Migrating a website or an intranet to a new CMS or new information architecture comes with considerable challenges. Automated tools can support the process, but almost inevitably a great deal of discussion and work will need to take place around specific areas of the site. Having information management policies and strategies in place goes a long way in helping to define content quality, metadata and a taxonomy for the new site. Moving file shares and other repositories into a cloud storage and application environment such as Google Drive has emerged as another requirement.
Effective search will play a vital role in adoption if the migration results in any changes to the intranet structure, and will continue to do so until users have found their way around the information architecture (IA), new content types and repositories.
I’m in the middle of three such migration projects at present. How to manage the search experience during and after the migration process has come up as a common theme in all. Search usually receives little attention in a migration project, but as it becomes more business-critical, the importance of including search-related issues at the outset of planning grows.
There are two options. To spread the effort out over a period of months, one option is to move sections of content to the new site as the content is revalidated. With this approach, the IA and the search application need to be in place from the moment the first content is migrated so that it can be tested in situ. Usability testing can be carried out on a sectional basis, but there will not be much value in searching the gradual accretion of content until most, if not all, is migrated. Screen design and some aspects of search performance related to metadata, etc. can also be appraised, but the total search experience needs the total content.
The second option is to work on each section of content but move it in a single process. This gives longer for the design and implementation work to be carried out, but makes it difficult to pin down problems with the search functionality. The results for a given query may include some obviously spurious hits but discovering the reasons why might take a lot of time and effort. If the problem is related to inconsistent metadata, a review of a substantial section of content might need to happen. It is also impossible to assess latency and other IT-related performance issues (e.g., security) until the entire site is migrated and perhaps linked via search to other applications.
Moving content into a cloud application also brings up the issue of which search application to use. On-premises search applications may be difficult, if not impossible, to link to a cloud application. Google Drive comes to mind here. In my experience search requirements are rarely included in the initial specification because the driver is one of reducing cost and "improving collaboration," with little or no thought for how the content will be discovered. Sorting out search analytics will also be an interesting task.
Planning for Success
It all comes down to very careful planning, which starts before scoping out the migration project. Waiting to optimize search until all the content has been migrated carries large risks. Search may be even more important for certain sections of content. The degree of search importance, or perhaps complexity, should be a factor in establishing the migration schedule, the training and other support needed during the migration.
Planning needs to extend beyond the notional close of migration because the full launch will probably be several months off. Search will need to be stress tested in the pre-launch period with close monitoring of search analytics, especially if the scope and quality of the content has been enhanced. The deletion of some content may also come as a surprise to some search users, not only because of its disappearance, but because of the effect on the ranking of results.
Managing Risk and Expectations
Performance should improve with enhanced content quality and better metadata search. However, a change in application interface or unexpected results appearing on the first few pages of search results could have a serious impact on business performance. This can undo all the positive expectations of the migration.
There are only two rules in migration. One is to start planning as soon as possible, and the other is to accept that it’s not easy to plan. In my experience no one know more about migration planning than David Hobbs, so start with his Web Migration Handbook.