The nuclear option? Migration planning is more subtle than that.

Have fantasies of starting your website from scratch? I think we all do. For all the complaints that other people have about our websites, we know about an even larger set of problems. Especially for older sites with distributed content entry, you probably have a lot of low quality content. So what are some reasons we might want to think about just starting afresh with our content?

• Focus on what is important
• For the content you do need, you can spend time on getting it right
• Drop low quality information
• Lower maintenance costs
• Avoid some difficult internal conversations

The fantasy goes something like: 1) Nuke the current site and then 2) build a new site from scratch. The problem is that, aside from small sites, this fantasy rarely matches reality. Don't get me wrong: dropping non-performing content should always be a priority. That said, here are some reasons the fantasy of starting from scratch doesn't work:

• It presumes a redesign-forget-redesign mentality (in general, you should be managing your site for ongoing quality making constant incremental improvements) — for instance, instead of starting from scratch to avoid tough discussions about what content should remain, drive those decisions to the forefront early on by discussing general rules about quality that could be used on an ongoing basis.
• Your website probably has at least pockets of high quality information (your site probably isn't complete junk!).
• Site visitors often do need to eventually get to details (providing context and focusing the experience is important, but at some point the visitor may need to get those product specs).
• You may have source or definitive materials (for example reference data, original research, and technical product information).
• Your business may cover a lot of ground, so rewriting key information for all of your business may need to be phased in.
• It takes a lot of effort to write high quality content.

That last bullet probably sets off some alarm bells in your head, so let's talk more about that. I firmly believe that organizations should focus on what is important, and not just what is easy. But you can't built a house overnight. If you start from scratch, then you need time. If you have the time, then great. But usually time and resources (and the site complexity) don't allow a total nuke of content.

You can meet the goals listed above (focus, get it right, drop low quality, lower maintenance costs) with a hybrid approach:

1. Set the vision of your site.
2. Discuss what content is required to attain that vision + conceptually how the site needs to be organized.
3. Define rules for what existing content has to be completely rewritten, dropped, or other moved as-is (you may also look at other dispositions such as leaving the content where it is) + what brand new content is needed (to meet that vision).
4. Some of those rules will mean that you don't need to analyze some content in more detail — for instance, you may decide to drop entire microsites that were created over time. For the content that does need to be analyzed more, create an inventory that you can then apply the rules above to (consider multiple sources of information).
5. Estimate the manual effort it will take to pull that off, and repeat the above process until you get to a manageable set of rules. 
6. Write the required new content + other rewrites / moves / modifications as defined.
7. Apply those rules on an ongoing basis (for instance, define a lifetime for microsites and regularly decommission them on that schedule).

In this way, we address those problems with a completely from-scratch approach, for instance by moving over the high quality information that does exist, with the added benefit of keeping the information high quality over time. Also, we don't wind up with an overly-ambitious plan that we cannot execute on. Note that as a side benefit this is a methodical way to determine that actually you do need to start from scratch from a content perspective!

With this process you still may drop a dramatic amount of content. But, although we might like to, starting purely from scratch probably doesn't make sense for any but small sites.

Put another way, content migration planning is much more subtle than a decision to either a) ram in the content from the existing site or b) don't move anything at all. 

[ The hybrid approach of content inventories suggested by Kris Mausser in Why A Content Inventory Isn’t Always The Best Strategy, and the comment by CMKL referencing this video on the nuclear option, prompted this post. Also see Gerry McGovern's Web content migration: disastrous strategy and Paul Boag's Just say no to the madness of content migration. For a case study of using rules to drop content, see the Exploding Topic Pages slideshare. ]