This is the eighth part in a 10-part series by Coalition Technologies on replatforming.
Table of Contents
Why Is Data Migration Necessary?
Data migration is the process of moving your valuable information from platform A to platform B, and there are several reasons why you might need to migrate or replicate your data. For example, your current database may find it difficult to store and scale your valuable information as data rapidly accumulates. If your current database proves to be incapable, then it may be necessary to move your data to another database that provides better scalability for your valuable information.
You can also choose to migrate your data from an on-premise site to a cloud computing environment. Be sure to find an efficient and secure method of migrating your data, if you are wanting to move to a public cloud, private cloud, multi-cloud, or hybrid cloud environment. You can transfer data from a local data center to the cloud via:
- Offline migration: A storage appliance is used to transfer data and then physically shipped to the cloud storage location.
- Online migration: Data is transferred over the Internet or a dedicated WAN connection.
Why Should You Begin with Data Migration?
Data migration is a phase within your replatforming project that will require significant effort since up to 40% of the project cost is typically allocated to data migration. By tackling data migration first, you can eliminate potential risks and ensure that your project is completed and tested by the time you go live. Although user experience (UX) optimizations and marketing enhancements are optional, it is essential that you have your data in line before you decide to go live.
Creating a Data Migration Plan for Your Business
In the data migration process, there are a variety of challenges as it is far more complicated than simply copying and pasting data. For a smooth migration, you should consider developing a data migration strategy that specifies all of the steps within the process. A well-prepared data migration plan decreases the possibility of unexpected costs, unplanned downtime, and customer dissatisfaction after the migration is complete. Be sure to reference your data migration strategy to avoid failure as it points out potential issues from the start of the data migration process.
Five Factors to Consider When Formulating a Data Migration Plan
When developing your migration plan, you should keep a couple of factors in mind for each step of the project.
1. Determine the Executors of the Data Migration Strategy.
Before starting the migration process, determine if the data platform that you are going to switch to has a reliable support service. In the event data migration does not run as smoothly as anticipated, it will be helpful to have a support service on standby to answer any questions and provide you with assistance as 83% of data migrations fail or go beyond their estimated completion date.
If the data migration process will be carried out by an external organization, your team should still monitor the project and someone intimately familiar with your company’s data should be in charge of the project from the start. Keep in mind that an external agency will likely need more time to migrate data than an internal team. B2B businesses will also need more detailed analyses since B2B businesses usually store a larger amount of data.
2. Discover the Data You Are Migrating
It is important to determine the type, volume, and complexity of the data as well as the scope of retiring the old system. You can prioritize scoping decisions based on the value your information brings to your organization. You will likely want to migrate data on customers, products, pricing, and price lists (or custom pricing).
3. Choose What Data to Migrate and What Data to Not Migrate
Migration projects run smoothly when you can effectively decide what to and what not to migrate. Here are few questions to keep in mind when cleaning your data:
- Does my data make sense? Are phone numbers correctly formatted?
- Are you holding onto outdated or incorrect information (i.e. mailing addresses, email addresses)?
- Are there any duplicates (i.e. usernames)?
According to Experian, about 30% of their customer data is believed to be inaccurate. Cleaning data leaves you with high-quality information that improves consumer analytics, increases revenue, and enhances email marketing and retargeting campaigns. Clean data can be stored in a separate location to reduce the chance of mixing your information with data that hasn’t been cleaned yet.
4. Phase and Test Migration
You can successfully mitigate risks depending on how carefully phased and organized your data migration process is. Consider running regular tests and starting with a single store or subset of data. For example, the 10-40-60 method is a testing mode where you:
- Copy 10% of your clean data from your backup server to your new platform. Since new data is usually cleaner, consider testing historical (the oldest) data first before working your way forward in time.
- Validate and test your work, then create a benchmark structure to see if your tests are good enough to proceed forward.
- Retest after you’ve met your success criteria, this time using 40% of your total data.
- Carry out validations and testing in order to meet your pre-determined success rate.
- Run the test again with 60% of your total data.
- Leave 60% of the data in your staging site after you’ve performed your 10-40-60 testing.
- Move the remaining 40% of the data. (You can evaluate and process the remaining 40% of the data before deploying it to minimize risk.)
5. Pre-Launch and Post-Launch Testing
Before you decide to go live, you should consider completing the data migration process at least 4-5 weeks beforehand. Keep in mind that this is an average timeframe that may change based on the size of your project. Having adequate time ensures that your data is in line and running smoothly.
- Internal Testing: Test and validate about 10% of your data.
- External Testing: Have customers visit and test your website.
Continue testing until you reach your benchmarks and repeat the testing cycle whenever updates are made.
In the next part of our Replatforming Series, read about how you can integrate the data after the data migration process.
