On-Page and Off-Page SEO: What You Control and What You Earn

Dominik Breitbach · Founder and Lead SEO Strategist

Dominik Breitbach founded taismo GmbH, an SEO and GEO agency from Munich. He works on both halves of search this guide covers: What a page says about itself, and what the web says about it.

⏱ Reading time: 21 min🔄 Last updated: 26 July 2026

On-page SEO covers every ranking signal you place on your own pages: Search intent, content, headings, internal links, meta data and markup. Off-page SEO covers the signals other websites send about you, mainly backlinks, brand mentions and reviews. You own the first half outright and can only earn the second, which is why on-page work comes first in almost every project. This guide covers both halves and the order that makes the work pay off.

👉 Not sure which half is holding you back? An SEO audit separates the two and puts a number on each.

What on-page and off-page SEO each control

The split is a question of ownership. On-page SEO happens inside your own domain, where you publish and change things the same day. Off-page SEO happens on domains you do not control, so you influence it but never decide it. That one difference drives budget, sequence and waiting time.

Google’s own account of how Search works splits the same way. Neither half substitutes for the other: A perfect page nobody references stalls, and a heavily referenced page that answers the wrong question converts nobody.

On-page SEO: Everything you control on the page

On-page SEO makes a single page the best available answer to a specific question, and makes that fact obvious to a machine. It has six working parts.

Search intent decides the format

Every query carries an expected answer type. A definition query wants a paragraph, a comparison query wants a table, a supplier query wants prices and proof. Google ranks the format the searcher expects, so the first on-page decision is structural rather than lexical. Open the first ten results for your target keyword and count which page type dominates. If nine of ten are checklists and you publish an essay, no wording saves it. Your target audience decides which intent to serve first.

The title link and meta description earn the click

Your meta title is the main source Google uses to build the title link, though Google states plainly that it also draws on headings and anchor text and rewrites titles it considers unhelpful. The same holds for the meta description: Google treats it as one input for the snippet, never a guarantee. Describe the page accurately and put the specific detail first. A rising click-through rate at the same position proves the pairing works.

Headings carry the structure machines read

Use one H1, then H2s that each state a claim instead of echoing a question. A heading reading “Pricing” tells a crawler nothing. “SEO retainers in Germany start around 1,500 euros per month” can be lifted whole. Keep the hierarchy unbroken, because a skipped level turns a clean outline into a flat list. AI systems quote passages that stand alone, so a self-contained section under a descriptive heading is easy to cite.

Depth means answering the follow-up questions

Depth is not length. A page has depth when it answers the questions that logically follow the first one. An article on link building that never explains how to evaluate a linking domain stays thin at 3,000 words. Google’s helpful content guidance asks the same thing: Would someone leave feeling they learned enough to accomplish their task? Use secondary keywords and long-tail variants as a coverage map, and treat WDF*IDF or LSI keywords as rough topic checklists, not formulas Google runs. Keyword stuffing has been a spam policy violation for years. Depth decays too: Content freshness matters most where facts move.

Internal links distribute relevance you already have

An internal link is the one ranking signal you can create on demand. It passes context through the anchor text, moves authority from strong pages to weak ones, and tells Google which page owns which topic. Group pages into clusters: One broad cornerstone page per topic, supported by detail pages linking back. Clusters also prevent keyword cannibalization, where two of your own pages chase one query and split the signals. Link from inside sentences, and give breadcrumb navigation to every page more than one level deep.

Images, URLs and markup close the remaining gaps

Google uses alt text to understand what an image shows, and alt text doubles as anchor text whenever the image is a link. A readable URL carrying the topic beats an ID string, and a self-referencing canonical stops parameter variants from splitting one page into duplicate content. Note what canonicalization does: It consolidates signals onto one chosen version, and it is no penalty mechanism, whatever you may have read. On top sits structured data, the JSON-LD layer naming what the page is, who wrote it and who published it, which is the precondition for rich snippets.

The on-page factors that carry the most weight

Ranked by the effect we see once the fix lands, seven factors do most of the work:

1. Intent match: The page delivers the format the query expects.
2. Answer position: The core answer appears in the first paragraph, not on screen three.
3. Heading structure: Descriptive H2s that each carry one claim.
4. Topical coverage: Follow-up questions answered on the same page.
5. Internal linking: Contextual links from strong pages into the target page.
6. Title and description: Accurate, specific, written for the click.
7. Structured data: Valid JSON-LD that types the page, its author and its publisher.

Answer position earns second place for a reason we measured on our own site. Reviewing all 90 entries in our glossary in July 2026, we found nine that opened without defining the term and fourteen that buried the definition past the eighth word. Those entries lose featured snippets and AI citations, because the quotable sentence sits where no machine looks. The fix took minutes per page and no new content.

How to check your on-page SEO

Checking on-page SEO means holding one page against one query and finding the point where the two stop matching. Six steps do that, they need no paid tool, and they take about twenty minutes per page.

1. Step 1: Read the queries the page already gets. Open the performance report in Google Search Console, filter to the URL and sort by impressions. A page with impressions and almost no clicks is found and then rejected, which points at the title and the description. A page with neither is not competing for that query at all, which points at the content.
2. Step 2: Read average position from the table, not from the chart. Google documents the difference: In the chart, average position means “the average position of the topmost result from your site”, while a table row gives the position of that specific URL. Judging one page by the chart mixes it with every other page you own.
3. Step 3: Open the live results for the strongest query. Count the page types in the first ten. If your format appears twice and a different format appears eight times, the wording is not what is holding the page back.
4. Step 4: Read your first two sentences on their own. Hide the rest of the page. If those two sentences fail to answer the query, nothing further down compensates, and no AI system finds a passage worth quoting.
5. Step 5: Copy the headings into an empty document. The H2 list alone has to carry the story. A heading reading “Overview” or “Details” states no claim, so a crawler learns nothing from it and a reader skips it.
6. Step 6: Inspect the URL instead of your source code. The URL Inspection tool shows the version Google stored: The crawled HTML, the canonical Google picked, which can differ from the one you set, and the structured data it detected. That view settles disagreements your browser cannot.

Two things this pass will not give you. It says nothing about crawling, rendering or loading speed, which belong to the technical layer covered further down. And it says nothing useful about the score your SEO plugin prints next to the page. Google publishes no page score. A plugin grade is the vendor’s checklist, useful for catching an empty description or a missing alt attribute, worthless as a number to maximize.

Run the pass on your ten highest-impression pages first. That is where two points of extra click-through rate return more traffic than a new article, on demand you have already earned.

On-page SEO by page type

The six building parts stay the same on every page. What changes is which one decides the outcome, and that follows from what the page is for.

E-commerce splits the keyword work across two page types. The category page carries the head term, “running shoes”, while the product page carries the long tail, “brand model width 2E”. Google’s e-commerce documentation organizes the job along the same lines, around product data, structured data, site structure and pagination. Product markup is what lets price and availability show up in the result itself instead of only on your page. The blunt on-page consequence: A description copied from the manufacturer gives a search engine no reason to prefer your listing over every other shop that copied the same paragraph. Write your own for the products that carry revenue and leave the dead stock alone.

Local landing pages fail in their own way, through similarity. In our July 2026 audit of taismo.de we compared our own city pages against each other and found two of them overlapping at 79 percent word-level similarity. That is the doorway pattern, and Google’s spam policies name it. More paragraphs do not repair it. One genuinely local fact per page does: The industry that dominates that region, a client from it, the competitors you actually meet in those local results. We rewrote all five city pages rather than expand them.

The content management system decides none of this. WordPress, Shopify and a custom stack move the fields to different screens, and an SEO plugin writes your title and description tags and validates your markup. No plugin decides whether the page answers the query.

Off-page SEO: The signals other sites send

Off-page SEO is everything happening away from your domain that still shapes how search engines judge it. Our glossary entry on off-page optimization carries the full definition. Off-site SEO and off-page optimization are the same discipline under different labels. What matters here is which parts you can influence, and they fall into four signal families:

• Backlinks: Links from other domains, still the strongest single off-page signal.
• Brand mentions: Your name appearing in context, with or without a link.
• Reviews and ratings: Public judgments on Google, industry portals and marketplaces.
• Entity signals: Identical name, address and phone number everywhere, which turns you into an entity in the Knowledge Graph instead of a string of text. Conflicting data splits the entity and weakens all of it.

One logic unites them: Search engines treat independent third parties as harder to fake than your own claims about yourself. That is why the shortcuts fail. Buying links, exchanging them at scale or planting them in user-generated content count as link spam under Google’s spam policies, and Google asks that any link placed for advertising carry a rel value of sponsored, ugc or nofollow. The same applies to third-party content on your own domain.

What makes a backlink count

Counting links stopped working long ago. Three properties decide whether a link does anything at all:

1. Topical relevance: A link from a site publishing on your subject outweighs one from a general directory. A machine tool manufacturer gains more from an engineering trade magazine than from a lifestyle blog with ten times the traffic.
2. Site authority: The linking domain’s own standing, which tools approximate as domain authority. Treat that number as a third-party estimate, never a Google metric.
3. Editorial placement: A link inside a paragraph a human chose to write beats one in a footer or forum signature. Placement separates a citation from a placement fee.

The value flowing through a link is what the industry calls link juice. It splits across every link on the source page, so a page stuffed with outbound links passes less than an article carrying three careful references.

Two categories are worth avoiding outright. Paid link networks and article farms are the pattern the Penguin update was built to catch, and Google now devalues such links quietly rather than announcing a penalty, so the failure never shows up in your account. Parasite SEO, borrowing a strong third-party domain’s authority, became an explicit spam policy violation once Google addressed site reputation abuse.

Off-page SEO techniques that earn links

Off-page SEO runs on a short list of techniques that produce links and a long list that only produces activity. Six belong on the short list, ordered by the return we see per hour invested:

Digital PR is the label for pairing the last two with outreach, and the order is what most teams get wrong. The asset comes first: A journalist links because they needed a number and you had it. Outreach without an asset produces a contact list and nothing to send it. Volume works against you here as well. A hundred identical mails to a scraped list land in spam folders, while ten mails that name the specific article you are adding to get read. For local businesses the same effort goes further, because regional publications and associations are reachable in a way national outlets are not.

Anchor text is descriptive, and you do not get to choose it

Google’s link documentation is unusually direct about anchors: “Good anchor text is descriptive, reasonably concise, and relevant to the page.” It names the failures too, from generic wording such as “click here” through anchors so long that the meaning drowns to keyword stuffing inside the anchor, which its spam policies treat as a violation. For off-page work that carries an uncomfortable consequence. You write your internal anchors and you inherit your external ones. A profile in which a large share of incoming anchors carry the exact commercial keyword did not assemble itself, and it reads that way to a machine. Editors write whatever fits their sentence, so earned links arrive with untidy anchors: A company name, a bare URL, half a sentence. Tidy anchors are the fingerprint of a purchase.

Off-page signals beyond backlinks

Links are the loudest off-page signal, not the only one. Four others do measurable work.

Brand search volume is the quietest and the strongest. When people search your company name directly, that demand appears in every dataset a search engine holds, and it compounds: A brand people recognize gets clicked more often from the same position.

Reviews shape both the click and the local pack. A four-star profile with 200 recent reviews outperforms a five-star profile with six, because volume and recency carry the credibility. For any company with an address, reviews are often the biggest local lever.

Consistent business data across directories, profiles and your own site lets a search engine merge every mention into one entity. Different phone numbers in different places do the opposite: They split the entity and weaken all of it. This is unglamorous work with an outsized payoff, because it costs nothing but attention and it gates everything an AI system can confidently say about you.

Editorial coverage is where content marketing and off-page SEO meet. Original data, a real case study or a tool people need earns links without asking, the idea behind skyscraper content. Paid formats such as advertorials and influencer placements need the sponsorship markup Google asks for, and they buy reach rather than ranking.

One defensive note: Negative SEO, a competitor pointing spam links at your domain, worries clients far more than it should. Google states that most sites never need the disavow tool, because its systems already ignore those links.

E-E-A-T: Where the two halves meet

E-E-A-T stands for experience, expertise, authoritativeness and trust, and our glossary entry on E-A-T carries the definition. What belongs here is the split, because each of the four has an on-page half you write and an off-page half the rest of the web decides.

Google’s guidance on helpful content puts trust above the other three: “Of these aspects, trust is most important. The others contribute to trust.” The same page asks three questions of any piece of content, who wrote it, how it was produced and why it exists, and calls the “why” perhaps the most important of them. All three are on-page fixes you can ship this week. A real byline linked to a page describing that person, a visible publication date, named sources: That answers who and how in full.

The reputation half is the one you cannot ship, and it explains a pattern every agency runs into, where a thin page from a recognized practitioner outranks a thorough page from a domain nobody has heard of. E-E-A-T is also not a dial inside the ranking systems. Google publishes it in the guidelines its quality raters work from, and those raters judge whether the systems return good results rather than scoring your page. Read it as a description of what good looks like, then build both halves toward it.

Where technical SEO sits between the two

Both halves rest on an assumption that is anything but automatic: That a search engine can reach the page, render it, understand it and store it. Holding up that assumption is the job of technical SEO, the third layer, and it fails silently. A page blocked by a stray noindex receives no traffic, however good the content and however strong the links pointing at it.

The practical rule is a sequence. Content Google cannot crawl ranks nowhere, so technical defects outrank on-page improvements whenever both are present. Once crawling, rendering and indexing work, the technical layer stops being the bottleneck. Our separate guide to technical SEO works through crawling, rendering, Core Web Vitals and structured data in the order that decides the outcome.

👉 Running all three layers alongside a full-time job rarely works. We keep them running for clients in a monthly SEO retainer, with the first measurable results inside three months.

Which half is holding you back

Most budget arguments between on-page and off-page end the moment somebody reads the symptom correctly. Six symptoms cover nearly every site we open:

One comparison settles the question faster than any audit. List the five URLs ranking above you for your target query and note two things per URL, its referring domains and how fast it answers the query. Three outcomes are possible. Those pages hold several times your links and answer no better than you do: Links are the constraint, and rewriting will not move you. They hold roughly what you hold and answer visibly better: The constraint is your page, so rewrite the opening, cover the follow-up questions and measure again in four weeks before spending anything on outreach. Both are true at once: You are on a crowded head term, which is the exception described in the next section.

Run the comparison per URL and not per domain. Domain-level metrics hide the case that decides most projects, where a strong site fields a weak page. That page is beatable with on-page work alone, and it is the fastest win in any conversion-relevant part of a site.

Which to fix first: The order that works

On-page work comes first in almost every project, for three reasons unrelated to preference.

It is faster. An on-page change takes effect at the next crawl, often within days. Off-page results arrive over months, because you wait on other people’s schedules.

It is cheaper. Rewriting an intro paragraph costs an hour. Earning a link from a trade publication costs a relationship, an angle worth covering and several follow-ups.

It multiplies whatever follows. A link pointing at a page that fails to answer the query still fails to convert, and the ranking gain gets clawed back by the first competitor with a genuinely better page. Fix the destination first.

The exception is real and worth naming plainly. Where every competitor has already done the on-page work well, off-page authority becomes the deciding difference. That is the situation on head terms in crowded verticals, and it explains how an established domain outranks a better page. There, on-page work gets you to parity and off-page work gets you past it.

On-page and off-page SEO inside AI answers

The split survives the shift to AI search with the weighting moved. AI Overviews and assistants such as ChatGPT and Perplexity assemble answers from passages, so the on-page unit that matters shrinks from the page to the paragraph. A section defining a term cleanly under a descriptive heading gets quoted. The same facts spread across four paragraphs do not.

Off-page signals decide something different here: Whether your brand appears in the answer at all. Language models draw on how often and how consistently an organization is described across the open web, so consistent naming, a cleanly resolving entity and coverage on sources the models already trust raise the odds of being named as the source instead of paraphrased anonymously.

This is the discipline we run as generative engine optimization, and it recombines both halves rather than adding a new toolkit. GEO asks for quotable on-page passages, clean structured data and an off-page footprint consistent enough that a model can tell who you are. A well-built grounding page delivers all three.

The combined checklist

Work down each column and stop at the first line you cannot tick.

Two numbers tell you whether the work lands. On-page progress shows up as impressions and average position on the pages you touched, visible in Search Console within two to four weeks. Off-page progress shows up as referring domains and branded search volume, over two to three quarters. Judging off-page work on a four-week timeline is how clients talk themselves out of what would have worked.

On-page and off-page SEO FAQ

What is on-page SEO and off-page SEO?

On-page SEO is every optimization you make on your own pages, including content, headings, internal links and meta data. Off-page SEO is every signal generated outside your domain, mainly backlinks, brand mentions and reviews.

What is off-page SEO?

Off-page SEO is the work of earning signals from other websites that tell search engines your site is credible. Backlinks from relevant domains are the strongest signal, followed by brand mentions, reviews and consistent business data.

What is the difference between on-page and off-page SEO?

The difference is ownership. You decide your on-page SEO and can change it today, while off-page SEO depends on other people linking to you, which takes months and cannot be bought safely.

How do you do on-page SEO?

Match the page format to the search intent, answer the core question in the first paragraph, structure the page with descriptive headings, cover the follow-up questions, then set internal links, title, description, alt text and structured data.

How do you improve off-page SEO?

Publish something worth citing, convert unlinked brand mentions into links, keep your business data identical everywhere and collect reviews continuously. Skip paid link networks: Google devalues them silently, so you pay and gain nothing.

Why is on-page SEO important?

On-page SEO is the only half you fully control, and it decides whether the traffic you earn converts. It also multiplies off-page work, because a link pointing at a page that answers the wrong question wastes the link.

What are off-page SEO techniques?

The six techniques that reliably earn links are mention reclamation, citations and profiles, broken link building, expert commentary, original data and guest articles. Mass directory submissions and link exchanges produce activity instead of rankings.

Is off-site SEO the same as off-page SEO?

Off-site SEO, off-page SEO and off-page optimization name the same discipline: Every ranking signal created outside your own domain. The three terms are interchangeable and no search engine treats them differently.

How do you check on-page SEO?

Filter the page in Search Console, hold its queries against what the page really answers, then read your first two sentences, your heading list and the internal links pointing in. Finish with the URL Inspection tool to see the version Google actually stored.

Sources

• Google Search Central: “In-depth guide to how Google Search works”, Google, 2026.
• Google Search Central: “Creating helpful, reliable, people-first content”, Google, 2026.
• Google Search Central: “Spam policies for Google web search”, Google, 2026.
• Google Search Central: “Qualify your outbound links to Google”, Google, 2026.
• Google Search Central: “Influencing your title links in search results”, Google, 2026.
• Google Search Central: “Control your snippets in search results”, Google, 2026.
• Google Search Central: “Google Images SEO best practices”, Google, 2026.
• Google Search Central: “How to specify a canonical URL with rel canonical”, Google, 2026.
• Google Search Central: “Introduction to structured data markup in Google Search”, Google, 2026.
• Google Search Help: “Disavow links to your site”, Google, 2026.
• Google Search Central: “Link best practices for Google”, Google, 2026.
• Google Search Central: “Ecommerce best practices for Google Search”, Google, 2026.
• Google Search Central: “Product structured data”, Google, 2026.
• Google Search Central Blog: “Our latest update to the quality rater guidelines”, Google, 15 December 2022.
• Google: “Search Quality Rater Guidelines” (PDF), Google, 2026.
• Google Search Console Help: “Performance report (Search results)”, Google, 2026.
• Google Search Console Help: “URL Inspection tool”, Google, 2026.