Far too often, organisations approach data migrations with eyes wide shut, possessing little visibility into what they are getting themselves in to.
As discussed in the article, 'Why You Need A Migration Readiness Assessment', a Migration Readiness Analysis (MRA) framework can shine much needed light into a migration project before it commences. Like a dry run preparatory to a major household move, an MRA enables you to move one or two critical data elements to illuminate project challenges, scope and risk.
This article discusses how you can leverage this advance-intelligence to jump-start your data migration project and mitigate those risks.
Improved strategic planning and business buy-in
A two week MRA exercise will uncover many of the data and mapping risks inherent to a specific migration project. Based on the results of data analysis and profiling, you will have documentation on source data issues including issues around data and metadata, entities and attributes.
Based on the results of an exploratory, dry-run mapping of one or two chosen entities, you will see the mapping issues encountered in mapping to the target, and how the various source data issues were handled.
This information can be leveraged in many ways. First, it enables you to develop an optimum migration strategy and step-by-step project plan. Second, it can help you engage the appropriate business users in the project. Migration is more than just an IT challenge. It requires business user involvement, since it is they who understand the semantic meaning of the data. Without their active involvement, the project will fail to deliver the expected value.
By spotlighting data and mapping issues up front for all to see, an MRA can help shift the emphasis of the project to the business users, and get the managers of the target system and its users to buy-in to the project.
More accurate scoping and estimation
Being able to accurately scope and estimate a migration project is essential to its success. Bringing in a migration specialist to physically map one or two entities lets you see exactly what efforts are involved to move the data successfully across to the target.
From this, you are able to extrapolate what it will take to move all the other entities. In working with the business users who understand the data, you can assign high, medium or low classifications to each entity to come up with a good timeline and cost estimate, and have confidence in your projection.
Optimised staffing and teamwork
It is highly unlikely that a company will have migration specialists on staff, and "re-purposing" inside people to play leading roles in a migration isn't necessarily the right approach. In fact, it will likely subtract value from other key projects while not doing justice to the migration.
The key is to engage the right outside migration specialists and then match them to the appropriate inside people (experts in your data and your processes). You will need a business-oriented subject matter expert for each source system as well as one for the target system. Happily, they are usually the same set of people, and an MRA will help determine who they are.
You will also need a technical person who understands how to get the data out of each source system and how to put it into the target. The migration specialists will significantly reduce the amount of time the inside people will have to devote to the migration process, and starting this off with an MRA will set the stage for successful teamwork going forward.
Streamlined conversion process
An MRA also sets the stage for streamlined data conversion. The typical conversion process consists of three discrete phases: analysis (mapping, discovery and technical), build and test. Usually, less time is spent on analysis than on the other phases.
Because the analysis is typically insufficient, things tend to fall apart during build and test, and you have to go back to analysis and start over. Conversely, it is easy to get into analysis-paralysis mode and waste time over-analysing. In either case, serious project slippage can result.
Performing an MRA using advanced analysis and automated profiling techniques helps put proper emphasis on the analysis phase while jump-starting the build and test phases. An MRA is a good first step towards discovering relationships, uncovering quality issues and validating assumptions.
From this base, you can proceed with refining designs, applying best practices and building leverageable in-house expertise. The result is a faster conversion with less need to return to analysis.
System cutover and legacy retirement planning
Eventually, users are going to have to move over to using the migrated data in the target system. During an MRA, you can start planning how and when to wean users from the legacy system. As you have business users already involved in the MRA, you can better plan where people will go for their data during cutover, what will happen to transactions, etc.
You will also be able to plan when to shut down the legacy system. As long as it is up, after conversion has taken place, you essentially have two systems running the business. An MRA can help you project how much time and effort it will take to retire the old system.
Improved risk mitigation
We have discussed the data risks and mapping risks inherent to migrations, but there are many other risks that can contribute to cost and budget overruns. These include personnel, technical and implementation risks.
Performing an MRA can help you identify these risks and create backup plans for such contingencies as target system implementation delays, sudden personnel departures, and technical glitches.
Small investment, immense return
An MRA is a strategic framework for a migration project with benefits that resonate throughout the entire migration process: improved planning and budgeting, better understanding of challenges and risks, better resource utilisation, and enhanced risk mitigation.
It helps an organisation deal with unknowns and uncertainties at both ends of a migration project - in the source data and in the requirements of the target system.
It fosters teamwork and gets business users involved in a migration at the earliest possible point, thus helping to drive overall project success. And it helps put a data migration project in proper perspective within the overall new application implementation or system migration initiative.
Given the failure rate for major data migration projects - an 80% likelihood of exceeding time and monetary budget, and a 33% likelihood of failing all together - leveraging the MRA framework is not a nicety, it's a business imperative and the best and least costly way to avoid becoming a statistic.