SquishPixel Blog · SEO · Front-end · Tools

A 12-point post-launch SEO audit that catches silent site problems

DEV Community published a post by a developer who ran an SEO audit four weeks after launching a site and found every item on the list was broken. The site was Speech Kobo, built with Next.js App Router in four languages. The author lays out the checks in the order they would run them.

The list starts with basics: make /page and /page/ resolve to one URL, settle the www and protocol versions of each address, and confirm that a missing page returns a real 404 status. A page that shows "not found" but returns a success status is a soft 404, which Google treats as thin content. The author says this can hurt the whole site.

Next come robots.txt and the sitemap. The author warns about a staging Disallow: / rule left in a production deploy and says to check it after every deploy. The sitemap should match the real page count, because stale entries make Google waste crawl budget on redirects and 404s. It should also include every language variant.

The author calls orphan pages the check nobody does. These are pages in the sitemap that no internal link reaches. They found twelve, and the fix was adding three index cards. Other items cover hreflang tags, which tell search engines which language version to show. Each must link back reciprocally, and the author advises generating them from data, not by hand. The list also covers self-referencing canonical tags, structured data that matches what users see, rendering without JavaScript, proper 404 handling, and checking environment variables on the live site.

The author's main point is that none of these problems throw an error. The site loads and works, so the damage builds up unseen. They estimate 45 minutes for a systematic run.

Why it matters

Most of these faults cannot be spotted by browsing your own site. A wrong canonical tag, a broken hreflang map or a stray robots.txt rule can keep pages out of search results while everything looks fine. Because the effects compound, small errors left alone get more costly over time.

The post also suggests turning the audit into a repeatable checklist, not a one-off. Several items, such as robots.txt and the sitemap, can change with each deploy, so they need checking again, not once.

Source: DEV Community. This brief summarises their reporting; read the original for the full detail.