer relevant? Are there regulations dictating how certain data must be kept? None of these decisions happen quickly.
There is Just So Much Data
There are more domains than you think.
You are probably thinking of the obvious ones: customers, vendors, products. But there are more data domains than you think, often with multiple sources in different formats that need to be unified. On top of that, your new system also needs reference and configuration data just to function.
Another challenge related to data migration is ownership. As discussed above, your business stakeholders need to be highly involved to determine how the data should look in the new system. That is required. There is no realistic scenario where a vendor can do that for you, so be ready for that.
However, you need to pay close attention early on in the proposal and contract phase, because more and more, implementation vendors are trying to avoid data migration activities at all costs. What this means for you is that if you don't pay attention, you may sign a contract that notes you as responsible for nearly all data migration activities. Then, at the end, once all the transformations are done, the hard work, they will be there to load it into the system.
Some businesses may be okay with this. But that doesn't mean you will be. Odds are, even if you are a smaller business, your data migration needs will be complex. If you find yourself owning the data migration, you must strongly consider bringing in talent. There are several ways to go about this, like hiring or contracting a skilled data migration manager. You can also hire a second vendor that specializes in data migration. Many companies do both. They may bring on a data migration manager to oversee the overall workstream, who then manages all resources involved, including the second vendor that specializes in data migration.
It may not be obvious to those not familiar with ERP projects, but data migration is one of the riskiest parts. It can make or break projects. The system will not function with improper data. If the workstream is riddled with issues, project timelines and budgets break.
Don't get caught off guard. Advocate for yourself. Make sure you know what you are getting into, then make the necessary adjustments to ensure that whichever path you take, you are as derisked as possible.
