Google Search

Google Search is not a magic black box. It follows three predictable stages: crawling, indexing, and serving. Think of crawling as Google sending a robot to photocopy every page it finds. Indexing is filing that photocopy in a giant cabinet organised by topic. Serving is pulling the right filing card when someone asks a question. If your page is not crawled, it never gets filed. If it is not indexed, it cannot be served. That is the entire pipeline. I used to think publishing a page was enough, but you must ensure each stage passes. The most common failure is at indexing: the page is crawled but blocked by a noindex tag, a login wall, or a JavaScript resource that Google cannot parse. A client's travel blog had all posts tagged noindex by mistake after a CMS update. Took me three hours to find. Another issue: robots.txt disallowing certain URLs. The [SEO checklist](/seo-checklist/) I follow includes verifying the robots.txt file. So the plain-English take: treat each stage as a gate. If the gate is closed, your page stays invisible.

This matters whenever a page you care about is not in Google's results. I run this check weekly for a client in the travel niche. Their new destination guides were not appearing. I used Search Console's URL Inspection tool and saw "Discovered - currently not indexed". That means Google found the page but decided not to file it yet. The root cause was thin content: the guides were under 300 words. Google has a quality threshold. Publishing a page does not guarantee indexing. The decision rule I use: before worrying about keywords or backlinks, confirm the page is indexed. If it is not, fix content depth (aim for 800+ words with original analysis), remove any noindex tags, and ensure internal links point to it. Another case: JavaScript-heavy e-commerce product pages. A client's entire catalogue was built with React. Google could crawl the initial HTML but the product data was loaded via JavaScript after page load. The result: Google saw an empty shell and did not index the product URLs. The fix was server-side rendering or dynamic rendering. So when it matters: anytime you depend on organic traffic for specific pages, and those pages are missing. [Organic SEO](/organic-seo/) practitioners talk about this constantly but many ignore the index check.

In the search results page, Google displays various features: organic links, ads, featured snippets, knowledge panels, images, videos, people also ask, and more. But the crucial point is that every feature is populated from the index. If your page is indexed but not relevant, it may appear for some queries but not for others. I once had a page ranking for a fringe query but not for the main query I wanted. The reason: the page's main heading and first paragraph matched the fringe query exactly, but the main query words were buried in the body. Google uses prominent placements. So showing up is not just about being indexed; it is about signalling relevance in the right places. Also, different features require different markup. For example, a recipe page might show up in a rich result with stars and cooking time if you add structured data. But if the structured data is incorrect, Google may ignore it. My rule: start with the standard organic link, then layer structured data only when the base content is solid. [Semantic search](/semantic-search/) has changed how Google interprets intent, making keyword stuffing less effective.

There is a tension between making pages crawlable and making them fast. Heavy JavaScript frameworks slow down crawling because Google has to render the page multiple times. A client used a JavaScript framework for a blog. The blog posts were not indexed. I had to decide: switch to server-side rendering or accept delayed indexing. We chose static site generation. That made the pages crawlable instantly, but the dynamic features (like live comment counts) were lost. Tradeoff accepted. Another tension: keyword optimisation versus readability. Using exact-match keywords in headings and alt text helps ranking, but can sound robotic. I prefer to write naturally and check that the keyword appears in the first 100 words, the title, and one subheading. If it does, I stop. Over-optimisation can trigger a quality filter. I saw a client's page drop after adding the target keyword to every paragraph. The content became unhelpful. The decision rule: optimise the title, main heading, and opening paragraph; then let the rest be natural. Use synonyms. Also, there is a budget issue: crawl budget. If your site has thousands of low-value pages, Google may not crawl your important pages frequently. Spend your budget on pages that matter. [SEO best practices](/seo-best-practices/) advise focusing on high-quality pages.

I used to believe that submitting a sitemap guaranteed indexing. It does not. A sitemap is a suggestion, not a command. Google may still skip pages it deems low quality or duplicate. I also thought that all JavaScript was rendered equally. It is not. Google's rendering capabilities have improved but still have limits. I wasted weeks building a single-page application for a portfolio site, only to find that Google could not index any of the project pages. I had to rebuild with static generation. The biggest mistake: I assumed that "Google Search" was static. It is a living system. Algorithm updates (like core updates) can change how pages are ranked overnight. I used to optimise for a specific ranking factor, only to have it become less important after an update. I now focus on content quality and user intent. [White hat SEO](/white-hat-seo/) is about playing the long game, not chasing short-term tricks.