Search traffic falling after a redesign is common enough that people treat it as inevitable. It is not. It is almost always caused by the same handful of omissions, and all of them are preventable with work done before the new site is built.
Crawl the old site before you change anything
You cannot preserve what you have not recorded. Before any development starts, produce a complete list of existing URLs, and for each one: its title, its meta description, whether it currently receives traffic, and whether anything links to it from outside.
Search Console and analytics together give you the traffic and link picture. A crawler gives you the full URL inventory including the pages nobody remembers — and those forgotten pages are frequently the ones with external links pointing at them.
Keep URLs identical wherever you can
The cheapest way to preserve ranking is to not change the address. A redesign is a visual and structural exercise; it does not require new URLs.
Teams change them anyway, usually because a new CMS has different conventions, or because someone wants a tidier structure. Neither is worth the traffic. If a URL currently ranks, keep it.
Where URLs must change, map every one
A redirect map is a spreadsheet: old URL, new URL, one row per page. Rules that cover whole sections are fine where the structure genuinely maps; the exceptions need to be listed individually.
Three rules that matter:
- 301, not 302. A permanent redirect passes ranking signals; a temporary one is treated as a detour.
- One hop. Old URL straight to final destination. Chains of redirects dilute the signal and slow the page.
- Not everything to the homepage. Redirecting a hundred retired pages to the homepage is treated as a soft 404 and preserves nothing. Send each to the closest equivalent page, or let it 404 honestly if there genuinely is no equivalent.
Carry the metadata across
Titles and meta descriptions on the old site were often written deliberately. A new CMS generating them from a template will overwrite that work silently.
Also carry structured data, image alt text, and canonical tags. Canonicals especially: a new site that canonicalises every page to the homepage will remove itself from the index entirely, and it is a surprisingly easy mistake to make — one line in a shared layout does it.
The launch checklist that catches the real failures
- robots.txt does not block the site. Staging usually blocks crawlers; shipping that file to production is the classic redesign disaster.
- No stray noindex tags left over from staging.
- Canonical tags point at the page itself, not the homepage.
- Sitemap regenerated with the new URLs and submitted.
- A sample of redirects tested manually — every top-traffic URL, by hand, verifying a single 301 to a live page.
- Analytics and Search Console still installed on the new site.
What to watch afterwards
Expect a small dip for a week or two while search engines recrawl. That is normal. What is not normal is a sustained decline, or pages dropping out of the index entirely.
Check indexed page count and crawl errors weekly for the first month. Most redesign traffic losses are recoverable if caught early — and almost all of them are a redirect that was missed or a robots directive that should not have shipped.
About the author. Kislay Chandan has worked with schools since 2014. Rather than gathering requirements from the outside, he worked inside a school to see what actually goes wrong day to day, and built what he found into the software. He still works with Shemrock Glitters in Patna — most changes to the product start with something that came up there.
Gallery
