After having done many large-scale WCM re-implementations, we've seen many ways Content Migration can be the Achilles heel of a site redesign project. It is typically THE MOST underestimated portion of a WCM implementation.
Here are some basic tips to consider at THE BEGINNING of your WCM re-implementation project:
Will you be attempting automated or manual migration?
If you are doing automated migration, you will likely only be able to do the items that have 200 or more of the same structured type of content. For instance, you can migrate all the News Articles or all the Products, but the rest will likely require manual migration. You will likely only be able to do the center body content, not the right rail or left rail modules. Also, automated migration IS A REAL DEVELOPMENT CYCLE. It will require some thinking/design up front and it will require iterative development and testing. Most likely you will need a separate environment to do it as well.
If you will be migrating manually:
You will not be able to start manual migration until the WCM input templates are > 80% complete and tested. Many times these are the same as the output templates for products that provide 'inline editing', but with additional tags specific to the WCM that provide help text and additional features for the editor. If you try manually migrating before the templates are 80% complete, you will risk having to re-add the content because template structure may change. Manual migration of a system that is < 100% complete will typically take 30-45 minutes / page, depending on the items that are required to be done (below).
Plan for 2 passes:
Whether you are migrating automated or manual, you should plan on doing multiple passes through all the content. Most times, it is not possible to do everything in one pass if you have shared content and pages with a complex navigational or relationship structure. The first pass is to load the content, the second pass to link the content together (you can't fix an embedded link or image until the target content is loaded), and potentially a third pass to set the navigation hierarchy and reset the meta-data if that's not possible in the first or second pass.
If you attempt to migrate between 80-100% complete:
You will require a technical resource. Most likely the WCM input final touches are the last things to be built. This means that a person who is okay with seeing bugs and working around them will be required. Typically, this means the resource will be more expensive. If you want to hire a low-wage temporary worker, you really can't have them migrate the content until the system is 100% complete and tested.
Typical migration steps that are forgotten until you start:
Manually migrating content isn't just copying and pasting pages one at a time, as some people guess and plan for. After you get the body content in, you need to:
- Format the text (especially important if you are just pasting into a text blob OR if your style has changed dramatically). Also the quality of the WCM text editor will improve or hinder this significantly.
- Upload all embedded images and re-add them to the pages (remember the URLs to the images are likely to have changed)
- Assign SEO URLs
- Assign where page fits in navigation structure
- Fix all embedded links (again, remember all the URLs have changed)
- Assign any meta-data to the page that isn't content that the user sees (such as URL, Title, keywords, etc)
- Assign right-rail modules that you'd like to have (even if they are the same modules as before, they will be connected differently in the new WCM and most times, there are new ones in a re-implementation)
- Handle any new features that have been added as a result of the new WCM such as personalization, analytics, etc. (you went to a new WCM for a reason, right?)
Did you change the IA?
If you are planning on changing the Information Architecture and/or Visual Design of your pages (page structure or styling), it will make it much harder to automate your migration and will most likely require manual steps. This is something you should know and plan for up front. Don't tell your UX team they have the ability to change the entire site UX and then tell your content migration team you want them to devise a way of migrating content automatically.
Did your Template Granularity Strategy change?
Granularity Strategy refers to how granular your templates are within your CMS. The more granular your templates, the more control your development staff has over enforcing the functionality and UX, but your editors need to do more clicks (think "Add center body list" module, "Add center body image" module). If your templates are less granular, it means the editor has more free-form control over the page, but potentially can make each page look drastically different from the other (think "Page with Center Blob of Text"). If your previous WCM had 10 modules on the page and you are planning to build your new WCM with 2 modules on the page (to reduce the complexity and increase flexibility for editors), then you will have a much more difficult time migrating the content, whether automated or manual. The same holds true going from a LESS granular to a MORE granular template strategy. This is something you can recognize up front and plan for when determining how long the migration will take you.
There are tools that can help you do content migration (i.e. make more automated), but they will still follow the rules above.