Taking the leap: Migrating to a new platform

Change data capture (CDC) migrates only new and changed data from the source platform to the target platform in real time or near-real time. “CDC is a great option if you need timely data and want to minimize the impact on your source system,” Shah noted. Finally, using a hybrid approach combines more than one methodology; for example, use a big bang approach for historical data and then switch to CDC for ongoing synchronization.

For most organizations, “day forward” is the best approach, according to Carmichael, with new content put on the new platform and old content migrating over as a second priority with a longer project timeline. “It is better not to completely discard old data, because it can have value for predictive analytics,” she commented. “That content converted into knowledge can be a game changer.” Old information that has been accessed and used can be brought into the new system, while dormant data can be reviewed for retention policies and brought over as necessary. “It is also important to have a human in the loop at some point,” Carmichael added.

Best practices for migration

The key best practice is upfront planning, including having an understanding of the potentially far-reaching impact of migration. “When an organization reaches a pain point where they need to sunset a system, they often become too focused on that one issue,” observed Carmichael. “They do not necessarily understand the full scope of the migration process.” She recommends spending time bringing people into the process and creating consensus. “Change management is a critical part of moving to a new platform.”

Developing a detailed plan for the migration takes time, but is a good investment. “Rushing into things can lead to delays and compatibility issues down the line,” Shah maintained. “The plan should cover everything from timelines to resources. Check to see that the new platform supports all your data types and functionalities, and try to schedule any downtime for off-peak hours.” Keeping everyone engaged in the planning process can help manage expectations and avoid misunderstandings.

As might be expected for KM applications, clean data is critical to system performance and a smooth transition. “Reconsider metadata as you move ahead,” advised Carmichael. “Some older repositories may not even have metadata, and a surprising number of organizations are still paper-based. There are tools that create metadata on a large scale that can expedite this process. Without the right metadata, system performance and information access will not match expectations.”

Working with the platforms

Although large-scale migrations may require the skills of systems integrators, moving from one KM platform to another can also be done in-house with assistance from the platform vendor. “Our customer success team assists organizations that have in one way or another outgrown their existing platform and need to change,” said Betsy Anderson, manager of the customer success team at Bloomfire. “For example, the platform may not scale either in terms of number of users, volume of data, or organizational shifts.” In addition, data has become more complex as knowledge programs have become more essential to enterprise-wide initiatives, and legacy systems may not be flexible enough to accommodate the greater complexity.