Data migration is complicated. When companies underestimate the attention and planning required, strategic enterprise-wide initiatives like product lifecycle management (PLM) implementations can be badly derailed. Schedules slip, go-live dates are missed and additional resources are required to get back on track while keeping sponsors happy and teams motivated.
In our first entry of this three part series, we addressed the first important question for ensuring a successful data migration: Who is the Data Migration Team? Once you have a clear picture of the business and IT resources needed to execute migration tasks and drive the process, focus can shift to the needs of the business.
The second part of this series offers practical tips to help build alignment across all impacted parties with a clear plan and established business expectations for migration activities.
Question Two: Are Migration Expectations Clear to the Business?
Business resources play a fundamental role during migration. They are exposed to risks if delays are encountered, and they reap the benefits when the system goes live. The Data Migration Process Manager should work with the business early in the project to:
- Confirm that business resources understand the key steps for migration
- Define the finish line (What is success?)
- Offer options for data freezes
- Build in room for ambiguity
Confirm Understanding of the Key Steps
To get started, clarify the process for migration and business teams by explaining the four major activities:
- Data Identification/Extraction – Determine what data is moving and where is it currently located
- Data Cleansing – Fix any incorrect data
- Data Transformation – Adjust data to adhere to new governance standards
- Data Load – Import into the new system
Once the approach is clear, the Process Manager should work with the business to understand the current state of data and begin defining scope. Initial scope assumptions should be refined with project management and affected business groups or division leads to develop a detailed plan.
Define the Finish Line
To define what a successful data migration will look like, it is necessary to understand the scope of the data to be migrated, the level of "cleanliness" required, and timing for migration. The time and resource constraints of the business should influence the proposed scope. The Process Manager needs to incorporate each group’s scope and timing feedback into a cohesive plan to establish the overall project goals. It is this plan that all parties must buy into in order to achieve success.
Offer Options for Data Freezes
Every data migration plan should address the “freezing” of data. It would be ideal to freeze data for the entire data migration period and complete cleansing and transformation exercises in a vacuum. But in reality, business must continue. While there is no single prescriptive model to best meet all needs, consider the following options when defining the migration timeline.
Multiple extract model with a short-term data freeze – minimizes the amount of time that the business is completely out of a system
- Initial extracts are used to cleanse the majority of data while the business continues working in legacy systems
- Shortly before the new system is to go live, business data is frozen and a final extract and cleanse is performed during the cutover window
Single extract model with external tracking of changes – allows business to continue as usual with additional change tracking for all updates
- All data is extracted at a single point in time and then cleansed
- Business utilizes external tracking tools to log any changes to business data between the extract and go-live dates
- At go-live, migration lead and data entry clerks manage the translation of modified data
Combination model with multiple extracts and tracking of changes. As a safety precaution, all plans should address the need to make emergency changes during a data freeze. No matter how short the freeze window, the business should have a plan to manually track and manage work while systems and data are frozen in final preparation for go-live.
Build in Room for Ambiguity
Data migration is never a smooth process. Despite careful planning, there are always bumps in the road and unknowns that must be overcome. The Data Migration Process Manager should help the business recognize that data comes in many forms from various sources, and that data discovery doesn’t happen all at once. One type of data may be stored in Excel documents in one location, an Access database in a second location, PDF files in a third location, and in a legacy, home built system in a fourth location.
Beyond data types, you may encounter data that contradicts the established governance rules. The migration of data in each of these situations requires different skill sets and time requirements which can be hard to estimate accurately. With so many factors, even the best planning effort can be inaccurate. It is the job of the Data Migration Process Manager to coach the business to recognize and accept a certain amount of ambiguity in the process.
With the business’s buy in for the plan, the final step is to make sure the team is prepared to execute against the plan with the tools they need. This will be the topic of the final post in this series.