The three rules for clean blog pagination

All three rules appear – directly or indirectly – in Google’s own documentation. They don’t contradict each other, they complement each other.

1. Self-canonical on every archive page

Every component page canonicalises to itself: /page/2/ to /page/2/, not to page 1. A canonical from all archive pages to the first page tells Google “the subsequent pages are duplicates” – they’re not, they list different posts. Google explicitly recommends in the official documentation giving every paginated page its own canonical URL.

2. Normal HTML links between the pages

Googlebot follows a-href links. Pagination that only works via JavaScript click handlers or a “Load more” button is a dead end for the crawler – Google names the solution in the same documentation: supplement or replace buttons with real links.

3. No noindex on archive pages

The widespread SEO plugin option “don’t index archive pages” dates from a time when people feared duplicate content. Today it costs you twice: first, permanently non-indexed pages drop out of the crawl rhythm; second, it weakens the link path to older posts. Leave archive pages indexable – they rarely rank, but they carry.

Pagination: the rules at a glance

Measure	Recommendation	Why
Self-canonical per archive page	Mandatory	Prevents Google from treating subsequent pages as duplicates
Normal HTML links between the pages	Mandatory	Only this way can Googlebot follow all pages
noindex on archive pages	Avoid	Cuts the link path to older posts
rel=”prev” / rel=”next”	No effect since March 2019	No longer an indexing signal, but harmless
Posts per archive page	10–20	Usual corridor for load time and usability

How to check your pagination

Five steps to check the state of a blog’s pagination without extra tools:

1. Open the page source of an archive page (right-click → View page source) and search for rel="canonical": does the URL point to itself or to page 1?
2. Check the SEO plugin settings: is an option like “don’t index archive pages” or “noindex paginated pages” activated? If so, deactivate it.
3. Check whether /page/2/ and further pages are reachable via real a-href links – not only via a “Load more” button without a link fallback.
4. Check archive depth: how many clicks away from the home page are the oldest posts? Tidy up categories and tags, link important evergreen posts directly.
5. Leave rel=”prev”/rel=”next” in the theme if present – removing isn’t a must, adding has brought nothing since 2019.

Frequently asked questions on pagination

Keep reading

Where accessibility and SEO are the same work

A large share of the WCAG criteria overlaps with what belongs to on-page optimisation anyway. Six points that serve both at once:

• Alt texts: screen readers read them aloud, Google uses them for image search. An alt text of 5 to 15 words serves both.
• Heading hierarchy: exactly one H1, then H2/H3 without jumps. Assistive technology navigates by heading, Google reads the structure.
• Contrasts: WCAG 2.1 AA demands 4.5:1 for body text and 3:1 for large type.
• Keyboard operation: forces real button and a elements instead of click handlers on divs – exactly this HTML is what Google crawls more reliably.
• Understandable link texts: “To the returns policy” instead of “Click here” – tells screen readers and Google what’s behind the link.
• Clean, semantic HTML: nav, main, table – the shared basis of accessibility and crawlability.

WCAG 2.1 AA as the audit frame

The BFSG names no pixel values. The practical frame of reference is EN 301 549, which points to WCAG 2.1 AA – around 50 success criteria:

1. Test automatically: Lighthouse, WAVE or axe DevTools. Free, done in 10 minutes.
2. Assess honestly: automatic tools only find about 30 to 40 percent of violations. A Lighthouse score of 100 does not mean “accessible”.
3. Check manually: operate the whole page with only Tab, Enter and arrow keys. Run through the checkout with a screen reader (VoiceOver comes preinstalled on every iPhone).
4. Prioritise: checkout and forms first – that’s where the legal and economic risk arises. Then templates, individual pages last.

BFSG rebuild at a workshop supplies shop

“At an online shop for workshop supplies with around 3,000 items, we straightened the heading structure because of the BFSG, backfilled alt texts for the product images and made the checkout keyboard-operable. After three months, mobile visibility was noticeably better – not because of a secret switch, but because the basics were finally right.”

— Viktor Pásztor, SEO freelancer

Frequently asked questions on accessibility and mobile SEO

Keep reading

Proven, plausible, speculation: the distinction that counts

GEO is a field full of claims. Before you invest money or time, sort every measure into one of three tiers. This sorting is the actual value of this article.

Proven – stated in Google’s documentation:

• At page level, two prerequisites: indexed and snippet permission. Plus the inclusion of the website in the Search Console.
• Unique, experience-based content is the strongest lever. Summaries of what’s already been said are not.
• Structured data is not required for generative search features. A special schema does not exist.
• Google Search does not evaluate llms.txt.

Plausible – a justifiable derivation, without proof of effect:

• Entity consistency: identical company facts across the website, schema, legal notice and business profiles. Grounding systems compare sources; contradictions weaken the assignment.
• A grounding page that bundles your company’s citable core facts. Useful for humans, possibly useful for machines.
• Allow OAI-SearchBot: according to OpenAI, websites that lock out the crawler do not appear in ChatGPT’s search answers. Allowing it is therefore a prerequisite – a citation probability does not follow from it.

Speculation – circulating, but not verifiable:

• llms.txt as a visibility lever.
• “Chunking” content, rewriting “for the AI”, purchased mentions. Google explicitly names all three as unnecessary.
• “AI ranking scores” from third-party tools. Google makes it clear: no external tool has access to internal ranking or AI systems.

A guide that mixes these three levels sells you speculation at the price of facts.

Controlling AI crawlers: GPTBot, ClaudeBot, PerplexityBot, Google-Extended

Every major provider now separates training, search and user requests. The overview:

Provider	Crawler	Purpose
OpenAI	GPTBot	Training the models
OpenAI	OAI-SearchBot	Search results in ChatGPT
OpenAI	ChatGPT-User	Page retrieval on user request
Anthropic	ClaudeBot	Training the models
Anthropic	Claude-SearchBot	Search quality
Anthropic	Claude-User	Page retrieval on user request
Perplexity	PerplexityBot	Search results in Perplexity
Perplexity	Perplexity-User	Page retrieval on user request
Google	Google-Extended	Training and grounding for Gemini and Vertex AI

The most common misconception first: Google-Extended does not control AI Overviews. The token concerns the training of future Gemini models and grounding in Gemini apps and Vertex AI. Google explicitly writes: the token influences neither inclusion in search nor ranking. AI Overviews run via the normal Googlebot. Whoever wants out of there would have to suppress snippets – and thereby also lose classic search.

Whether you allow the other providers’ crawlers is a business decision, not a technical one. Allowing means: a chance of being named in their answers, often without a click. Blocking means: your content isn’t used there, but you’re invisible. By the way, not every retrieval respects robots.txt. Perplexity itself documents that Perplexity-User generally ignores robots.txt rules for user-initiated retrievals.

Entity consistency: the underestimated part

Grounding systems compare statements across multiple sources. If your website, legal notice, Organization schema and Google Business Profile show different company names, addresses or self-descriptions, every contradiction weakens the assignment. The measure is unspectacular: the same facts everywhere. A uniform company name including legal form, identical contact details, the same key figures.

This is no new AI craft, but well-maintained on-page work with clean schema – except that sloppiness here could become more expensive than before. The effect is not proven. But it’s the most plausible unproven measure in the whole field, and it costs almost nothing.

Case study from a B2B project

A B2B client wanted to know why ChatGPT names the competition for industry questions, but not him. The check revealed: robots.txt blanket-blocked all non-Google crawlers – a leftover from an old security configuration. Plus three different company names in circulation: one in the legal notice, one in the schema, a third in industry directories.

The measures: OAI-SearchBot and PerplexityBot released per the client’s decision, GPTBot blocked at the client’s request, the company name unified everywhere, core facts bundled on a visible, linked page. There was no citation promise – no provider guarantees visibility. But the documented prerequisites are now met, and the decision on every crawler lies with the client in writing.

FAQ

Keep reading

GEO: the new goal is citation instead of click

Since Google has been rolling out AI Overviews and the AI Mode, search answers many questions directly – with a generated answer that cites sources. What that means for clicks has been measured: for searches with an AI summary, US users clicked a classic result in only 8 percent of visits in March 2025, without a summary in 15 percent (Pew Research Center).

GEO describes optimising for exactly these citations. The industry classifies it similarly: WordStream and Seer Interactive describe GEO as an evolution, not a replacement of SEO; market observers like Colling Media citing Gartner expect a noticeable decline in classic search queries in favour of AI answers by the end of 2026 – the exact number is a forecast, the direction is considered undisputed.

What GEO practically demands overlaps strongly with good SEO: precise answer blocks (40–60 words, with a number), clean structure, consistent entities – meaning the same company name everywhere, the same facts, linked profiles. What’s new is the focus: you write citable units, not just ranking pages.

To be honest, part of it is: how strongly GEO measures increase the citation probability is harder to measure than a ranking. Whoever sells you “position 1 in ChatGPT” is selling a promise without a measuring instrument. What’s proven: without indexable, structured, consistent content, you certainly won’t be cited.

The three-way comparison

This is how the three approaches stand side by side:

Criterion	SEO	GEO	MCP SEO
Goal	Ranking + click	Citation in AI answers	Automating routine
Measurability	high (positions, clicks)	developing (GSC doesn’t yet report AI traffic separately)	high (completed tasks, time, costs)
Time horizon	Months	Months, building on SEO	immediately effective as a way of working
Who needs it	everyone with search demand	everyone whose customers use AI answers	everyone paying for SEO routine
Relationship	GEO requires SEO · MCP SEO accelerates both

The assessment in the initial call

In the free initial call, I go through the three questions in the same order this comparison sets out. First SEO: is the website technically clean, indexable, with content that has substance? If not, any GEO discussion is premature – without a ranking basis, there’s nothing that could be cited.

Only then does GEO come to the table: are the most important answers formulated as standalone, citable blocks, with consistent entities across the website, schema and business profile? And lastly MCP SEO: which of the recurring tasks – crawls, meta data, redirects, reports – is worth automating, without a single change going live without approval? This order is no coincidence: it follows exactly the dependency from the comparison table above.

Frequently asked questions about SEO, GEO and MCP SEO

Keep reading

Checking in the Search Console whether a URL is in the index

Forget the site: query – it’s a rough estimate, not a report. The reliable source is the Google Search Console, and it costs 0 euros.

Single URL: enter the address into the URL inspection. You’ll see in seconds: is the URL on Google? When was it last crawled? Which canonical URL did Google choose?

Whole website: the “Indexing → Pages” report lists all non-indexed URLs with reasons. Two statuses come up constantly: “Crawled – currently not indexed” (a quality or duplicate problem) and “Discovered – currently not indexed” (not yet crawled, often a question of crawl prioritisation).

Crawl budget: only a topic from five-figure URL counts

The most overrated topic in technical SEO. Websites under about 10,000 URLs practically don’t need to worry. It becomes relevant from around 1 million URLs – or from about 10,000 URLs with daily changing content, for example large shops with faceted navigation.

If your website has 300 pages and one of them isn’t indexed, that’s not down to crawl budget. It’s blocked, too thin or a duplicate.

Indexing signals at a glance

Signal	Controls	Effect
noindex	Indexing	Page stays out of the index – requires it to be crawlable
Canonical	Duplicates	Recommendation to Google, not a block
robots.txt (disallow)	Crawling	Prevents retrieval, but not necessarily indexing
XML sitemap	Discovery	Should only contain index-worthy URLs with status code 200

The four signals solve different tasks. Whoever mixes them – for example using robots.txt as a deindexing tool – produces exactly the mistakes from the next section.

Common mistakes with indexing and crawling

When which channel is the right choice

SEA, if you need results fast: product launch, seasonal business, or to test keywords before betting twelve months of SEO on them. A click price of 5 euros at a 2 percent conversion rate means: 250 euros per lead – you have to be able to afford that, permanently.

Social media, if your offer needs explanation or is new and nobody is searching for it. Or if your audience buys visually: fashion, furnishings, food.

Email, as soon as you have a list. The channel with the consistently best return, because you don’t pay anew for the reach every time.

SEO, if there’s measurable search volume and you think in timeframes of years. What SEO delivers in detail is its own topic – this is about the classification in the overall picture.

Honestly: when SEO is the wrong choice

Short-term promotions. A discount weekend, an event in six weeks – by the time SEO kicks in, the promotion is over. That’s a case for SEA and email.

No search volume. If nobody searches for your offer, there’s nothing to rank. A novel product needs demand first – via social, PR or ads. SEO comes after.

No budget for the dry spell. Whoever can’t last the first 6 to 12 months quits before the effect sets in. Then better start with SEA, and begin SEO when there’s headroom.

Common mistakes in the channel mix

Mass AI content is the new black hat

The classic techniques are known and reliably detected. How reliably is shown by Google’s own numbers: in 2022, SpamBrain detected five times more spam websites than the year before and 200 times more than in 2018 – 99 percent of visits from Google search remained spam-free (Google Webspam Report 2022).

Period	Detected spam websites compared	Share of spam-free search visits
2018	Reference value	–
2021	Basis for the 2022 comparison value	–
2022	5 times more than 2021, 200 times more than 2018	99 %

Source: Google Webspam Report 2022.

Today’s relevant grey area is a different one: hundreds or thousands of AI-generated pages without their own substance, produced solely for rankings. With the March 2024 spam update, Google created its own category for this: “scaled content abuse” – and explicitly classifies not the tool, but the purpose. Writing with AI assistance is allowed; rolling out masses of pages without added value is not.

This is the point where it’s decided today whether a service provider works cleanly. Anyone promising you “500 pages in a week” is selling you a risk. How I work with AI instead – automating the routine, approval before every change, substance from your knowledge – is described under MCP SEO.

Negative SEO: the sober checking plan

The fear of link attacks is bigger than the danger: Google has explained for years that spam links are largely devalued automatically – which is why, according to Google’s own documentation, the disavow tool is intended for one case only: when a manual action for unnatural links exists or realistically looms. Anyone submitting a disavow for every odd backlink is fighting a problem Google solved long ago.

If you suspect an attack, check in this order:

1. Search Console → Manual actions: if nothing is listed there, there’s no acute penalty problem.
2. Date the ranking loss: does it coincide with a Google update? Then the cause usually lies with your own site, not the competitor.
3. Look at the backlink profile (Ahrefs/Semrush): a sudden surge of identical anchor texts from spam domains is conspicuous – document, monitor.
4. Check for content theft: have copied content on third-party domains removed via a DMCA report if necessary.
5. Disavow only in case of a manual action – then targeted, with documentation.

Common mistakes with black hat, white hat and negative SEO

The iron rule: markup = visible content

Everything you mark up must be visible on the page. A price in the markup that isn’t on the page; a rating without visible reviews; questions in FAQPage markup that appear nowhere in the text – all of that violates Google’s structured data policies.

The consequence is real: Google imposes manual actions for structured data spam. Then the entire domain loses its rich results. Rule of thumb: write the visible content first, generate the markup from it – never the other way around.

Why FAQPage markup stays anyway

Rich results were never the only consumer of structured data. FAQPage markup delivers question-answer pairs machine-readably – exactly the format AI search systems process when assembling answers. A cleanly marked-up answer of 40 to 60 words is a ready-made citation unit for AI Overviews and AI Mode.

On top of that, the entity effect: every valid markup feeds the search engine’s understanding of who you are. My recommendation: keep FAQPage markup – but only for visibly answered questions.

Validating: two tools, two questions

Two tools, two different questions:

1. Rich Results Test (Google): is the markup suitable for a Google rich result? Only checks supported types.
2. Schema Markup Validator (Schema.org): is the markup syntactically valid – regardless of what Google makes of it?

Use both: an FAQPage block has been invisible in the Rich Results Test since May 2026, but can still be checked in the Schema Markup Validator. After that, keep an eye on the Search Console.

@id linking in practice

This is what the linking looks like in practice:

“For a B2B software provider, we consolidated the scattered Organization blocks across around 300 pages onto one central @id and linked the authors as Person entities. Four months later, the provider was named as a source in AI Overviews for the first time – before that, it didn’t appear there at all.”

— Viktor Pásztor, SEO freelancer

What was decisive was not a single schema type, but that in the end every page pointed to the same Person and Organization @id – instead of creating dozens of unconnected copies.

Frequently asked questions about structured data

Keep reading

Status codes at a glance

Status code	Meaning	SEO effect
200 OK	Page is reachable and delivers content	Normal state, no action needed
301 Moved Permanently	Permanent redirect	Transfers signals to the target URL, standard for moves
302 Found	Temporary redirect	Target URL only takes over signals with a delay
404 Not Found	Page not found	Rechecked by Google several times before it drops from the index
410 Gone	Deliberately removed, no replacement	Clear signal, usually disappears from the index faster than 404

The table shows the codes that are actually relevant in daily SEO work. All other 4xx and 5xx codes are usually technical errors, not a strategic decision.

302: almost always the wrong choice

The search volume for “302 seo” is small, the problem behind it big: 302 is usually set accidentally because it’s the default in some systems. The consequence is no drama – Google treats a permanently standing 302 like a 301 after some time – but the transition period is unnecessary blur: the old URL stays in the index longer, the new one takes over more slowly. Mnemonic: 302 only if you know the return date.

The relaunch case: where visibility really gets lost

Most ranking losses after a relaunch have the same cause: the redirect concept was missing or only covered the main pages. How much is at stake is shown by a study by Ahrefs across a good two million websites: 66.5 percent of all links set to these domains since 2013 are now dead.

Categories, old blog slugs, parameter URLs – everything that carried backlinks or rankings runs into 404. That’s why every relaunch should be preceded by a complete crawl of the old domain plus a GSC comparison, and every URL gets one of the four decisions from the decision tree above. The overall framework for this work is provided by technical SEO.

Frequently asked questions about status codes and redirects

Check indexing – the blocker before everything else

Open the Google Search Console and look at the “Indexing > Pages” report. There you’ll see how many of your URLs are in the index – and why the others are missing: noindex tags, redirect chains, duplicates without a canonical URL, crawling errors.

You check individual important pages with the URL inspection. If it says “Crawled – currently not indexed”, Google considers the page too weak or too similar to others. That’s a quality signal, not a coincidence.

Rule of thumb from my audits: websites with more than 500 URLs almost always have indexing problems nobody knew about. This check is among the first steps in every SEO audit.

Internal links – underestimated and completely in your hands

Internal links show Google which pages matter to you and pass authority from strong to weak pages. Unlike backlinks, they cost you nothing but care.

Three rules are enough to start. First: every important page gets links from at least 3 thematically fitting other pages. Second: the anchor text describes the target (“keyword research for online shops” instead of “click here”). Third: no orphaned pages – every URL reachable within a maximum of 3 clicks from the homepage.

You’ll find your strongest link givers in the Search Console under “Links”: the pages with the most external references should link to your most important target pages. What that looks like in detail is shown on my on-page optimisation page.

Reduce load time – mandatory, but no miracle cure

Core Web Vitals are a ranking factor, but a small one. A page doesn’t jump from position 30 to 3 because it loads faster. But it can lose positions if it’s significantly slower than the competition.

The target values: LCP under 2.5 seconds, INP under 200 milliseconds, CLS under 0.1. The biggest brakes: uncompressed images, third-party scripts, missing caching. Measure with PageSpeed Insights, fix the three biggest items, measure again.

Position and intent decide the click-through rate

How strongly position and search intent decide the click-through rate is shown by a look into the Search Console of a project I manage.

Keyword	Position	Click-through rate	Impressions
chiasamen mit apfelessig	1.5	13.6 %	–
höhle der löwen abnehmen	9	0.49 %	29,449

Source: Google Search Console, monikagungl.com (public reference), 08/2025–07/2026. Raw search volume without a top position and fitting intent is almost worthless – the long-tail keyword at position 1.5 gets a significantly higher click-through rate than the keyword with almost 30,000 impressions at position 9.

“A Shopify merchant with a mid-four-figure URL inventory wondered about stagnating rankings. The crawl revealed: filter parameters had created thousands of duplicates, and Google was spending its crawl budget on variants instead of products. After the cleanup, the number of indexed relevant pages rose significantly within two months – the rankings followed.”

— Viktor Pásztor, SEO freelancer

Frequently asked questions about improving rankings

Keep reading

Alt texts: accessibility first, then SEO

The alt text exists for people who can’t see the image – screen reader users, but also anyone with a dropped connection. SEO is the side effect. Whoever reverses the order writes alt texts like “images seo optimisation agency berlin” – useless for the blind, risky for rankings. In its own image search documentation, Google names the same order: first an honest description, then visibility.

• Describe what’s visible, in one sentence: “A mechanic trues a wheel in the assembly stand.”
• 80 to 125 characters are a good corridor.
• Decorative images get an empty alt attribute (alt=””), so screen readers skip them.
• No “image of…” – the software already knows it’s an image.

This is how file name, alt text and fixed dimensions interlock in the markup – purely as an example, without a real image loading here:

<img src="rennrad-carbon-rahmen-schwarz.avif"
     alt="Black carbon road bike frame photographed from the side"
     width="800" height="600" loading="lazy">

Dimensions and srcset: never deliver more pixels than needed

A 2,400-pixel-wide image in a 400-pixel container is the most common performance mistake in my audits. How much is at stake is shown by the Web Almanac 2024 by HTTP Archive: at 1,054 KB, images are the largest single item in the median page weight of a desktop homepage – ahead of JavaScript, fonts, CSS and HTML. In cases like the one described above, the browser loads a multiple of the data needed for the visible container.

The solution is srcset with sizes: the same image in several widths (e.g. 400, 800, 1200, 1600 pixels), the browser picks the right one. Modern CMS generate the variants automatically – just check that the sizes attribute matches the layout. This includes fixed width and height attributes in the HTML: they prevent layout shifts and keep your CLS value stable.

AI images: labelling obligation under the EU AI Act

Since August 2026, Article 50 of the EU AI Act applies fully: AI-generated or AI-manipulated images must be recognisable as such. Concretely: AI images need machine-readable labelling (many generators embed C2PA metadata – don’t strip it when compressing). For realistic-looking people, places or events, the labelling must additionally be visible, for example in the caption.

Stock images too: Check licence and personal rights. “Royalty-free” doesn’t mean rights-free – editorial licences often don’t allow commercial use, and depicted people need a model release. Document source, licence type and scope of use per image – a spreadsheet is enough.

An image project in numbers

Numbers weigh more than claims. An example from an ongoing client project shows what the individual levers – format, srcset, lazy loading rule – deliver together in practice.

“For a bike parts shop, we converted around 12,000 product images to AVIF, fixed srcset and set the hero image to eager. The LCP dropped from 3.4 to 1.9 seconds in the CrUX report, and image impressions in the Search Console rose by about 30 percent within three months.”

— Viktor Pásztor, SEO freelancer

None of the levers is spectacular on its own. Taken together – format, right size, correct loading priority – they make the difference that shows up in the Core Web Vitals.

Frequently asked questions about image SEO

Keep reading

The break-even calculation, laid open

Take a realistic B2B example: target keywords with a combined potential of 2,000 clicks per month, click price €6.

SEA: 2,000 clicks × €6 = €12,000/month media budget, plus management. From day 1, precisely scalable, ends with the budget.

SEO: Support from €1,000/month with me. Result: little in the first months, from month 6 to 12 a growing organic share. If the site reaches even a quarter of the click potential (500 clicks/month), it arithmetically replaces €3,000 of media budget – three times the support costs. At half: six times.

The break-even – the point where cumulative SEO costs fall below cumulative ads costs for the same clicks – depends almost entirely on the click price: at €1 clicks it takes long or never pays off; at €5 it’s typically in the second year; at €10+ often already in the first. Rule of thumb: the more expensive your clicks, the faster SEO pays off. The assumptions are deliberately conservative – recalculate with your real click prices, they’re in your Google Ads account.

Criterion	SEO	SEA (Google Ads)
Cost	support from €1,000/month	in the example 2,000 clicks × €6 = €12,000/month media budget
Time to effect	6 to 12 months until a growing organic share	from day 1
Sustainability	the stock remains, even when paused	ends with the budget
Controllability	no guaranteed result	can be switched on and off instantly
Typical break-even vs. ads costs	month 12 to 24, depending on click price	–

Break-even calculation based on the example in this section: 2,000 clicks/month potential, €6 click price, SEO support from €1,000/month. The actual break-even depends on the respective click price.

The hybrid strategy: ads as a data source for SEO

The smartest use is not either-or. This is how I combine both channels in client projects:

1. Ads first on the purchase keywords: 4 to 8 weeks of campaigning deliver real conversion data per keyword – before a single SEO text is written.
2. SEO on the winners: The keywords that demonstrably bring inquiries in ads get pages and content. No 12-month bet on assumptions.
3. Scale ads back step by step where organic rankings take over – the freed-up budget moves to keywords that are organically out of reach.

This way SEA finances the learning phase and SEO permanently lowers unit costs. The channel comparison with social and email is its own topic – this was only about the calculation between the two search channels.

Common mistakes in the SEO vs. SEA comparison