A site cleanup that leaves fewer, stronger pages.

A site cleanup starts with a full list of URLs, not a crawl alone. We combine the Screaming Frog crawl, the XML sitemaps, the Search Console page indexing report, analytics landing pages and the store’s own product and collection export. Pages that only appear in one of those sources, such as an old product Google still indexes but nothing links to, are often the ones that need attention.

Each URL then gets data beside it: organic clicks for the last 12 months, external links, orders, word count and its indexing status. With that in place, the decision is mostly made by the numbers. A page with links or sales stays or is merged; a page with no clicks, no links and no orders in 12 months is a candidate for removal. Google’s guide to helpful, people-first content is the test for the pages in between.

Every URL ends up in one of four groups. Keep: the page works. Improve: the page should rank but is thin content, so it gets copy, filters or products added. Merge: two pages do the same job, so one absorbs the other with a 301. Remove: the page has no value and no replacement, so it returns a 404 or 410. Google’s page on HTTP status codes explains how each one is processed, and it is why a removed page should never answer 200 with an empty template.

Duplicate product pages are merged, not deleted. We pick the URL with the most links and sales, move reviews and copy onto it, point the other at it with a 301, and update internal links so the store never links to the old one. Where both versions must stay for shoppers, such as a bundle and a single item, the canonical tag tells Google which one to index.

Discontinued products follow one rule. If a close replacement exists, the old URL is redirected to it. If not, it goes to the parent collection when that collection still sells something similar. If neither applies, the page returns a 404 with links back into the store. Redirecting everything to the home page is the one option we never use, because Google tends to treat that as a soft 404 and shoppers find it confusing.

Redirect chains get flattened. We follow every redirect from the full URL list and rewrite each one to point straight at its final page, one hop only. Google’s page on redirects covers the types to use; its crawlers follow up to 10 hops, but shoppers and other search engines are less patient. On Shopify, as of October 2026, a store can hold up to 100,000 URL redirects (20 million on Plus) and redirects only work when the old URL returns a 404, according to the Shopify Help Center. So we check limits and hidden products before importing anything.

We did the same job on this site. Before the rebuild launched, every old aboveclicks.com URL was mapped: 13 redirects, 4 rules returning 410 for old files with no value, and the old soft 404 page replaced by a real 404. Kashif runs the same crawl and redirect checks on store SEO work, and the Search Console results on our jewelry page show what a cleaner store can earn over time.

Last, the sitemap and internal links are cleaned up to match. Sitemaps list only the URLs that should be indexed, menus and footers link to the final URLs, and we re-crawl to confirm there are no chains, no internal links to redirects and no 200 pages left empty. Search Console is checked again after 4 and 8 weeks.