Implementing a new Knowledge Repository requires converting and migrating existing knowledge artifacts from numerous repositories and organizing those artifacts around taxonomy domains. The industry rule of thought is to complete the migration prior to release (Big Bang) rather than releasing partial completed knowledge domains (Swiss-Cheese). The positives for the Big Bang approach is largely focused upon fostering user buy-in with the new knowledge repository. It allows old repositories to be shut down and moves all usage to the pristine new repository. When users have the options to choose between a new repository and existing repositories, historically users will stay with the tools they are most comfortable with.
With large migrations, taking six to twelve months, the Big Bang approach provides zero value until production launch. Releasing the new Knowledge Repository prior to 100% of the migration completion may make sense. We had the opportunity to try both approaches for comparisons within identical cable customer care call centers. Location ABC went with the Big Bang and Location XYZ the Swiss-Cheese. Both locations have approximately a million customers and 900 support Agents. Both locations offer the same services to their customers but have different processes managing their service offerings and different equipment used by their customer. The domains are segmented by High Speed Data (HSD) housing troubleshooting artifacts for Digital Phone and Broadband services, Video Troubleshooting, and Video/HSD Sales & Billing.
Location ABC migrated each of the three Taxonomy segments as a whole and shut down correlating legacy repositories with the launch of each. They started with the HSD content, then Video/HSD Sales & Billing, then launched the Video Troubleshooting content. The migration took approximately ten-months; HSD five-months, then Video/HSD Sales & Billing three-months, and finally Video Troubleshooting two-months. Location XYZ, migrated content from all three areas simultaneously; focusing on prioritizing the value of the content. After four-months, Location XYZ’s knowledge repository consisted of 65% of its estimated content needs. Instead of waiting another three to four months until conversion completion, we decided to launch the repository in the current 65% state. The legacy repositories were not shut down but a feedback mechanism was in place, and the migration team continued adding content daily.
The results of both approaches were overwhelming successful. Location ABC realized user buy-in quickly since the content was complete and legacy repositories were shut down. Location XYZ took longer to obtain the same success since users were forced to navigate between the new and the old repositories. One unexpected benefit for Location XYZ was the user feedback mechanism. Users overwhelmingly provided an average of 100 weekly feedback items allowing the KM team to focus upon those request over the predetermined 35% of estimated content planned for migration. This new variable eventually discontinued the remaining 35% of content planned for migration. The team solely focused on feedback and new content initiatives (new products & services, promotions, etc.). Implementing bi-monthly surveys for each Location and usage/feedback metric tracking, no notable metrics distinguished one method over the other. The only two notable variables extracted from the two approaches are: 1) the Swiss-Cheese approach for Location XYZ offered a shorter time to production and its quick adoption moved 75% of users voluntarily off legacy content portals, and 2) the Swiss-Cheese approach doesn’t allow the immediate shut down of legacy content portals and we soon found those portals no longer being maintained, and stakeholders soon referenced outdated and unmaintained artifacts.
Comments