Tag: technical-seo

  • Internal linking as a system, not an afterthought

    Internal linking as a system, not an afterthought

    Most sites I look at don’t have a keyword problem. They have a navigation problem.

    Important pages sit two or three clicks farther from the homepage than they should. Related articles never mention each other. Category hubs are thin. The only way a crawler (or a brand-new reader) finds a useful piece is through the sitemap or the search box. Then the team publishes more, and wonders why none of it builds on itself.

    Sound familiar?

    Internal linking is how you tell people and search engines which pages matter, how your topics connect, and which URL is the main source for a subject. Even Google’s guidance on AI features lists making content easy to find through internal links as one of the fundamentals that still apply. Not because links are trendy. Because a page nobody can discover can’t be indexed, ranked, or cited.

    So in this post I’m going to treat internal linking as a planned system. You’ll see which pages deserve the most links, how I write anchors, how to keep crawl paths short, and how to keep the whole thing from rotting as your site grows.

    What an internal linking system is for

    A good internal linking system does three jobs at once.

    It gives crawlers a path. Google finds pages by following links. When I run a site audit, I almost always find “orphan” URLs that exist in the sitemap or in analytics but aren’t linked from a single indexable page. If you want a page in the index, someone already on your site should be able to click their way to it.

    It tells everyone what a page means. When several related pages link to one URL with descriptive anchors, you’re saying, “this is the page we mean when we talk about this topic.” That helps traditional rankings. It also helps retrieval systems that want one clear, specific source instead of five posts that all sort of say the same thing.

    It helps the reader finish what they came to do. Someone who showed up for a definition might need a process next. Or a local page. Or your service page. Those paths should be written into the content on purpose, not left to a “related posts” widget that’s basically guessing.

    Here’s my honest take. If your site has been publishing for months with little to show for it, look at your links before you look at your content calendar. A lot of the time the missing work is already on the site. It’s just hard to reach. (I come back to this in what a real site health audit covers.)

    Start with a short map, not a plugin

    You don’t need a 400-row spreadsheet to get going. You need a map of the pages that should be collecting authority, and an honest look at what points to them right now.

    Here’s how I build that map:

    • Pick your 10 to 20 workhorse URLs. Service pages, key location pages, pillar articles, and the product or category pages that need to be found. These should get the most thought-through internal links on the site.
    • Check what already points to each one. Note which pages link to it, whether it’s in the main navigation, and which obviously related articles don’t mention it yet.
    • Crawl the site. Flag orphans, pages you can only reach from the footer, and paths that take four or more hops from the homepage to a money page.
    • Group content by topic, not by publish date. A post about title tags should point to a post about meta descriptions, and both should point to the on-page or services page that’s the main offer. “Previous post / next post” links based on dates aren’t a system.
    • Pick one main URL per topic. If two articles are competing for the same job, merge them. Or make the relationship obvious: one is the overview, the other is a narrower follow-up, and the links say so.

    Navigation and breadcrumbs matter. They’re just not enough on their own. Your main nav can only hold so many items before it stops being useful.

    The links that usually move things? The ones inside the body of relevant pages, sitting right where a reader would actually want to click.

    Write anchors that describe the destination

    “Click here.” “Read more.” A naked URL.

    All three waste the link.

    Your anchor text is a tiny description of where you’re sending someone. “SEO and AI search services” tells a person and a crawler a lot more than “learn more.” “How to write title tags that Google will actually use” beats “this post” every time.

    The rules I follow:

    • Use the words a knowledgeable reader would use to describe the destination. Not some stuffed variation of a keyword you’re trying to force in.
    • Link from a sentence that already makes a real point. The link should feel like the natural next step, not decoration.
    • One clear link to the right URL beats three repetitive links to the same page in one article.
    • Read the paragraph out loud. If the anchor sounds weird, it’s forced. Rewrite it.

    Keep crawl paths short, and keep important pages indexable

    Links only help if they can be followed.

    A page behind a login, blocked in robots.txt, tagged noindex, or buried inside a JavaScript interface that never renders links in the HTML isn’t really part of your system. It’s just sitting there.

    These are the patterns I watch for:

    • Orphans. Give every orphan at least one editorial link from a related, indexable page, and add it to the right hub or category.
    • Footer-only links. A footer link can help discovery. But a page that matters should also get links from contextual content, and from the navigation if it belongs there.
    • Facets and parameters. Ecommerce and directory sites can spin up thousands of filtered URLs. Link to the canonical, useful versions. Don’t let every sort order and checkbox combo turn into an internal destination.
    • Redirects and chains. Point internal links at the final 200 URL, not an old slug that hops twice before landing. It’s basic hygiene, and it keeps you from leaking whatever meaning the link was supposed to carry.
    • Pagination and archives. Category pages and date archives can make good hubs if someone actually curates them. They shouldn’t be the only road to your best work.

    People sometimes ask me for the official “maximum clicks from the homepage” rule. I don’t think there’s one worth treating as law. My test is simpler: if someone who already knows your site still can’t find the page without using search, the path is too long or too buried.

    Maintain the system as you publish

    Internal linking breaks down the moment you treat it as a one-time project.

    Every new page should get links from the pages already in its topic group. And it should link out to the hub or service page it supports. If you can’t name those connections when you hit publish, I’d ask whether the new page actually has a job.

    For most sites, a light routine is plenty:

    • When you publish, add two or three contextual links from older, related pages. Not a sitewide blast.
    • Once a quarter, recrawl your important URLs. Look for new orphans, broken internal links, and hubs that have gone stale.
    • When you update or merge a page, fix the anchors that pointed to the old version in the same sitting. Not three months later.
    • Keep an eye on automated “related article” modules. They can help, but they usually link by tag or recency, not by meaning. Use them as a backup, and check they aren’t pointing at thin or off-topic pages.

    None of this is glamorous.

    It’s also how a site starts to feel like a body of knowledge instead of a pile of posts.

    What not to do

    Don’t install a plugin that stuffs the same exact-match anchor into every post in a category. It looks like manipulation, and it’s almost always worse than fewer, better links.

    Don’t build an “HTML sitemap” page as a stand-in for editorial links and then leave the rest of the site disconnected. A list of URLs isn’t a reading path.

    Don’t nofollow your own internal links to try to “sculpt” PageRank. You’re far more likely to confuse the crawl than outsmart it.

    And please don’t link every mention of a word to the same sales page no matter the context. Readers notice. It makes the whole site feel cheaper.

    How this work shows up in a real engagement

    When I work on internal linking at Seopya, it’s usually part of the technical and on-page work, not a standalone “link install.” After the crawl, I map the hubs, fix orphans and chains, and add contextual links where they genuinely help a reader get where they’re going.

    That all sits inside our SEO and AI search services. A site people can’t navigate is a site that’s hard to rank or cite with any confidence.

    You should be able to see the result. Important URLs are easy to reach. Related pages acknowledge each other. And a new article has a home on the day it goes live.

    Key takeaways

    • Internal links are how people and crawlers find your pages, and how your site signals which URL owns a topic.
    • Start with a short list of important pages, then add links from related content. Navigation alone won’t cut it.
    • Write anchors in plain language that describe where the link goes.
    • Short paths, final 200 URLs, no orphaned money pages.
    • Every time you publish, merge, or update, touch the links too.

    FAQ

    How many internal links should a page have?

    There’s no magic number. Link when there’s a real next step for the reader or a real relationship between topics. A long article can carry more links than a short one. A short page stuffed with links is usually dodging clearer writing.

    Do links in the footer or sidebar count?

    They can help with discovery, especially for utility pages. But they’re weaker than a contextual link in the body of a related article, and they shouldn’t be the only links to a page you care about.

    Does internal linking help with AI Overviews or LLM citations?

    Indirectly, yes. Google includes internal linking among the fundamentals that still support AI features in Search, because a page has to be discoverable and indexable to be eligible at all.

    Will internal links “insert” you into an AI answer on their own? No. Nothing does that.

    The next step

    Want a simple place to start? Pick ten important URLs. Crawl to see how they’re reached today. Then add links to them from the related pages that should’ve been pointing there all along. Fix any broken or redirected internal links on those paths while you’re at it.

    That’s an afternoon of work, and it often does more than another blog post would.

    If you’d rather have help building the system on your site, get in touch. I’m glad to walk through what a crawl would show and which links would actually be worth adding.

  • What a real SEO site health audit actually covers

    What a real SEO site health audit actually covers

    Here’s a pattern I see all the time. A business publishes a blog post every week for a year. They pick up a few links. They tweak their homepage copy twice. And traffic just… sits there.

    Usually the problem isn’t the content.

    It’s something underneath it. Pages Google never indexed. A redirect chain that bleeds every internal link. A template that crawls along on phones. Three versions of the same URL quietly fighting each other. You can publish on top of that for a long time and never see anything compound.

    That’s why every serious project at Seopya starts with a site health audit. I don’t mean a one-click “score out of 100.” I also don’t mean a fifty-page PDF of warnings that nobody opens twice. I mean a structured look at whether search engines can find, crawl, render, and trust the site you already have.

    And yes, this matters for AI search now too. Google’s documentation on AI features says a page has to be indexed and eligible for a snippet before it can show up as a supporting link in AI Overviews or AI Mode. Other answer engines can only cite pages they can actually reach and read.

    So let me walk you through what a real audit covers, in the order I actually do it, and how to turn the findings into a short list of fixes instead of a pile of noise.

    What a site health audit is for

    A site health audit answers one question: what is stopping good pages from being found and understood?

    Everything else is secondary.

    Google breaks its own process into three stages (crawling, indexing, and serving results) in its guide to how Search works. I follow that same path in an audit, because a problem early in the chain caps everything after it. A beautifully written page that can’t be crawled won’t rank. It won’t get quoted in an AI answer either. It’s invisible, no matter how good the writing is.

    When I’m done, you should walk away with three things:

    • A clear picture of which pages search engines can reach, which ones they index, and which ones they ignore.
    • A short, prioritized list of technical problems, ranked by how much they hurt important pages (not by how many URLs a tool happened to flag).
    • A baseline you can measure the next six to eight months of work against.

    That’s it. If an audit doesn’t give you those, it was a report, not an audit.

    Step 1: Crawl the site the way a search engine would

    I always start with a full crawl. Screaming Frog, Sitebulb, whatever you like. The important part is setting it up to behave like Googlebot Smartphone.

    Why smartphone? Because Google uses mobile-first indexing, so the mobile version of your page is the one that counts. If your mobile template hides content, links, or structured data that the desktop version shows, that gap is a finding all by itself. You’d be surprised how often it shows up.

    What I look for in the crawl

    • Status codes. Broken internal links (4xx), server errors (5xx), and redirect chains or loops. Every internal link should go straight to a final 200 URL. No detours.
    • Orphan pages. Pages that sit in the sitemap or analytics but that no internal link reaches. If you don’t link to a page, you’re telling crawlers it doesn’t matter. They tend to believe you.
    • Crawl depth. How many clicks it takes to get from the homepage to your important service and product pages. Money pages buried five clicks deep are common, and they’re one of the easier things to fix.
    • Rendering. Whether key content and links actually exist in the rendered HTML. This one bites hardest on JavaScript-heavy themes and page builders.

    Robots.txt and crawl directives

    Next, I open robots.txt and read it line by line. It’s a short file. One wrong Disallow can still hide an entire section of your site.

    Keep one distinction in mind, which Google spells out in its robots.txt introduction: robots.txt controls crawling, not indexing. If you want a page out of the index, you use a noindex directive, as covered in Google’s guidance on blocking indexing. And the page has to stay crawlable so Google can actually see that directive.

    Blocking a URL in robots.txt and adding noindex to it at the same time? That’s one of the most common contradictions I run into. Google can’t read the noindex on a page it isn’t allowed to fetch.

    Step 2: Check what’s actually indexed

    A crawl shows you what exists. Google Search Console shows you what Google decided to do with it.

    The Page indexing report lists your indexed pages and sorts the rest by reason: “Crawled – currently not indexed,” “Discovered – currently not indexed,” “Duplicate without user-selected canonical,” “Excluded by ‘noindex’ tag,” and so on.

    Don’t panic when you see a big excluded number.

    The skill here is reading the report, not reacting to it. Plenty of URLs are excluded on purpose: filtered pages, tag archives, old parameters. What you care about is whether any important page landed in the wrong bucket. So I export the report, line it up against the crawl, and flag every page that should be indexed but isn’t.

    Canonicals and duplicates

    Duplicate URLs split your signals. The usual suspects:

    • HTTP and HTTPS versions
    • www and non-www
    • Trailing slash and no trailing slash
    • Uppercase paths, tracking parameters, and print or pagination versions

    Google’s documentation on consolidating duplicate URLs lays out your options: redirects, rel="canonical", and consistent internal linking.

    In the audit, I check that every important page has a self-referencing canonical, that canonicals point to indexable 200 URLs, and that the canonical Google picked (you can see it in the URL Inspection tool) matches the one you declared. When Google disagrees with you, it’s worth finding out why.

    XML sitemaps

    Your sitemap should be a clean list of the URLs you want indexed. Nothing redirected. Nothing noindexed. Nothing returning an error. Google’s sitemaps overview covers the format and the limits.

    Here’s a trick I use on every audit. I compare three lists: the sitemap, the crawl, and the indexing report. Where they disagree is usually where the problem lives.

    Step 3: On-page hygiene across templates

    This isn’t copywriting yet. That comes later.

    At this stage I’m doing a hygiene check across templates, because one broken template can repeat the same mistake on hundreds of pages. Fix the template and you fix them all.

    • Titles and meta descriptions: missing, duplicated, truncated, or identical across a whole section.
    • Headings: one clear H1 per page that matches what the page is for, with a sensible H2/H3 outline underneath.
    • Thin and near-duplicate pages: location or service pages that differ by a single word. (You know the ones.)
    • Images: missing ALT text, huge files, and filenames like IMG_4021.jpg.
    • Structured data: whether the markup is valid, matches what’s visible on the page, and describes the organization, author, and page type accurately. Google’s introduction to structured data is clear that markup should reflect what’s actually on the page.

    That last one is where old-school SEO and AI-era SEO meet.

    Clear titles, honest headings, consistent business details, and accurate markup help Google understand a page. That same clarity makes it easier for answer engines to pull out a correct answer and credit the right source. Same work, two payoffs.

    Step 4: Performance and page experience

    Speed isn’t a ranking trick. Think of it as a floor you need to stand on.

    Google reports Core Web Vitals on three metrics:

    • Largest Contentful Paint (LCP) for loading
    • Interaction to Next Paint (INP) for responsiveness
    • Cumulative Layout Shift (CLS) for visual stability

    INP replaced First Input Delay as a Core Web Vital in March 2024, as the Chrome team announced on web.dev. The published “good” thresholds are an LCP of 2.5 seconds or less, an INP of 200 milliseconds or less, and a CLS of 0.1 or less, measured at the 75th percentile of real visits.

    When I audit performance, I start with field data: the Core Web Vitals report in Search Console and Chrome UX Report data. That’s what real visitors experience. Then I use lab tools like Lighthouse to figure out why a template is slow. It’s usually some mix of giant hero images, render-blocking scripts, bloated page-builder output, and third-party tags nobody remembers adding.

    Google’s own page on Core Web Vitals and Search treats these metrics as part of page experience. I think that’s the right way to see them. They matter for users. They don’t replace relevance.

    Step 5: Turn findings into a priority list

    This is where most audits fall apart.

    A crawler will happily hand you thousands of “issues.” Your job is to figure out which ten actually matter. I score every finding on two questions:

    How many important pages does it affect? And how directly does it block crawling, indexing, or understanding?

    Then I sort them into this order:

    1. Blockers first: anything that stops important pages from being crawled or indexed. Robots.txt mistakes, stray noindex tags, broken canonicals, server errors.
    2. Signal leaks next: redirect chains, duplicate URL versions, orphaned money pages, and internal links pointing at redirects.
    3. Template fixes: one change that cleans up hundreds of titles, headings, or image attributes at once.
    4. Performance: the slowest templates with the most traffic, fixed at the source.
    5. Polish: the long tail of minor warnings, handled as normal maintenance.

    Every item gets an owner, a rough estimate, and a way to confirm it’s fixed. Usually that’s a re-crawl or a validation request in Search Console.

    That list is the deliverable. Not the raw export.

    What not to do

    A few traps I’d steer you away from:

    • Don’t chase a tool’s health score. A 98 out of 100 can hide a noindexed service page. A 70 can belong to a site that’s doing just fine.
    • Don’t start with content or outreach. New pages and new links just pour into the same leaks. Fix the plumbing first.
    • Don’t “fix” everything at once. If you change URLs, redirects, and templates in one big release, you’ll never know what helped and what hurt.
    • Don’t treat the audit as one and done. Sites change every time a plugin updates or someone adds a page. A lighter health check every quarter catches regressions before they get expensive.

    How this shows up in a real engagement

    At Seopya, the audit is the first phase of our SEO and AI search services. It happens before we write a word of new content or send a single outreach email.

    You get the prioritized fix list. We work through the blockers together. Only then do we move into content strategy, link building, and AI search visibility, because all of that depends on a site search engines can read cleanly. Building on a cracked foundation just makes a bigger crack.

    FAQ

    How long does a site health audit take?

    It depends on how big and complicated the site is. A small brochure site can be reviewed in days. A large ecommerce site with faceted navigation and international versions takes longer. Honestly, speed matters less than whether the findings get prioritized and verified.

    Can I run an audit myself with free tools?

    Yes, a basic one. Google Search Console, PageSpeed Insights, and the free tier of a desktop crawler will surface most blockers on a small site.

    The hard part is interpretation. Knowing which warnings are intentional and which ones are actually costing you visibility takes practice.

    Does technical SEO matter for AI search and AI Overviews?

    It does. Google’s AI features documentation says that to be eligible as a supporting link in AI Overviews or AI Mode, a page has to be indexed and eligible to show in Search with a snippet. There aren’t separate technical requirements. Other answer engines also need to reach and parse a page before they can cite it.

    Good technical health won’t guarantee a citation. Technical problems can absolutely rule one out, though.

    The next step

    If you’ve been publishing for months and don’t have much to show for it, don’t publish faster. Look underneath.

    Crawl the site. Compare it to what Google has indexed. Fix the blockers on your most important pages. Then build on top.

    And if you’d like an experienced set of eyes on it, take a look at our SEO and AI search services or get in touch. I’m happy to walk you through what an audit of your site would involve, and what it can and can’t tell you.