Once an enterprise has decided to implement Salesforce, the next question that most people have involves how to integrate their legacy data. It would be nice if Salesforce possessed psychic abilities and data migration could be “plug and play,” but no platform offers that capability — yet. Instead, you will need to develop a plan to prepare your data for migration to ensure a smooth transition. The attention you devote to planning your data migration can help ensure the integrity of your data, giving you better results after migration. Although the specifics of a data migration plan can vary by business or industry, there are certain steps that are universal.
Review Your Legacy Data
Before you can plan how to migrate your data, you need to know what you have. Determine what files need to be migrated and map their location. Are there files that can be archived on-premises, such as records that are not used but must be retained as support for tax filings? Are there siloed databases that need to be fully integrated? Do some employees have files on their PCs rather than the server? How clean is the data? It is also critical to know how much data you have that must be migrated.
Plan Your Data Extraction
The first phase of your data migration plan should address the extraction, or getting your data out of your legacy system. What tools, such as an ETL app, will you need for extraction? What is your timeline for extraction, and who will be the person responsible for overseeing the process? Is there a specific order in which data should be extracted? What steps will be taken to verify the data? What is your recovery plan should the extraction fail? Where will the data be held, if necessary, between extraction and loading, and do you need to implement any additional security measures during this stage?
Plan Your Data Transformation
Although it should be obvious, too many times, legacy data is not backed up and a recovery plan drafted before the data transformation stage. Make sure that you can recover from a failed transformation should the worst happen. Typically, the transformation phase is where your data will be cleaned and standardized. There are a number of excellent tools available at the Salesforce AppExchange that can help you with these tasks and save you considerable time over performing them manually. Even if you have already scrubbed your data prior to embarking on your migration, it is a good idea to clean it one more time before you load it.
Load Data into Salesforce
You will need to determine whether there are specific files, such as customer masters, that need to be loaded first and handle them accordingly. You cannot load legacy data without the use of a data loader, but you can find a variety of apps — some of them free — at the Salesforce AppExchange. Although some customers choose to load all their data in one massive step, it is better to break your data loads into smaller pieces. This lets you validate the data incrementally. It is far easier to fix any problems as you go than to try to fix your data after it has all been loaded. Set up a plan for loading your data that includes the order of transfer, periodic validations and time frame for completion.
Getting rid of duplicate records is an important part of scrubbing your data. You may have already de-duped your files manually or performed the task as part of your transformation process. If not, now is a good time to do it. Some of the ETL apps in the Salesforce AppExchange will handle the de-duping automatically. However, it is always a good idea to verify that the duplicates have indeed been eliminated before you go live.
Enrich Your Data
In a previous post, the ways you can enrich your data were discussed at length. In brief, enrichment involves populating missing data or enhancing data that is of poor quality. When you enrich your data, you are improving the accuracy of your analytics and enhancing the usefulness of your reports. Visit the Salesforce AppExchange to find helpful tools for filling in the gaps in your data.
Data migration may never top your list of your favorite ways to relax, but it does not have to be your least favorite. Prior, detailed planning can do much to ensure a successful migration. A few additional tips might also prove beneficial.
- Have a recovery plan for every step of the migration. Know how you will roll back the step should something go wrong.
- Validate periodically during each phase. Waiting to validate data until everything is done is asking for problems.
- Assign specific people to be responsible for the various steps, such as data validation or cleansing, and make sure that everyone knows who has authority over decisions involving the data.
- Review your business rules related to data. If necessary, alter them to make sure that the migration can succeed, even if you must change them after migration.
- Plan how you will manage your data after the migration to keep it clean. You might need to establish a routine “scrubbing schedule” or address issues with your source.
Data migration can be a complex project, but with adequate planning, you can reduce the chances of failure.