How CMS Migration Impacts SEO

Every CMS migration ends with a number. Four hundred and twelve posts in the old system. Same four hundred and twelve in the new one. Everybody signs off on that number. Then organic traffic slides for six weeks and nobody can name a cause, because the obvious checks all came back clean.

That count was never the thing worth checking.

The Post Count Matches and That Proves Almost Nothing

A matching count answers one narrow question. Did every record survive the move? Yes. Good.

Google was never reading your records though. It reads pages. And a page is the record plus everything the old system wrapped around it while nobody was watching.

The wrapping is what fails to come across.

Your Export Only Holds What Somebody Typed

Open the export file. Look at what sits inside it. Titles. Body copy. Featured images. Dates, authors, categories.

Notice the shape of that list. Every item on it is something a human once put into a box.

Nobody typed the canonical tag. Or the schema block. The breadcrumb trail appeared on its own. So did the related-posts module under every article.

Those things got generated at page load. Generation leaves nothing behind to export.

The Old CMS Was Writing Half of Every Page

Pull up any article on the old site. View source. Then hunt for that HTML in your export file.

The title tag pattern. Whatever fallback rule wrote your meta descriptions. Heading levels the theme assigned. Pagination links. The image srcset. Hreflang, if you run more than one language.

Roughly half the page. More than half on a template-heavy CMS.

Take a product listing as the clearest case. Words a merchandiser actually typed run to maybe eighty. The rendered page carries breadcrumbs, review markup, stock status, related items, filter links. Google reads every bit of that. Your export holds the eighty words.

The migration tool moves the other half.

Somebody Will Say the Redirects Cover It

Redirects matter. Map every changed URL and honour it properly, because a broken address costs you everything on that page. A pre-launch website audit catches most of the URL side.

But a redirect only preserves an address. It says nothing whatsoever about what gets served at that address.

Point the old URL at the new one and the request arrives fine. Then the new template renders something thinner than the thing that ranked. Same address. Different page. Google agrees with your redirect and disagrees with the result.

Templates Decide Which Words Count as Headings

Here is where a CMS migration gets quiet.

Your old theme rendered the post title as an H1. The new one wraps it in a styled div, since the designer wanted control over typography. Subheadings that were H2s arrive as bold paragraphs.

Nothing got deleted. Every word transferred. The document outline that told Google what the page was about has simply gone flat.

Run a heading check on twenty migrated pages before launch. Twenty is enough to expose a pattern.

The Links That Vanish Were Never in the Body

Count the internal links on your old site. Then count how many of them sit inside the article text itself.

Fewer than you would guess. The rest came from modules. Related posts. Category archives. Tag pages. That sidebar list of recent articles.

A migration carries body links across, because those live in the content field. Module links live in no field at all. They existed as a rule the old CMS ran on every request.

New CMS, different rule. Sometimes no rule at all. Crawl both versions and compare the totals yourself. The gap is rarely small.

Plugins Do Not Migrate. Their Output Was Doing the Ranking.

An SEO plugin keeps its data in its own tables. Titles you rewrote by hand. Descriptions as well. Noindex flags on thin archive pages. Canonical overrides on near-duplicate listings.

Switch platforms and that table stays behind. Then the new system generates its own defaults instead.

Defaults are the problem. A default title is the post title plus the site name. Yours was probably better, since somebody spent an afternoon on it.

Export the plugin data separately. It is a different export from the content one.

Every Date Resets and the Sitemap Announces It

Watch what the migration tool does with timestamps. Plenty of them stamp the import date onto every post. A decade of publishing history collapses into one afternoon.

Then the new sitemap goes out with a lastmod value of today against every URL on the site. Nothing in there was modified today. Google gets told the whole site changed at once.

That claim rarely helps you. A crawler that believes it burns budget re-fetching pages that never moved. One that stops believing it will ignore your lastmod dates later, when a page genuinely does change.

Check the sitemap on day one. Matching dates everywhere means the developer needs a message before Google finishes its first crawl.

Read the Old Source Before You Switch CMS Platforms

Do this while the old site still exists. Afterwards you are working from memory.

Pick your ten best pages by organic traffic. Save the raw HTML of each. Not a screenshot. The source.

Launch, then diff the new HTML against your saved copy. Anything missing was a decision somebody made without knowing they were making it.

Ten pages. One afternoon. It catches most of the damage before Google ever sees it.

How Long a CMS Migration Dip Lasts

Expect movement for four to eight weeks even on a clean migration. Google recrawls at its own pace. Bigger sites take longer, simply because crawl budget spreads thin across them.

A dip that recovers by week eight was ordinary turbulence. One still sitting there in month three is a missing signal, not a patience problem.

Check Search Console coverage weekly through that window. Watch the indexed page count more closely than the traffic line.

Where the CMS Migration Argument Stops

None of this rescues weak content. Thin pages that ranked on an aging domain keep sliding after the move anyway. Then the migration takes the blame for it.

Platform choice also matters less than most vendors suggest. Mainstream systems all rank fine once somebody configures them properly.

And some migrations genuinely improve things. A faster front end helps. So does dropping a plugin stack that was fighting itself.

CMS Migration Work for Businesses Across Calgary

SEO Company To-The-TOP! has watched this pattern repeat since 2007. One specialist handles the work start to finish, so nothing gets handed to a junior halfway through a migration.

Businesses looking at SEO in Calgary usually call To-The-TOP! after the drop rather than before it. Earlier is better. Booking the audit while the old site is still live is what makes the comparison possible at all.

Bring the keyword research file too. Which pages deserve the careful treatment depends on which ones actually earn the traffic.

Nineteen years of this makes the pattern familiar well before a developer quotes the job. To-The-TOP! has watched migrations go smoothly as well. The smooth ones all had somebody recording the old site’s output beforehand, on purpose, while there was still something to record.

Paid search covers the gap in the meantime. A campaign under Google Ads management keeps the phone ringing while organic recovers. To-The-TOP! runs both sides for plenty of clients on exactly that basis. Call (403) 308-5949 before the developer starts, not after.

Common Questions About CMS Migration and SEO

Does changing CMS hurt SEO permanently?

No. Rankings usually settle within two months once the missing signals go back in. Permanent damage comes from leaving them missing, not from the switch itself.

Should I redesign and migrate at the same time?

Separate them if you can. Two variables changing at once turns diagnosis into guesswork. Migrate first. Then redesign a month later.

Will my meta descriptions transfer automatically?

Rarely. They sit in the plugin’s tables, not the content ones. Export them yourself before the switch, or accept whatever defaults the new CMS writes for you.

How many pages should I check after launch?

Start with the top twenty by traffic. Then sample ten more at random from the tail, since templates tend to fail differently on older posts.