Migrating from a Legacy CMS Without Losing a Year to Content Debt
A CMS migration usually gets scoped as a technical project: pick a new platform, map the content model, move the pages, redirect the old URLs. The part that actually eats a year is the content itself. Every legacy CMS accumulates pages nobody owns, copy nobody has reread since it was written, and structure that made sense to whoever built it and to no one since. Move all of that into a new system as-is and you haven't modernized anything. You've just given your content debt a nicer coat of paint, and a slower team to maintain it.
What content debt actually is
Content debt is what accumulates every time a page gets published as a stopgap, a redundant version gets created instead of updating the original, or an owner leaves and nobody reassigns their pages. None of it looks urgent in the moment. All of it becomes a tax on every project that comes after, migrations included.
A migration is usually the first time in years anyone has looked at the whole site at once, which makes it either the best opportunity to pay that debt down or the easiest way to lock it in for another five years, depending on whether the content gets evaluated or just copied over.
Start with an inventory, not a migration plan
Before deciding how content moves, someone needs to know what actually exists. Nielsen Norman Group draws a useful distinction here: a content inventory is a factual catalog of every page and asset, while a content audit is the judgment call on whether each piece is still worth keeping (NN/g). Skipping straight to the audit, or skipping both and going straight to migration, is how legacy problems make it into the new system unexamined.
A working inventory typically captures, per page:
- URL and current owner, if one still exists
- Last updated date
- Traffic or usage over the past year
- Whether it duplicates or contradicts another page
- Whether it maps to a real content type in the new system, or was a one-off workaround in the old one
This is tedious work, and it's exactly the kind of task worth applying some automation to. Crawling tools can generate the raw inventory in hours instead of days; AI can help cluster near-duplicate pages or flag copy that hasn't been touched in years. But the judgment about what to keep still needs a human who understands the organization, because a page with low traffic isn't automatically a page worth cutting. It might be the one page that matters to a small but important audience, like a compliance form or an accessibility statement.
Audit against what people actually need, not what currently exists
Gerry McGovern's Top Tasks approach reframes the audit question. Instead of asking whether each page is well-written, it asks whether it serves something people are actually trying to do, and treats everything else as noise competing for the same navigation and search real estate (A List Apart). Organizations that apply this lens to a migration often end up cutting a large share of existing pages, not because the writing was bad, but because the content was never load-bearing in the first place.
For Marker Seven's government and education clients, this step matters even more. Compliance content that looks redundant may be legally required. A course description that looks outdated may still be the page prospective students search for by name. The Top Tasks lens works best applied by people who know which of those constraints are real and which are just habit.
Where migrations actually go wrong
A few patterns show up repeatedly in CMS migrations that carry legacy content debt straight into the new platform:
- Lift-and-shift as a strategy. Moving outdated assets, duplicate pages, and unused content types into the new CMS because auditing felt like it would slow the timeline down, which just relocates the cleanup to a future project (TechRadar).
- Broken redirects and lost search visibility. Migrations carry real risk of dropped rankings and dead links when URL structure changes without a mapping plan, especially on large sites with years of inbound links pointing at old paths.
- No owner for the audit itself. Content inventories and audits need a named owner who keeps them current, and that responsibility tends to get treated as a side task nobody has time for, which is exactly how the next round of content debt starts accumulating on day one of the new system (NN/g).
- Underestimating how long content decisions take. The technical migration often has a clear timeline. The content decisions rarely do, because they require pulling in stakeholders who own pages they haven't looked at in years.
What a clean migration actually requires
None of this means every migration needs a multi-month audit before a single page moves. It means treating the content decisions with the same rigor as the technical ones, and building the timeline around both. In practice that looks like:
- Inventory the full site before scoping the migration timeline, not after
- Audit against real usage and real tasks, not gut feel about what "seems important"
- Assign an owner to every page that survives the audit, before it goes live in the new system
- Map redirects deliberately, especially for pages with existing search visibility or inbound links
- Build a lightweight review cadence into the new CMS from day one, so debt doesn't start rebuilding immediately
A migration is disruptive enough to force decisions that would otherwise get deferred indefinitely. That disruption is worth using on purpose. The alternative is a new platform carrying the same unexamined content it just replaced, which means the next migration, in a few years, starts from the same place this one did.