Why Isn't My Content Ranking in Google and AI? We Analysed 10,937 SEO'd Pages

Advanced on page SEO made simple.

Try POP today

Includes AI Writer

7-day refund guarantee

10,937

WEB PAGES ANALYSED

99%

Missing Schema

33/100

Median Score

The content flood since AI is real. But it's not the whole story.

Before ChatGPT launched in November 2022, roughly 1 in 20 new articles published online was AI-written. Today, that number is 1 in 2. AI-generated text now costs as little as $0.50 per 1,000 words, compared to ~$200 for a human writer - a 400× cost reduction. When the price of creating content drops by 99.75%, volume explodes. The web didn't get twice as hard to rank in. It got infinitely cheaper to flood.

But quantity isn't the real threat. If you've published a good page and watched it go nowhere, the reason is usually something you can't see. Every web page has a layer that Google, Claude, ChatGPT, and Perplexity read - and most people don't know it's there. When it's missing, even great content stays invisible.

So, does the AI content flood mean it's now impossible to get noticed in Google and AI?

THE SHORT ANSWER

No. And this study is the proof. The bar for most content online is far lower than you think - and that means the bar to outperform it is also lower than you think. The opportunity is massive. You just need to know these 7 signals to use in your content.

About this study

Last month - between April 10 and April 30, 2026 - we pulled the export data from POP reports run across 10,937 web pages by our own PageOptimizer Pro users. People who already know SEO and actively use a tool to optimise their content.

And here's what the data shows: even this group is starting from a shockingly low baseline. If the people who care most about SEO are this far back, imagine the everyday business owner just publishing and hoping.

None of these were cherry-picked - they're real pages, for keywords those businesses actually wanted to rank for. Most had been published already. Some for years.

This is what those pages looked like before anyone optimised them. The starting line, not the finish.

And that's the version Google was already reading.

The page you see is content. The page Google reads is math.

Google doesn't read your pages the way you do. It runs math.

It can't tell if your writing is good. It can't judge whether your argument is convincing. It looks for specific signals in specific places - keywords in the title, related terms in the body, schema (machine-readable code that labels what your page is) in the source code, structure that matches what's ranking - and decides where you sit based on which signals hit and which don't.

Good content is what converts visitors once they arrive. The math is what gets them there. You need both.

Most pages have the content. The findings below show how many are missing the math.

How we measure the invisible: Our proprietary POP score

The POP score is a number from 0 to 100. It measures how well your page has been written for Google and AI search - factoring in over 100 on-page signals (schema, semantic coverage, keyword placement, page structure, trust markers) and benchmarked against the pages currently ranking for your target keyword.

It doesn't grade your writing quality. It measures whether the math is there - the signals Google, Claude, ChatGPT, Perplexity, and Google's AI Overviews use to decide what ranks and what gets cited.

The score is powered by the POP Rank Engine™ - a US-patented algorithm built on 400+ controlled SEO experiments I've run over the past six years. Most SEO tools tell you what page-1 averages look like. POP tells you where you need to go to beat them. That's the difference between matching the field and beating it.

Here's what the bands mean:

Score	Band	What it means
0–25	CRITICAL	Invisible in competitive results
26–50	WEAK	Major signals missing, ranking unlikely
51–79	GETTING THERE	Some signals in place, not enough
80–100	STRONG	Built to rank in Google and get cited by AI

A score of 80 is where your content starts moving upwards. Pages below it rarely break through. Pages above it consistently get traction.

KEY FINDING

The average score was 41.6. The median was 33.1. Nearly half (46.8%) scored below 25.

These weren't beginner blogs. They were actively managed pages from people running an SEO tool. The starting line for the engaged segment of the web is less than half of what it takes to rank.

What we found: 7 gaps that show up on almost every page

Seven gaps. Same dataset. Most affect more than half of all pages. Here they are, in order we'll walk through them - starting with the easiest to fix.

The 7 gaps: keyword en H1, keyword en title tag, schema markup, trust signals E-E-A-T, términos semánticos LSI, longitud de contenido, estructura de página

Critical High Medium

1. Exact keyword missing from H1

63.3%

2. Exact keyword missing from title tag

59.5%

3. Missing schema markup

99.0%

4. Missing trust signals (E-E-A-T)

79.8%

5. Missing semantic terms (LSI)

99.4%

6. Content shorter than competitors

58.5%

7. Page structure missing elements

66.2%

New to SEO? E-E-A-T = the signals that show your content can be trusted (real author, real business, credentials). LSI = the related words around your topic that prove you actually cover it. Each finding below explains what to do, and the glossary at the end defines every term.

‍

Free Industry Breakdown

We analysed 11 industries. One averaged a POP Score of just 14.8 out of 100. Where does yours rank?

See how your industry compares, where the biggest gaps are, and which basics competitors still miss.

Full Google Sheet with industry data

All 11 industries ranked by POP score

Exact weak points by industry: schema, trust, headlines, and content depth

See which industries are still under-optimized

Benchmark your industry against the rest

Pulled from the 10,937 web pages in this study. Instant access, free.

Thank you! Your submission has been received.

Oops! Something went wrong while submitting the form.

Jump to results

Which is the best LLM for SEO content?

Get the full rankings & analysis from our study of the 10 best LLM for SEO Content Writing in 2026 FREE!

Get the complete Gsheet report from our study
Includes ChatGPT, Gemini, DeepSeek, Claude, Perplexity, Llama & more
Includes ratings for all on-page SEO factors
See how the LLM you use stacks up

Thank you! Your submission has been received.

Oops! Something went wrong while submitting the form.

Finding 1: H1 and Title Tag Gaps - The Most Overlooked Basics

H1 Tag (Page Headline)

The main visible headline of your web page. There should be exactly one per page. Google uses it as a primary signal to understand what the page is about. If it doesn't include your target keyword or topic, Google has to guess - and it usually gets it wrong.

Title Tag (SEO Title)

The clickable text that appears in Google's search results and in your browser tab. Not always the same as your page headline. Google uses this heavily to determine whether your page matches a search query.

You'd think this would be the easy one. Adding your target keyword to your page headline and title tag is SEO 101. And yet, across 10,937 web pages:

63.3%

of pages missing exact keyword in H1 headline

59.5%

of pages missing exact keyword in title tag

The most common reason? Content written for readability or branding that softens or rewrites the target phrase. A service page optimised for "emergency plumber London" might have the headline "Plumbing Help When You Need It Most." That's good copywriting. It's bad SEO.

WHAT TO DO

Make sure your exact target keyword (or the primary topic phrase) appears in both your page's main headline (H1) and your title tag - ideally near the beginning of each. This single fix can improve rankings within days of re-crawling, particularly on pages ranking on page 2 or 3.

Finding 2: Schema Markup - The Biggest Gap Nobody Talks About

Schema Markup (Structured Data)

Code added to your web page that explicitly tells search engines and AI search systems what your page is - not just what it says. Think of it as labelling your content so Google can understand: "This is a local business. This is written by a qualified professional. This is a product with a price and a review." Without it, search engines have to infer all of this - and they often get it wrong.

Schema isn't a direct ranking factor - adding it won't move you up the SERPs by itself. What it does: helps with indexing, unlocks rich-result real estate (stars, FAQs, breadcrumbs in search), and increasingly helps AI tools decide whether to cite your page.

99%

of pages missing at least 1 recommended schema type

10.4

average number of missing schema types per page

45%

of pages had NO schema detected at all

Schema isn't advanced SEO anymore. It's table stakes - and 99% of pages in our study aren't at the table.

The most-missing schema types aren't exotic rich-result formats. They're foundational:

WHAT TO DO

Start with the foundational types your competitors are already using. POP's rule: if at least 2 of your ranking competitors have a schema type, you probably should too. Foundational markup (Organization, Person, Article, BreadcrumbList, WebSite, ImageObject) covers the gap on most pages. You can generate it inside POP with one click, or use a CMS plugin like Yoast or RankMath.

WHY THIS MATTERS FOR AI SEARCH

When tools like ChatGPT, Perplexity, or Google's AI Overviews decide whether to cite your content, they rely heavily on structured data to understand what your page represents and whether it's trustworthy. A page with no Organization or Person schema is, to an AI, an unsigned document. Schema is how you sign your work.

‍

Which is the best LLM for SEO content?

Get the full rankings & analysis from our study of the 10 best LLM for SEO Content Writing in 2026 FREE!

Get the complete Gsheet report from our study
Includes ChatGPT, Gemini, DeepSeek, Claude, Perplexity, Llama & more
Includes ratings for all on-page SEO factors
See how the LLM you use stacks up

Thank you! Your submission has been received.

Oops! Something went wrong while submitting the form.

Finding 3: E-E-A-T Signals - Google's Trust Filter Is Broken for Most Sites

E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness)

Google's framework for evaluating whether content can be trusted. It asks: Does this content come from someone with real experience? Is the author an expert? Is the website authoritative? In 2026, these signals matter for AI search too - both Google's AI Overviews and tools like ChatGPT use trust signals to decide what content is safe to recommend.

79.8%

of pages missing at least 1 trust/entity schema type

6.3%

of pages had a usable Google entity inventory

Without these signals, a page is essentially anonymous to search engines - and anonymous content is, by default, low-trust content. E-E-A-T doesn't move you up the rankings. It anchors the rankings you already have. If you're missing it, you're not climbing - you're at risk of falling.

WHAT TO DO

Build the basics first. Privacy Policy. Terms of Service. About Us page with real people. Contact details (address, phone, email). SSL certificate. Social links. Organization schema. Author schema for the person who wrote the content.

These won't make you rank - but they'll stop you falling. Most real businesses already have most of this. Make sure search engines and AI tools can see it.

PARTICULARLY BAD FOR: SERVICE, LEGAL, HEALTH AND LOCAL BUSINESSES

Our data shows that legal, home services, and local business pages have the highest trust-schema gaps in the entire dataset. These are precisely the categories where Google applies its strictest trust filters - known as YMYL ("Your Money, Your Life"). If you're in one of these sectors and your schema doesn't declare who you are, what you do, and where you do it, you're operating at a structural disadvantage.

Finding 4: LSI and Semantic Terms - The Language Google Uses to Understand You

LSI Terms / Semantic Entities

LSI (Latent Semantic Indexing) terms are the contextually related words Google expects around your topic. Entities are the specific things Google recognises - people, places, organisations, concepts.

Different mechanics, same job: they give your content context.

Write about a kitchen, and the surrounding terms decide what page you have. "Fridge, oven, dishwasher" makes you an appliances page. "Family, recipes, holidays" makes you a lifestyle blog. Same word - two completely different things, decided by what surrounds it.

Having the word isn't enough. You need the surrounding terms that prove the concept - and that's also how AI tools verify what your page covers and whether it's worth citing.

99.4%

of pages had at least 1 missing LSI/entity term

20.2

average number of missing LSI terms per page

Most pages we analysed weren't writing badly. They were writing about a topic without writing through it.

What's revealing is where these terms are missing. It's not just in the body copy - it's in the supporting signals most content teams ignore entirely:

THE INSIGHT

Modern on-page optimisation is not "put the keyword in the title." It's distributing the right topical language across every meaningful element of a page - including the places content teams almost never think about: image alt text, image file names, link anchor text, and subheadings.

‍

Finding 5: Content Length - More Than Half of Pages Aren't Covering the Topic

Content Length vs. Competitor Target

POP doesn't just count your words - it compares your word count to the word counts of the pages currently ranking for your target keyword. The target isn't arbitrary; it's derived from your actual competitors who are beating you.

2,194

average competitor-derived target word count

1,534

average actual word count of pages in study

58.5%

of pages more than 20% below competitor target

0.59

median current-to-target ratio (59 cents on the dollar)

Longer is not always better. About 25% of pages were more than 20% longer than their competitor target - meaning some content is bloated, not lean. Write the right amount to cover the topic in depth. Not more.

WHAT TO DO

Don't chase a magic word count. Match the range the pages currently ranking for your keyword are in. If they're all at 3,000 words and you're at 500, you're not getting there. If they're at 1,200 and you're at 4,000, you're bloated and outside what Google rewards. The target is the range, not the maximum.

Finding 6: Structural Signals - Formatting That Google Rewards

Page Structure Signals

Beyond the words, Google evaluates the format of your content - how it's organised, how navigable it is, and what elements it uses to help users find answers quickly. Lists, images, tables, FAQ sections, and anchor links all signal to Google that your content is user-friendly and substantive.

Structured content is also disproportionately favoured by AI tools - when ChatGPT, Perplexity, or Google's AI Overviews need to extract an answer, well-formatted lists, FAQs, and tables get pulled first.

Structure isn't just "more is better." Over half of pages in our study had too many H2 subheadings - creating heading bloat that confuses both users and Google. The goal is matching the structure of pages currently ranking. Not piling on elements your competitors don't have.

WHAT TO DO

Add what your competitors have that you don't: more internal links, more images with proper alt text, unordered lists, an FAQ section. Cut what they don't need: H2 bloat, paragraph bloat. The goal is to match the pattern of what's ranking - not max out any single element.

Finding 7: Page Type Breakdown - Your Money Pages Are Your Worst Pages

YOUR MONEY PAGES ARE YOUR WORST PAGES

Category and landing pages - the most commercially important page type in any business's site - have the lowest average content score of all at 33.6. These are the pages driving traffic to your products and services. They have the worst H1 gaps (69.4%) and the highest short-copy rate (68.6%) of any page type. If you only optimise one page type on your site, start here.

Free Industry Breakdown

We analysed 11 industries. One averaged a POP Score of just 14.8 out of 100. Where does yours rank?

See how your industry compares, where the biggest gaps are, and which basics competitors still miss.

Full Google Sheet with industry data

All 11 industries ranked by POP score

Exact weak points by industry: schema, trust, headlines, and content depth

See which industries are still under-optimized

Benchmark your industry against the rest

Pulled from the 10,937 web pages in this study. Instant access, free.

Thank you! Your submission has been received.

Oops! Something went wrong while submitting the form.

Summary: The bar to beat the web is shockingly low

10,937 web pages. Seven categories of gap. One consistent story: the standard of content online, even among people actively trying to get it right, is far lower than it needs to be.

That's genuinely good news. You don't need to be exceptional to outperform most of the web. You just need to be technically correct - and that's entirely learnable.

THE CORE INSIGHT FROM THIS STUDY

Most content fails not because of bad writing, but because it's missing the machine-readable signals Google and AI tools use to understand, trust, and promote it. Fix the signals, and the writing you already have performs significantly better.

Your content isn't optimised for Google and LLMs unless you've measured it against what's actually ranking. That's the gap. That's the opportunity. That's what PageOptimizer Pro was built to do.

How to fix your page. Your action checklist:

Put your exact keyword in your H1 and title tag. The fastest single fix in SEO. Affects 63% of pages.
Add the foundational schema types - at minimum, Organization (who you are), Person (who wrote it), Article (what kind of content this is), BreadcrumbList (where this page sits on your site). 99% of pages were missing at least one type their competitors already use.
Build the E-E-A-T basics - Privacy Policy, About Us, Contact, author and Organization schema. These don't make you rank. They stop you falling.
Distribute semantic terms everywhere - alt text, image filenames, anchor text, subheadings. Not just the body copy. Average page missing 20 semantic terms.
Match competitor word count. There's no magic number - aim for the range your ranking competitors are in. 58.5% of pages are more than 20% below that target.
Add the structural elements your competitors use - internal links, images, lists. Cut what they don't: H2 bloat, paragraph bloat.
Measure against what's actually ranking. Not instinct. Not best-practice guides written in 2019. The pages currently in the top 10 for your keyword.

IF YOU'RE USING AI TO WRITE YOUR CONTENT

The gap this study describes is wider for AI-written content, not narrower. Earlier this year we ran 13 major LLMs - including ChatGPT, Claude, Gemini, and Perplexity - through POP's on-page scoring system. The average model scored 58.4 out of 100 on a standard SEO prompt. Only one entry broke 80.

AI tools can write. They still can't optimise what they write. If your content is AI-generated, the gaps in this study are almost certainly present in your pages too.

→ Read the full LLM study

How to close the gap with POP

Everything above - the H1 gaps, the schema gaps, the missing semantic terms, the structural issues - is exactly what POP identifies and tells you how to fix. Your first POP report takes under five minutes. No technical setup. Five steps.

50% of my 377 articles that we used POP on are in the Top 5 in the SERPs

Within a few weeks of using POP, pages that had been stuck on page 2 and page 3 of the SERPs for over a year began moving to page 1. I couldn't believe it! As of right now, 50% of my articles that we used POP on are in the Top 5 in the SERPs, this is 370 articles, of which 122 are in Position 1. So if you want to start making money and getting more traffic to your website, you have to get POP.

Paul L

Create your POP account

Head to PageOptimizer Pro and sign up. No setup, no configuration. You're ready in minutes.

Submit your content

Enter the URL of the page you want to optimise, or paste your content directly. POP analyses your page against what's actually ranking for your target keyword.

Get your POP score and your to-do list

POP gives your content a score out of 100 and generates a list of exactly what to fix - headings, keyword placement, schema, semantic terms, page structure, and more. No guesswork. Just a specific list of changes.

Make the updates and publish

Work through the recommendations and republish. Pages that hit a strong POP score consistently see better rankings, more AI citation visibility, and more organic traffic.

Re-score and refine

SEO isn't one-and-done. As competitors update their content and the algorithm evolves, POP helps you keep your pages optimised over time.

PREFER TO CREATE CONTENT FROM SCRATCH?

POP AI Writer builds articles optimised from the first word - not content you have to fix after the fact. The only AI system in our 2026 LLM study that scored 100 - every other major model failed to break 80.

→ See POP plans

RATHER HAVE SOMEONE DO IT FOR YOU?

That's what POP White Glove is for. Our team runs the entire optimisation process on your behalf - audits, schema, rewrites, reporting, the lot. Fully white-labelled for agencies.

→ Learn more about POP White Glove

FAQ

Because Google doesn't rank writing - it ranks signals. Our study of 10,937 audits found the median page scored 33 out of 100 against what's actually ranking. Good writing converts visitors once they arrive, but it's the machine-readable layer - schema, semantic terms, structure - that gets them there. Most pages have the writing. They're missing the signals.

Missing the exact keyword where it counts most. In our data, 63.3% of pages didn't have their target keyword in the H1, and 59.5% were missing it from the title tag. It's the most basic signal in SEO, and the majority of pages still get it wrong - usually because the headline was written for branding instead of search.

Not directly - and this trips people up. Schema won't move you up the rankings by itself. What it does: helps Google index and understand your page, unlocks rich results, and increasingly decides whether AI tools like ChatGPT and Perplexity cite you. 99% of the pages we analysed were missing at least one recommended schema type. It's not a ranking lever, but its absence makes you invisible to AI search.

Not the writing itself - the optimisation behind it. In 2026 we tested 13 AI models on a standard SEO writing task. The average scored 58 out of 100, and the best mainstream LLM topped out at 72.8 - below the 80 mark where content starts to move. Not one general-purpose model broke 80. AI tools can write. They can't yet reliably optimise what they write.

There's no magic number. The right length is the range the pages currently ranking for your keyword are in. In our study, 58.5% of pages were more than 20% below their competitor-derived target - but about 25% were too long. Match the field. More isn't better; right is better.

E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness) is how Google judges whether your content can be trusted. Here's the nuance most people miss: it doesn't push you up the rankings - it stops you falling. 79.8% of the pages we analysed were missing at least one trust signal.

Because exact-match keywords are only part of the picture - and the smaller part. 99.4% of the pages in our study were missing the semantic terms that prove genuine topical depth: the related words, entities, and concepts Google expects around your subject. You can repeat your keyword a hundred times; if the supporting language isn't there, Google still can't tell you actually cover the topic.

Your money pages. Category and landing pages - the ones that drive revenue - scored worst in our entire study, averaging 33.6 out of 100. Everyone pours effort into blog posts and neglects the commercial pages. Start where the money is.

Measure it against what's actually ranking - not against your instinct or a checklist from 2019. The only reliable signal is a side-by-side comparison with the pages currently winning for your keyword. If you're not measuring against them, you're guessing.

Both - and arguably more for AI. When ChatGPT, Perplexity, or Google's AI Overviews decide what to cite, they lean heavily on structured data and clear topical signals. The same gaps that keep you off page one keep you out of AI answers. A page with no schema and thin semantic coverage is, to an AI, an unsigned document it has no reason to quote.

Glossary

New to SEO or AEO? Plain-English definitions for every technical term used in this article.

POP Score

An overall score from 0 to 100 measuring how well a page is optimised for Google and AI search. It factors in over 100 on-page signals. Pages below 80 are unlikely to rank in competitive niches.

H1 Tag

The main visible headline of a web page. Should be exactly one per page. Google uses it as a primary signal to understand what the page is about. Should include your target keyword.

Title Tag (SEO Title)

The text that appears in Google's search results as the clickable headline and in your browser tab. A critical ranking signal that should include your target keyword.

Schema Markup

Code added to a web page that tells search engines and AI tools what the page is - not just what it says. Helps with indexing and rich-result eligibility, and increasingly with AI search citations.

E-E-A-T

Experience, Expertise, Authoritativeness, Trustworthiness - Google's framework for evaluating content quality. Doesn't directly move rankings up, but anchors the rankings you've already gained.

LSI Terms / Semantic Entities

The related words, concepts, names, places, and topic vocabulary that naturally belong around a subject. Used by Google and AI tools to verify that your content genuinely covers a topic.

AEO (Answer Engine Optimisation)

Optimising your content to be cited, quoted, and recommended by AI answer tools like ChatGPT, Perplexity, and Google's AI Overviews.

YMYL (Your Money, Your Life)

Google's category for content that could significantly impact a person's health, financial stability, or safety. Pages in finance, health, and legal sectors are held to stricter quality standards.

Entity

In the context of search, an entity is a uniquely identifiable thing - a person, business, location, or concept - that Google has an understanding of. Establishing your brand as a recognised entity improves trust and AI search visibility.

Image Alt Text

A written description of an image added in the page code. Used by screen readers for accessibility and by Google to understand what an image depicts. A frequently overlooked place to include semantic terms.

Kyle Roof

Kyle Roof is a Co-Founder & Lead SEO at POP, SEO expert, speaker and trainer. Kyle currently resides in Chiang Mai, Thailand with his family.

Questions or comments? Visit our Help Center for support.

Table of Contents

Finding 1: H1 and Title Tag Gaps - The Most Overlooked Basics

Finding 2: Schema Markup - The Biggest Gap Nobody Talks About

Finding 3: E-E-A-T Signals - Google's Trust Filter Is Broken for Most Sites

Finding 4: LSI and Semantic Terms - The Language Google Uses to Understand You

Finding 5: Content Length - More Than Half of Pages Aren't Covering the Topic

Finding 6: Structural Signals - Formatting That Google Rewards

Finding 7: Page Type Breakdown - Your Money Pages Are Your Worst Pages

Summary: The bar to beat the web is shockingly low

We Built POP for Query Fan-Out Before It Had a Name

We Tested 13 New “Best-for-Writing” LLMs for SEO Content in 2026

Educational infographic showing a digital brain symbolizing a large language model emerging from a monitor, with icons for keyword analysis, content optimization, semantic search, ranking algorithms, and user intent, illustrating LLMs' role in modern SEO.

LLM SEO Explained: What LLM Means in SEO Context

Why Isn't My Content Ranking in Google and AI?

About this study

The page you see is content. The page Google reads is math.

How we measure the invisible: Our proprietary POP score

What we found: 7 gaps that show up on almost every page

We analysed 11 industries. One averaged a POP Score of just 14.8 out of 100. Where does yours rank?

Which is the best LLM for SEO content?

Finding 1: H1 and Title Tag Gaps - The Most Overlooked Basics

Finding 2: Schema Markup - The Biggest Gap Nobody Talks About

Which is the best LLM for SEO content?

Finding 3: E-E-A-T Signals - Google's Trust Filter Is Broken for Most Sites

Finding 4: LSI and Semantic Terms - The Language Google Uses to Understand You

Finding 5: Content Length - More Than Half of Pages Aren't Covering the Topic

Finding 6: Structural Signals - Formatting That Google Rewards

Finding 7: Page Type Breakdown - Your Money Pages Are Your Worst Pages

We analysed 11 industries. One averaged a POP Score of just 14.8 out of 100. Where does yours rank?

Summary: The bar to beat the web is shockingly low

How to fix your page. Your action checklist:

How to close the gap with POP

Create your POP account

Submit your content

Get your POP score and your to-do list

Make the updates and publish

Re-score and refine

FAQ

Glossary

Related articles:

Read next:

Advanced on page SEO made simple.