The SEO Migration Trap: Why Magento Stores Lose Their Best-Performing Category Pages After a Relaunch
by Rahul Basu
October 15,2026
Picture a garden furniture retailer relaunching on Magento. The new store goes live on a Thursday night. By Monday it all looks fine. Orders are coming through, checkout behaves, and the design team is quietly pleased with itself. Three weeks on, the e-commerce director spots something that was never on the launch checklist. Organic revenue has slipped, and not evenly. A handful of category pages that used to bring in a big slice of search-driven sales have gone quiet.
Nobody broke anything. The redirects are switched on and no old URL returns a 404. Yet Garden Furniture Sets, which sat third for its main term for years, isn't on the first page at all.
That's the SEO migration trap, and Magento stores walk into it more often than they should. This article covers why category pages are the thing at risk, how the damage happens, and what a Magento agency should have settled before the first template goes live. Replatform, upgrade or just reshuffle the catalogue, and the same rule holds. Magento migration SEO is mostly decided on paper, in the weeks before launch.
Look at rows three and six in particular. Merging bistro and balcony sets is defensible only if the new page's copy covers both searches, and the filtered teak URL needs a rule of its own, because a standard category rewrite would lose the parameter.
Then the acceptance criteria, which are what stop a map from being decoration. Agree them up front. For every Tier 1 URL, the old address answers with a single 301 hop to the mapped destination. The destination returns 200, carries a self-referencing canonical and no noindex, sits in the new XML sitemap, and has a title, H1 and intro copy that either match the old page or have been deliberately improved. Menus and footers link to the new addresses directly. Store views carry correct hreflang if you run more than one language. Test it all on staging by crawling the old URL list against the new build. Ingold's Magento migration process starts with an analysis of the existing shop and gives clients a staging version to test before launch, and the map is what makes that testing measurable. We've also put the full checklist into a downloadable SEO migration risk register, with each risk scored, owned and tied to an acceptance test.
Finally, version the map and keep it. It's the document you'll reach for six weeks later, when a category is down and somebody asks what happened to it.
Set review dates before launch rather than improvising them afterwards. Weekly for the first month is a sensible rhythm, with the Tier 1 list in front of everyone and each review run as a small SEO migration audit against the same map. If one group is still falling while the rest recover, escalate it then, not at the quarterly review.
And write down what "normal" means for your store in advance. Agree with finance, or with the e-commerce director, how large a dip in organic revenue is tolerable and for how long, before the migration gets blamed or cleared.
Why Do Category Pages Often Carry More Commercial Value Than Individual Products?
Think about how people actually shop. Almost nobody types a product code into Google unless they already own one. They type "rattan garden dining set" or "outdoor lighting", a phrase that means I want to buy something like this and I'd like to see what's out there. The page built to answer that is a category page, and it's the one that tends to rank. Product pages catch the narrow stuff: a model name, a specific size, a brand plus a number. Useful, but small, and unstable. Products sell out, get discontinued and are replaced by next season's version. Category URLs hang around for years, collecting internal links, external links and a long history of clicks, and that accumulated weight is hard to rebuild once it's gone. There's a structural point too. A category page is the shelf that feeds everything on it. Products get found and linked through listings and pagination, so when a category slips down the rankings, every product inside loses part of its way in. The damage spreads wider than the one URL that moved. For anyone reporting on revenue, this changes what to watch. A top category isn't a page, it's a revenue line. Treat it that way and Magento category page SEO stops being a copywriting chore and becomes risk management: which listings earn the money, what they depend on, and what a migration might quietly take away.What Goes Wrong When the Category Structure Changes?
Three things tend to happen, and they often happen together. The first is the paths. Magento builds category URLs from each category's URL key and its place in the tree, plus a suffix such as .html if one is set. Rebuild the tree, rename a key, drop the suffix, or arrive from a platform with different habits, and nearly every category address moves. Say /garden/furniture/dining-sets.html becomes /garden-furniture/dining-sets/. One URL, and nobody loses sleep. Multiply it by a few hundred categories and it's a project. The second is consolidation. The instinct is usually sound. A catalogue that has grown organically ends up with overlapping categories, and a cleaner structure helps shoppers. Trouble starts when the tidy-up folds together pages that were each ranking for something different. Bistro sets and balcony sets are separate searches. Merge them into one "small space" page and the new URL may inherit the old authority, but it no longer matches either query as precisely, and the long-tail rankings drift away. Google's guidance does allow consolidated content to redirect to a single new page. It works when that page really answers what the old ones did. The third is lost landing pages. Seasonal categories, old campaign pages, brand pages, categories that were once hidden from the menu but kept ranking. They rarely appear in a crawl of the live navigation, so nobody maps them, and on launch day they return a 404 or land on the home page. This is why a crawl alone makes a poor inventory. Google's site move documentation suggests gathering URLs from sitemaps, traffic data, server logs, Search Console and the CMS, and the reason is simple: each source finds things the others miss. A quieter version of the same problem lives in the navigation. A new mega-menu that drops the middle layer of categories removes internal links that were holding those pages up. Nothing is redirected wrongly, because nothing is redirected at all. The pages still exist, they're just no longer two clicks from the home page, and rankings follow.Why Are 301 Redirects Alone Not Enough?
Redirects are the minimum. They stop a link from dying, and they're the part everybody checks, which is exactly why a project can pass the redirect check and still lose rankings. Magento 301 redirects answer one question: where should this old address send people? They leave four others unanswered. Start with the mechanics anyway, since they go wrong more often than you'd expect. Adobe's documentation for category URL rewrites lets you choose between a temporary 302 and a permanent 301, and says search engines generally don't retain page rank through the temporary one. It also notes that query parameters are dropped during these redirects, which matters when we get to filters. Automatic redirects help with everyday URL key edits, but they won't help when the URLs changed on another platform or the whole structure was rebuilt from scratch. And Google's guidance on redirects points the same way as its site move advice: send people straight to the final destination. Googlebot will follow a long chain, but Google suggests keeping it to three hops at most and always under five. Old rewrites left behind by earlier migrations are the usual way chains appear. Canonicals come first among the four. The new page has to declare itself the preferred version of itself. Adobe's SEO settings include "Use Canonical Link Meta Tag for Categories", and a canonical left pointing at the old URL, the staging domain or a filtered variant is an easy mistake to ship. Google's own move checklist says to confirm that canonicals and noindex rules point at the new URLs when the redirects go live. A noindex copied over from staging is the classic way to make a whole section disappear. Then internal links. If the menu, the footer and the body copy still link to old addresses, every click passes through a redirect, and Google recommends updating links so the site isn't leaning on redirects at all. It's a signal problem as well as a speed one. The pages you most want to rank are being linked to through addresses you've told Google are retired. Then relevance. Rankings follow content. If the old page ranked with a decent intro, a descriptive H1 and a sensible title, and the new template drops all three, a perfectly executed 301 sends visitors to a thinner page. Google also advises against redirecting lots of pages to one unrelated destination, the home page being the obvious example. A redirect is a claim that the new page is the successor. The content has to back that claim up. Last, destinations. A retired category should go to its closest genuine equivalent, which is sometimes a parent and sometimes a different category altogether. Decide it page by page, record why, and have someone sign it off. And keep the redirects. Google's advice is to leave them in place for as long as possible, generally at least a year. If old URLs still have links pointing at them, keep them well beyond that.How Should Category URLs Be Prioritised Before Migration?
A mid-sized store can carry thousands of category and filter URLs. You can't give them all the same care, and you shouldn't try. Rank them by what they're worth. Four inputs do most of the work. Organic revenue per landing page over the last twelve months, which is long enough to smooth out seasonality. Organic sessions, which show scale even where revenue attribution is patchy. Backlinks, meaning the referring domains pointing at each URL, because those links are the hardest asset to replace. And conversions, to catch pages that draw less traffic but turn it into orders at a good rate. Pull them from analytics, Search Console and a backlink tool, join them on the URL, and sort. Then draw lines. Tier 1 is the short list that carries most of the organic revenue. Those URLs get mapped by hand, compared page against page before and after, and signed off one at a time. Tier 2 is the middle, mapped with rules and then spot-checked. Tier 3 is the long tail, handled by patterns, with sampling to catch mistakes. The cut-offs depend on the catalogue. What matters is that somebody chose them, in writing, before launch. A few cases deserve a flag. URLs with strong backlinks but little traffic are worth protecting for the links alone. URLs sitting on page two of the results are a quiet opportunity, since a relaunch can push them up if the new page is better or knock them out if it's worse. And if a Tier 1 URL doesn't have to change, leave it alone. Keeping an address as it is beats any redirect.What Happens to Filters and Faceted Navigation?
Filters are where a migration quietly changes how much of your site Google can actually get through. Layered navigation can turn one category into hundreds of URLs: colour, size, material, price band, in every combination and in any order. Google's documentation on faceted navigation describes the problem plainly. Facets can produce a near-infinite number of URLs, crawlers can't tell which ones are worth fetching until they've fetched them, and the result is wasted crawling and slower discovery of the pages that matter. That's already true on a stable site. During a move it gets worse, because Googlebot has to visit every old and new URL at least once to complete the move, and a thicket of filter URLs stands between the crawler and the pages you care about. Relaunches tend to go wrong in three ways here. Filters that were blocked on the old store become crawlable on the new one, and the index fills with near-duplicates. Filters that ranked, and earned money, get blocked or set to noindex because the new rules were applied bluntly. And old filter URLs with links or rankings get redirected without their parameters, since Magento's rewrite redirects drop query strings. The first is a duplicate content mess. The other two lose revenue. The fix is a decision, made before launch. Pull search demand for filter combinations from Search Console. Where people clearly search for "teak garden furniture", give it a proper landing page with a clean URL, its own title and its own copy, and map the old filter URL to it. For everything else, decide how to keep it out. Google says robots.txt is the stronger way to stop facet crawling, with canonical tags and nofollow acting more slowly and less reliably, so don't lean on canonicals alone if crawl waste is the worry. If you'd rather keep facet URLs open, the practical advice is to use the standard ampersand separator, keep filter order consistent, and return a 404 for combinations that don't exist. Pagination belongs in this conversation too. Deeper pages in a category shouldn't all point their canonical at page one, or products listed only on those pages can lose their route to being found. Check what the new theme outputs against the canonical settings you intended.How Should a Magento Agency Build an SEO Migration Map?
The map is the spine of the whole exercise. It's a spreadsheet, usually a dull one, and it decides whether a launch is safe. It's also the working core of any SEO migration audit, the thing every other check gets measured against. Any Magento agency should be able to show you theirs before development starts, not hand it over in the final week. Start with the inventory. Combine a crawl of the live site, the XML sitemaps, Search Console, analytics landing pages, server logs and backlink data. Then export Magento's own url_rewrite table, because it holds every redirect the store has collected, including the ones earlier migrations left behind. Those are the redirects that turn into chains if nobody looks. Each row needs the old URL, its tier and its numbers, the new URL, the redirect type, and a match type that records what kind of decision was made: exact match, merged, parent fallback or retired. A few rules help. One-to-one wherever the page has a true equivalent. Merges only with a reason written down. Parent fallbacks need approval from whoever owns the revenue. Pages being retired should return a proper 404 or 410, or lead to a useful alternative, and never to the home page by default. Table 1 shows the idea for a fictional garden furniture retailer. It's an illustration, not client data, but each row stands for a decision that real migrations have to make.| Old URL | New URL | Handling | Reason |
| /garden/furniture/dining-sets.html | /garden-furniture/dining-sets/ | Exact match, 301 (Tier 1) | Top organic revenue. Same intent, so title, H1 and intro copy carried across. |
| /garden/furniture/rattan-sets.html | /garden-furniture/rattan-sets/ | Exact match, 301 (Tier 1) | Strong backlinks. Direct single hop, no intermediate redirect. |
| /garden/furniture/bistro-sets.html and /garden/furniture/balcony-sets.html | /garden-furniture/small-space-sets/ | Merged, 301 (Tier 2) | Two thin categories became one. New page copy covers both searches. |
| /garden/lighting.html | /outdoor-lighting/ | Renamed, 301 (Tier 2) | Same category under a better key. Internal links updated to the new address. |
| /summer-sale-2023.html | /sale/ | Parent fallback, 301 (Tier 3) | No equivalent page, but 14 referring domains. Closest hub chosen, not the home page. |
| /garden/furniture.html?material=teak | /garden-furniture/teak/ | Filter promoted to landing page (Tier 2) | Real search demand. Needs its own server-level rule, because a category rewrite drops the parameter. |
| /garden/furniture.html?price=100-200 | /garden-furniture/ | Retired filter, 301 to parent (Tier 3) | No search demand. Blocked from crawling on the new build. |
| /catalog/category/view/id/42 | /garden-furniture/ | System path, 301 (Tier 3) | Legacy Magento route still receiving hits in the logs. |
Which Metrics Reveal a Genuine Post-Launch SEO Problem?
Start with the number the board cares about and work backwards. Organic revenue by category landing page is the headline measure, compared against the same weeks last year so seasonality doesn't mislead you. Group old and new URLs using the migration map, so a moved page is compared with itself and not with an empty row. It also needs tracking that survived the move, which is worth checking in week one, because broken analytics looks exactly like lost traffic. Clicks and impressions in Search Console come next. Falling impressions with steady click-through points at rankings or indexing. Steady impressions with falling click-through points at the listing itself, usually a changed title or meta description. The Performance report can filter by page, so feed it the Tier 1 list. Indexing tells you whether Google has accepted the new pages. Google's site move guidance says the old sitemap should gradually show fewer indexed pages as the new one shows more, which gives you a clean thing to watch. In the Page indexing report, look at Tier 1 URLs for statuses like "Crawled - currently not indexed", "Alternate page with proper canonical tag", "Excluded by 'noindex' tag" and "Blocked by robots.txt". Any of them against a page that should be live and canonical is a ticket, not a trend. Rankings are useful once there's a baseline. Track a fixed set of money terms for each Tier 1 category for a few weeks before launch, so you have something to compare with. Without it, a post-launch rank report is a snapshot with no context. Server logs and the Crawl stats report fill in the rest. You want Googlebot hitting old URLs, getting a single 301 and landing on a 200, not wandering through filter URLs or stalling on 5xx errors. Google also expects crawling to pick up after a move, so check the server can cope.How Can Retailers Distinguish Normal Fluctuations From Structural Damage?
Some wobble is expected. Google's guidance says outright that rankings and visibility can fluctuate during a move, and that for a medium-sized site it can take a few weeks or more for new URLs to replace the old ones in results, longer for bigger sites. A dip in week one proves nothing. The useful question is what shape it has. Normal turbulence is broad and gentle. Everything dips a little, in the same direction, and then the line starts to turn. The old URLs fade from the index while the new ones come in. The queries bringing traffic stay roughly the same. Structural damage is lumpy. A few URL groups fall hard while others hold. The line doesn't turn. A top category's impressions drop to almost nothing and stay there, or the old URLs vanish before the new ones are indexed. The query mix shifts, because the page now matches something different from before. That lumpiness is why the migration map earns its keep: it shows which group a problem belongs to, so you can go straight to the likely cause. Table 2 lists some common patterns.| What you see | Likely cause | Where to look |
| Impressions for one Tier 1 category fall sharply and don't recover | Wrong redirect destination, or the new page doesn't match the old content | Map row, destination copy, redirect trace |
| New URLs flagged as alternate pages with a proper canonical | Canonical still pointing at the old URL or a variant | Canonical tag on the new page |
| Whole sections drop out at once | Noindex or robots rule carried over from staging | robots.txt, meta robots, response headers |
| Impressions steady, clicks falling | Titles or meta descriptions changed | Old and new title and snippet comparison |
| Heavy crawling, key pages visited rarely | Filter or parameter URLs absorbing the crawl | Server logs, Crawl stats report |
| Gentle, all-round dip that's easing | Normal reindexing | Keep watching weekly |
