How to Perform a Successful Data Migration for Your Life Sciences Company

Data migrations are challenging and complex – and often an afterthought when implementing new systems. In this article, we cover the building blocks of a successful data migration for life sciences companies.

Companies are continuing to adapt their strategies – and technology resources – to the current environment. As companies pivot and move to new systems, their data will have to do the same. But in the life sciences industry, a successful data migration is not as simple as it sounds.

Life sciences companies need to first understand what this (often underestimated) project means—and why it’s much more than just an unavoidable and tedious task.

What is a Data Migration?

In broad terms, a data migration is the process of moving data from one location, format or application to another.

In the life sciences industry, data migrations occur as a result of these three situations:

• Merger, acquisition or divesture activity
• Initiatives to move to the cloud
• Replacing outdated, legacy technology

In fact, all of these situations are often at play for many data migration projects.

Let’s take a look at two real-life examples.

#1. Implementing a New Cloud-based System After Multiple Acquisitions

A global pharmaceutical company grew through multiple acquisitions and ended up managing multiple repositories and TMF solutions. To consolidate systems, replace outdated technology and move to the cloud, the company decided to implement the Veeva Vault Clinical eTMF. This implementation included a complex data migration to extract, transform and move data from multiple systems into one new system.

#2. Sunsetting Two Legacy Systems 

Another global pharmaceutical company realized it was hosting study startup data in one place and TMF documentation in another. This lack of centralized document control highlighted the need to migrate content from two legacy systems into one.

What do these two examples have in common? They were both a successful data migration. This is because they understood the importance of data migration in their greater software implementation project—and they rigorously planned for it.

But what happens when migration efforts go wrong?

At best, poorly planned migration efforts result in implementation delays and cost overruns. At worst, bad data compromises system integrity and corporate or regulatory compliance.

Here’s how to prevent that.

Steps to Successfully Perform a Data Migration

A migration plan should follow a process referred to as “ETL,” which stands for Extract, Transform and Load.

Following are the ETL steps for executing a data migration:

Step 1: Extract

Extract content and metadata from the source system or systems.

Step 2: Transform

Analyze the extracted content and metadata to identify how to match the source to the target system. This step involves “mapping” old values to new ones so the information can be transformed to the target system object model. This is the most difficult and time-consuming part of a migration as both technical and business leads work together to visualize the end results.

Step 3: Load

Import content and transformed metadata into the target system. Depending on the complexity of the source objects, the load may be done in a series of operations. For example, the first operation will load the documents, the second will load document renditions, the last will load document audit trails.

Step 4: Post Load Processing

Verify the accuracy of the migration through a number of custom queries run from testing scripts to confirm the transformations were completed accurately.

Migrating to Veeva Vault?

Try our free query tool to assist in migration verification activities.

Planning a Data Migration

Before performing the technical pieces of a migration, there are specific ways to plan it to ensure success. First, you must address three fundamental points:

1. Identify core data requirements and scope of migration
2. Identify realistic activities, schedules, assumptions and risks
3. Define how to make data work in the new system and enterprise data architecture

This is where most organizations fall short. They don’t take the time to establish these expectations and jump right into the technical approach. This lack of understanding leads to longer project times, frustrations and higher costs.

That’s why we recommend following a five-step process that wraps your migration efforts into the broader scope of your new system implementation project.

Step 1: Plan

This phase coincides with the Extract step as the source data is key in developing the migration plan.

Get with your team to address key objectives, approaches, roles and responsibilities and other deliverables.

Will all content be migrated from the source system(s) to the target system, or just a subset? Will the source system(s) be decommissioned completely?

Are you following a big bang or rolling migration strategy? Who is responsible for what? How are you testing?

Step 2: Analyze

This phase coincides with the Transformation step.

Consider what data truly needs to be migrated versus archived. Where can the use of default values be leveraged?

Analyze source system data and identify mapping rules upfront, so you don’t have to spend as much time cleansing and enriching data.

Step 3: Prototype

The Prototype is the dry run of the Load step.

Run a series of migration prototype sessions into a target system development environment, so you can visualize your data in the context of the new application. This will allow you to update migration requirements and mapping rules before formal testing. In more complex migration projects, there will be multiple re-runs of the Prototype.

Step 4: Test 

We recommend a three-prong approach to migration testing. This includes:

• Count-based testing
• Functional testing with migrated content
• Manual business verification

Step 5: Deploy

As discussed, migrations typically happen in conjunction with a new system deployment or upgrade. So, the execution of the migration must be a part of your overall system deployment activities.

By putting in the effort upfront to better understand your data, approach, and how migration fits into the overall scope of a new system implementation, you will be on a better path to a successful data migration.

Conclusion 

Data migration projects are complex, but success can be achieved with the right preparation. In the tightly regulated life sciences industry, successful migrations are crucial and require a high level of planning to ensure integrity and compliance.

Daelight Solutions has decades of experience performing validated migrations of all shapes and sizes, with an emphasis on planning and migration analysis to minimize tedious business-led enrichment activities. Contact us to let us know how we can help you.