A pre-launch technical SEO review for people who want to catch the boring problems before they become expensive ones.
Start with the page a crawler actually receives
Open a few important URLs without relying on the browser cache and check what comes back from the server. A homepage that looks perfect in your browser can still return a redirect loop, a soft 404, an authentication page, or almost no useful HTML to a crawler. I usually test the homepage, one service page, one article, a category or archive, and any page that was rebuilt during the project.
The first pass is intentionally unglamorous: status code, final URL, response headers, canonical, robots directives, and whether the important copy exists in the HTML. If one of those basics is wrong, there is little value in debating keywords yet.
Check crawl rules with the production domain in front of you
Staging rules are a classic launch-day trap. A team blocks a development site with noindex or a broad robots.txt rule, moves the site, and forgets that one setting came along for the ride. Check robots.txt, the meta robots tag, X-Robots-Tag headers, and any CMS-level “discourage search engines” switch.
Do not treat crawling and indexing as the same thing. A URL can be crawlable but noindexed. It can also be blocked in robots.txt before a crawler gets a chance to see the noindex instruction. Write down what you intend for each template, then compare the live configuration with that decision.
Follow redirects and canonicals instead of reading them in isolation
A canonical is useful only when the rest of the URL signals agree with it. If the browser lands on HTTPS, the sitemap lists HTTPS, internal links use HTTPS, but the canonical points at an old HTTP hostname, the page is sending mixed instructions. The same problem happens after domain migrations and redesigns.
Pick the preferred URL and make the path boring: internal links point there, redirects end there, the canonical references it, and the sitemap lists it. One clean hop is easier to reason about than a chain of historical redirects.
Look at templates, not just the homepage
Technical SEO problems are usually template problems. A homepage can pass every check while a product template outputs two H1s, a blog template forgets canonicals, or a filtered archive generates thousands of indexable combinations. Sample each template that can create URLs.
For every sample, check title and description output, headings, indexability, canonical behavior, structured data, pagination, internal links, and the amount of meaningful server-rendered content. This is also where you catch image dimensions, lazy-loading mistakes, and scripts that block rendering.
Use performance data as a debugging tool
A single performance score is not a diagnosis. When a page is slow, identify what is actually costing time: a hero image that is too large, a font that blocks text, a third-party script, a long main-thread task, or layout movement caused by missing dimensions. Fix the cause, not the number.
Test on a representative mobile connection and keep field data separate from lab data. Lab tests are useful for reproducing a problem; field data tells you what visitors have really experienced. Both matter, but they answer different questions.
Keep a launch record so you know what changed
Before launch, save a small baseline: indexable URL count, sitemap URL count, a handful of important rankings, Search Console coverage, key performance metrics, and screenshots of important templates. Then check the same set after deployment.
This sounds bureaucratic until something drops. When it does, a baseline turns “SEO is down” into a concrete investigation: these URLs disappeared from the sitemap, this template started returning noindex, or this page became much heavier after a script was added. That is the kind of evidence that gets fixes shipped quickly.
Before you move on, check these
- Test representative URLs for status codes, redirects and useful server HTML.
- Confirm robots.txt, meta robots and X-Robots-Tag match the intended index policy.
- Make canonicals, internal links, redirects and sitemap URLs agree on one preferred version.
- Audit every URL-producing template, not only the homepage.
- Investigate the cause behind weak performance metrics instead of chasing one aggregate score.
- Save a pre-launch baseline and compare it with the first days after deployment.
Keep reading
Paste a public URL into AEO & GEO Insights and compare the article with what the crawler actually finds: structure, indexing signals, evidence, entities and extractable answers.
