Data migration is a critical, high risk, high reward moment. If the migration process maintains the accuracy and accessibility of legacy data, the CMS implementation becomes largely successful, and the CMS will add value to the organization. However, the process may go wrong for several reasons, ranging from poor planning to inadequate domain expertise, and from technical glitches to poor hardware. Such incompetent data migrations may result in cost overruns, project delays, and even potential data corruptions, which in turn lead to other grave issues.
Here are some pointers to consider when embarking on a data migration exercise. Paying heed to these considerations would help in successful data migration.
- Understand migration complexities and prepare to succeed
Successful legacy data migration requires anticipating all eventualities, and planning a route map to avoid potential data loss, minimize user disruption, and reduce downtime.
The first step in preparing for data migration is a comprehensive assessment, to identify the traps in the migration path, the infrastructure required at the other end, and other critical factors. A sound migration assessment generates a migration plan, and also deliverables related to content inventory, platform architecture recommendations, content modeling for target system, and an assessment of potential risk factors. The migration team has to provision for all these before starting the actual migration process.
A comprehensive migration plan incorporates:
- Definition of the project scope. Create meaningful cost-estimates, and get top management buy-in.
- Inventory analysis to identify the data to be retained as well as discarded. The scope of the inventory analysis may extend to CMS, blogs, message board forums, application data, and other web content.
- Clarity on the migration process, including tools, procedures, and other resources.
- Identification of risks associated with data migration and action plan to mitigate such risks. Do not forget to back-up the data earmarked for migration. Several things can go wrong with the migration exercise, leading to lost or corrupted data.
- Timeline of actual migration phases.
- Testing and validation processes: A discovery and gap analysis will ensure that all the required data is on board. A one-to-one mapping may not always be possible, as disparate pieces may have to be moved to new locations in the new data structure.
At times, data migration can break down the integrity of relationships between digital assets. The reason is that we often forget to look at the underlying complexity of relationships that may not be apparent on the first sight. For instance, content restructuring might complicate metadata and link reassignment. Moreover, digital assets that live outside the source CMS take time to retrieve.
Even though most CMS systems have multifaceted data models to handle heterogeneous data relationships, the content migration plan needs to catalog all legacy systems and their relationships.
- Have appropriate personnel in place
Several ad-hoc factors influence data migration, and it requires having a team of skilled project personnel that undertakes complete ownership and responsibility of the data to take the process forward. Even the best laid plans can go wrong, and unforeseen issues crop up during the actual implementation process.
While it requires technical expertise, usually from the IT team, to conduct the actual migration, involvement of business managers is indispensable to ensure the migration of the right data.
Successful data migration requires the efficient involvement of:
- A project manager who oversees the project
- A CMS Administrator, who has in-depth knowledge of both the existing system and the ways the company plans to use the new system after migration
- Technical hands responsible for executing the actual migration process including extraction of data from legacy sources, transformation of data as required to suit the new system, and actual transfer of data and validation in the new environment
- Business managers, who bring to the table the business perspective of migration
- End-users, who best understand the nature, flow, and process of the data
Lack of technical or subject matter experts among internal employees makes it essential to bring in external consultants. Roping in consultants might be a good idea for complex projects, to avoid disruptions in the job profile of in-house employees.
- Assess the value of historical data
Not all information is worth migrating. Migrating obsolete information overburdens the migration process and makes it more complex. However, the migration team needs to strike a fine balance between the content to migrate and content to be left alone, and this requires a thorough and in-depth understanding of the business.
Relevance of content depends not just on the topic or subject matter alone. Use a tool that scours available content across repositories to identify duplicate content, historical content that is no longer useful, broken links, and more.
- Understand target system requirements
Any CMS would have special requirements for its data, such as compatibility, completeness, and cleanliness. Failing to meet such requirements will increase the cost and time of migration, and also sap the integrity of the migrated data, thereby derailing the entire process.
It is critical to cleanse the data first, by:
- Locating and eliminating obsolete or redundant data
- Identifying incompatible data that needs reformatting
Besides, there are other considerations such as load sequencing and scheduling, and configuration and customization points to look at while preparing for migration.
- Utilize appropriate migration tools
There are several ways and approaches to transfer data from its legacy silos to the new CMS. The popular options include manual loading, custom scripting, and bulk loading using software. The following pointers will provide you some insights on how to choose the right approach:
- Manual loading best suits small amount of data stored in homogeneous databases. When thousands of interrelated files are in the mix, manual loading becomes ineffective, time-consuming, and expensive and is also prone to errors such as incorrect data entry, misplaced data, or data omission.
- Custom scripting extracts all assets, including metadata and links, and transforms the same into a new CMS format, loaded into the new system with links reassigned. However, scripts, especially custom scripts, could malfunction and require modification resulting in time delays and cost overruns.
- Bulk loading software supports heterogeneous content storage environments, and eliminates the risks associated with data compatibility, cost, and scheduling. This method, however, adds to the cost and resource requirements. An API or web services-based method makes the transfer less complex. In the absence of such tools, it would require either direct access to the database or an export format, both requiring technical expertise. Another good option is developing a custom plug-in for the transfer.
There is no best method over another, and the pros and cons of each method is often related to specific circumstances and data type. The onus is on the migration team to weigh the advantages and disadvantages of each approach, depending on the particular needs of the organization.
- Do not overlook anything
Make sure not to overlook anything. Some content may appear trivial but would have great implications, if missed out. An example is 404 pages for websites, some of which may have quality links pointing to them.
For instance, make sure to map the URLs of the legacy site to the URLs on the new site. Things can go terribly wrong if this is not done with precision. Make sure to implement redirects using the permanent 301 redirects that pass most link equity from the old page to the new page.
Migrating legacy data is often a long drawn out event, and depending on the size of the organization, quantum of data, and complexities involved, can extend to several months.
To know more about Suyati’s CMS platforms and data migration solutions, write to email@example.com.