Discovery & crawling
Discovery means finding a possible URL, for example through a link or sitemap. Crawling is a crawler's visit to that URL, if it can and chooses to visit. Discovery does not mean that a page was crawled.
CASHIN INSIGHTS / SEO
When people search for a service like yours, your website should make it easy to understand what you offer and find the right page. This guide helps you check whether your pages answer customer questions, what to improve first and how to follow progress.
SEO can improve relevance and visibility over time; it does not promise a first position or a particular amount of traffic. Google does not guarantee crawling, indexing, rankings or citations in search features. Quality, clarity, accessibility and measurement are sound foundations, not a performance guarantee.
Discovery means finding a possible URL, for example through a link or sitemap. Crawling is a crawler's visit to that URL, if it can and chooses to visit. Discovery does not mean that a page was crawled.
After crawling, Google processes page content and signals and decides whether and how to include it in the index. A crawled page may not be indexed; inclusion is not guaranteed.
For a query, systems select and serve results based on the query, context and relevance. Being indexed does not guarantee an impression or a particular position, and results may differ between searches.
Before writing, state in plain language what a person is trying to do: learn, compare, find a provider, understand cost or get in touch. A search for “small business website design” expresses a different need from “how much does a business website cost?” A service page, guide, comparison and pricing page answer different jobs, even when some details overlap.
Give each page one primary purpose and a defined audience. If questions need separate answers, use distinct connected pages; related details can share sections.
Good boundaries prevent duplication and internal competition. A page that tries to be a guide, sales pitch, price list and glossary may serve none well. Decide what it answers.
A useful page starts with a direct answer, not a long introduction that hides the point. Explain context, options and the conditions that affect a decision. Use descriptive headings and relevant examples; explain jargon so readers can follow.
Originality means adding value: analysis, business data, original images, service experience or a useful local answer, not copying search results. Distinguish what is known from what depends on scope; never invent expertise, outcomes, figures or promises.
AI-assisted content still needs editorial review: verify facts, adapt it to readers and add human context where useful. Similar pages without reader value are not a strategy. Update material when services, prices or rules change; do not change dates just for appearance.
Organize the site around services, audiences and resources. A simple hierarchy helps people and crawlers find key pages.
A useful internal link is real HTML with descriptive text, such as “Guide to choosing a business website” rather than “click here”. Place links in context.
Map important pages and their relationships: a service page can link to a guide explaining the process, and the guide can link back when that next step fits. Broken links, orphan pages and redirect chains create friction. Redirect moved content to its relevant destination, not automatically to the homepage.
Important content should be available in the server's initial response, or remain reliably accessible after rendering. Inspect HTML without running JavaScript, especially for the title, content, links and language. Client rendering can work, but do not serve only a holding message or tiny summary when the article belongs on the page.
Check the HTTP status for every URL. A live page should return a successful response; a removed page with no suitable replacement should return a real 404 or 410, not an error-looking page with status 200. Redirects should lead to relevant destinations.
Review robots.txt and robots directives in both page markup and response headers. A crawl block can prevent access to content; noindex requests exclusion from the index, but a crawler must be able to see the directive. Check that service pages are not accidentally blocked or noindexed. Crawl blocking and index exclusion differ.
Set a consistent canonical to identify the preferred version when alternate URLs or parameters exist. Canonical is a signal, not an absolute command; keep links and sitemap consistent with it. An XML sitemap lists URLs you want discovered; it does not guarantee crawling or indexing. Include important canonical URLs and remove errors or blocked URLs.
Write a concise, unique title and meta description for each page that describes its content without overstating it. Search engines may choose a different title or snippet. A clear H1 and organized subheadings improve hierarchy and reading; there is no need to force a keyword into every heading.
Use relevant images at sensible file sizes. Alt text describes an image when its visual information matters; it is not a place for a keyword list. A decorative image can use empty alt text so assistive technology can skip it. Set dimensions or an aspect ratio to avoid layout shifts, and maintain readable contrast, text size and comfortable tap targets.
Test reading, navigation and forms on a phone; popups and tiny buttons can hurt usability. Accessibility and usability help visitors complete their task.
Core Web Vitals are field measures of user experience: LCP measures when the largest content appears, INP measures responsiveness to interactions, and CLS measures visual stability. Google's thresholds for a good experience are LCP at or below 2.5 seconds, INP at or below 200 milliseconds and CLS at or below 0.1, measured at the 75th percentile of visits. These are experience targets, not ranking promises.
Diagnose before guessing. Field data reflects users when enough data exists; lab tools help reproduce issues. Large images, slow servers, third-party code and unreserved media space can contribute. Fix bottlenecks and test again.
Do not chase a perfect score on every tool. Prioritise fast main content, responsive controls and stable reading. Test key mobile pages and consider maintenance before adding scripts.
For a local business, explain where it operates, available services and how customers can get in touch. Keep address, hours and contact details accurate across relevant profiles, including an eligible Business Profile. Do not list false addresses or service areas.
Local pages need distinct information, not repeated city templates; reviews should be authentic.
Local SEO is unnecessary for some businesses; match it to how customers choose: proximity, availability, specialization or delivery.
Google Search Console lets verified property owners review search performance, queries, impressions, clicks, URLs and indexing diagnostics available for that property. This does not mean anyone has checked this particular site or that a particular page is indexed; those facts must be verified in the relevant account at the relevant time. A Google site: query is not a substitute for an indexing report.
Start with questions: which pages should generate inquiries? Are they receiving impressions? Does the result describe them well? Do visitors take the next step? Analytics can examine configured, permitted journeys and conversions. Compare periods and account for seasonality.
Measure inquiry quality as well as visits. A submitted form is not necessarily a customer, and a click is not a sale. Define events in a privacy-respecting way, confirm tracking works, and use the data to choose one experiment or fix. Record each change and expected effect to learn from tests.
Google's official guidance for AI features in Search says no special AI-only schema or llms.txt file is required for inclusion. The same SEO fundamentals apply: people-first useful content, crawlable links, a page eligible for indexing and sound technical consistency. Do not publish duplicate formats or write specifically for a model.
AI features may show links to sources in some cases, but there is no guarantee that content will be included, cited or receive traffic. Answer questions well, provide original details and sources. Do not forecast on assumed citations or undocumented tactics.
There is no fixed timeline. Discovery, crawling, indexing, competition and site changes affect the pace, and business traffic may follow later. Track trends over time and be wary of promises of a specific date or position.
No. Indexing is appropriate for useful pages you want people to find. Thank-you pages, duplicates, internal search results and private content usually do not need to be indexed. Decide based on each page's purpose and use canonical or noindex where appropriate.
No. A sitemap helps search engines discover URLs, but does not guarantee crawling, indexing or ranking. Internal links, content, technical eligibility and search-engine decisions still matter.
Google does not require AI-specific schema or llms.txt for its AI Search features. Follow existing SEO fundamentals and make the page eligible for indexing; inclusion and citations are not guaranteed.
No. LCP at or below 2.5 seconds, INP at or below 200 milliseconds and CLS at or below 0.1 are good-experience targets, not ranking promises. Improve them where that benefits users, alongside content and accessibility.
A redirect sends a browser and visitor to another URL; canonical signals a preferred version among similar URLs. Search engines may consider other signals, so keep them consistent and do not treat canonical as an absolute command.
This guide is general professional information based on the linked Google Search Central resources. It is not a substitute for a site audit, account data or tailored professional advice.