The line item in a replatforming proposal that's almost always under-priced is "content migration." Two words, a couple of grand, and a sense from both the agency and the client that it's the boring part — somebody manually moves the words across, the design handles the rest. This is wrong, and it's the reason replatformed sites so often launch and underperform the site they replaced.

My position, clearly: content migration is the highest-impact data-quality work in any website rebuild. Get it right and the new site outperforms from week one. Get it wrong and you've paid $80k for a faster version of the same problem.

I'll explain what "right" means, what almost everyone misses, and what the realistic cost of doing it properly is.

The case for treating content migration as design work

Most migration proposals frame content as a logistics task. Extract from the old CMS, transform to match the new schema, load into the new CMS. ETL. Boring. Nominally done by an intermediate developer or a project coordinator.

The reframe that actually predicts success: content migration is the moment to renegotiate the relationship between every piece of content and every other piece of content. The old site grew organically — a service page added in 2023, a campaign landing page that became permanent, a blog category nobody touches, four product pages that should have been one. The migration is the only chance to re-architect that, because nobody's ever going to volunteer to do it after launch.

If the migration is treated as copy-paste, the new site inherits the architectural debt. If the migration is treated as a content design exercise, the new site is structurally better.

What an honest content audit produces

Before any content moves, the audit produces a spreadsheet. Every URL, every piece of content, with four columns of judgement that almost nobody fills in:

1. The traffic column

Pull 16 months of Google Search Console data. For every URL, the total clicks, the average ranking position, and the top three keywords. This tells you which content is doing real work and which is just sitting there.

The honest cut: anything that's earned fewer than 50 clicks in 16 months is a candidate for deletion or merger. Anything ranking page 4 or worse for its target keyword is also probably not pulling its weight. The exceptions are pages that don't need to rank — homepage, contact, about — which obviously stay regardless of traffic.

2. The conversion column

Cross-reference Search Console with GA4 goal completions. Which pages are converting? Which ones are entry pages for visitors who eventually convert elsewhere? Conversion data outweighs raw traffic — a page with 200 clicks and 12 conversions matters more than a page with 4,000 clicks and zero.

The grim reality of most small business sites: 80 percent of conversions come from 20 percent of pages. The 80 percent of pages doing 20 percent of the work are candidates for consolidation.

3. The quality column

This is the judgement call. For every piece of content, is it actually any good? A piece of content from 2021 that ranks well but reads as dated, references outdated regulations, or links to dead resources — that's content with a half-life. Migrating it as-is and shipping a 2026 site with 2021 prose is a missed opportunity.

The quality assessment is usually a junior copywriter or marketer's job. Two minutes per URL, three buckets: keep as-is, rewrite, delete. For a site with 80 pages this is 160 minutes. For a site with 800 pages it's a few days. Either way, it's the single most important step.

4. The intent column

What is each page trying to do? Inform, convince, capture? Pages that don't have a clear job are usually pages that don't get any job done. The migration is the moment to assign every surviving page a specific intent and structure it accordingly.

What gets thrown out (and the courage to do it)

The single hardest conversation in any content migration is the one where you tell the client that some of their content is dragging them down.

That blog post from 2022 with 12 comments and 30 backlinks — but covering a service the business no longer offers? Probably needs to be killed and 301'd to the current equivalent. The "About Us" page split into six sub-pages back when someone decided that was good for SEO? Probably needs to be consolidated. The "Resources" section with 40 PDF downloads behind a form that captures emails nobody reads? Probably needs to be either reactivated with a real strategy or quietly retired.

I've watched agencies lose this conversation because the client emotionally identified with content they wrote four years ago. The migration becomes a faithful preservation of every word, and the new site launches with the same bloat. Twelve months later they're paying someone else to consolidate it anyway.

The honest version of the conversation, said up front: a typical small business site can lose 30 to 50 percent of its pages with no traffic impact at all. Deleting them and redirecting them to better-targeted equivalents produces a measurable improvement in the remaining pages' authority. The math is in Moz's research and consistent across other SEO consensus — fewer, stronger pages outperform more, weaker pages.

The structural rewrite (what the new CMS lets you do)

Once the audit decides what survives, the next decision is how it's structured. This is where moving from one platform to another either delivers value or doesn't.

Squarespace, Webflow, WordPress: each has its own content model with its own constraints. A custom build on a modern headless CMS like Sanity, Payload, or Storyblok lets you design the content model around the business, not around the platform.

The shift that matters: structured content. Instead of a "Services" page that's a long-form rich text blob, the new model has a Service content type with explicit fields — name, summary, full description, pricing, FAQs, related case studies, related team members, related blog posts. The page is rendered from those fields, which means:

Fields can be reused on other pages (a service's summary on the homepage, the full description on the service page itself)
Content editors fill in fields, not rich text, which keeps the design consistent
The schema markup the site renders is automatically derived from the structured fields, not hand-written
Related-content sections are dynamic instead of manually maintained

The platform shift is what enables this. Doing structured content in Squarespace or out-of-the-box WordPress is fighting the tool. Doing it in a headless CMS is the natural pattern.

The metadata layer (the SEO equity bridge)

Every piece of content migrated needs its existing metadata preserved. This is the SEO equity bridge between the old site and the new one.

For every URL the audit kept, the title tag and meta description from the old site come across to the new one verbatim — unless there's a specific reason to optimise them, which there sometimes is. The default is preservation. Years of accumulated incremental SEO tuning live in those tags.

The same goes for:

Heading structure (H1, H2, H3 hierarchy) — should match the existing pattern unless the audit decided to restructure
Image alt text — should carry across, not be regenerated
Internal anchor links — should be updated to the new URLs, not left pointing at the old ones through redirects
Schema.org markup — should be reimplemented on the new platform, ideally improved (because most existing implementations are incomplete)

Each of these is a small piece of equity. Skipping any one of them is a small concession. Skipping all of them is the difference between a migration that recovers in 4 to 6 weeks and one that recovers in 6 months.

The image migration problem

Almost every migration underestimates the image work. Two specific failure modes:

Source quality. Squarespace and Webflow serve images at platform-optimised sizes, not original quality. A 1200x800 image on the source site might have been originally uploaded as a 4000x3000 file the platform downsized. If the migration scrapes the rendered HTML for image URLs, you get the platform-resized version. The new site needs the originals to render its own responsive sizes properly.

This means going back into the source CMS to download the original image library, not pulling images from the live site.

Image SEO. Image alt text, file naming, and rendered context all carry SEO weight. A clean migration preserves alt text and improves file naming where the originals were IMG_4892.jpg. The new site's responsive image rendering should output multiple sizes via srcset and serve modern formats (WebP, AVIF) via the framework's image optimisation pipeline. In Next.js this is the next/image component. In other frameworks it's the equivalent.

The image migration is often a week of work all on its own for a content-heavy site. It's almost never quoted as a separate line item.

The post-migration content review

Two weeks after launch, the content audit happens again. Same spreadsheet, same metrics, with comparison columns. Which pages are performing as expected? Which have dropped? Which surfaced new behaviour — high bounce rate, unexpected entries from a new ranking query?

This second audit is what catches the consolidation decisions that were wrong, the metadata that didn't migrate cleanly, the internal link that's still pointing at a deleted page through a redirect chain. The fixes are small individually and they add up to a meaningfully better trajectory.

Most migration projects don't include this. The agency ships the site, invoices, and walks. The post-migration review either doesn't happen or happens in a panic six months later when somebody finally notices the traffic plateau.

The real cost of doing content migration properly

For a typical Australian small business site with 80 to 200 URLs, doing the content migration properly — audit, judgement, structural rewrite, metadata preservation, image migration, post-launch review — is usually 30 to 50 hours of work. At normal Australian agency rates, that's $4,500 to $9,000 as a line item.

In most replatform proposals, content migration is quoted at $1,500 to $3,000. The gap is the corner that's being cut.

The honest bottom line

If you're scoping a website migration of any kind — Webflow, Squarespace, WordPress, Shopify, anywhere to anywhere — the content migration is not the boring part. It's the part that decides whether the new site outperforms the old one.

Ask the agency quoting you what the content migration line item includes. If the answer is "we'll move the content across," that's not enough. If the answer mentions audit, structural rewrite, metadata preservation, image migration, and post-launch review — that's the answer that predicts a successful migration.

A useful starting point: run a free audit on your current site. It'll surface the rendered metadata, the orphan pages, the image weight, and the specific content issues that a faithful migration needs to either preserve or fix. If the audit finds problems the migration proposal hasn't priced in, that's the conversation to have with the agency before signing — not three weeks into the build.

Why Content Migration Sinks $80k Replatform Projects