Our Top 5 Blog Posts of 2025 (And What Made Them Work)

AI Overviews Reduce Clicks by 34.5% was our most-viewed article overall, driving ~96,000 visits since it was published back in April, according to Ahrefs Web Analytics.

This research tapped into our first-party data to tackle the question every SEO has been asking this year: how much traffic are AI Overviews actually stealing?

We analyzed 300,000 keywords and found that when an AI Overview appears, the top-ranking page sees 34.5% fewer clicks compared to similar queries without one.

By comparing clickthrough rates before and after Google's AI Overview rollout, we measured the real impact: position #1 for informational keywords dropped from a 0.056 CTR to just 0.031 after AI Overviews appeared.

The data confirmed what many of us suspected—AI Overviews are significantly reducing traffic to top-ranking pages, even when those pages still hold the #1 spot.

An Analysis of AI Overview Brand Visibility Factors (75K Brands Studied) came in at number two. I wrote this piece of research to confirm, once and for all, whether PR and off-site brand mentions were the most important factor when it came to AI visibility.

It was a theory I kept hearing everywhere at the time, but I (like others, it turns out) wanted to see the cold hard data. Together with Ahrefs’ Data Scientist, Xibeijia, I analyzed 75,000 brands to assess how eleven different branded search factors correlated with AI Overview visibility—we’ve since repeated this study for ChatGPT and AI Mode.

Our biggest finding? Branded web mentions do correlate strongly (0.664) with AI Overview presence—more so than any other factor, including backlinks or domain rating.

We also found that if your brand sits in the bottom 50% of web mentions, you're essentially invisible to AI systems.

AI Content Creation: My Process for High-Quality, SEO-Friendly Articles was our third most-viewed blog, racking up ~30K views.

My teammate Despina Gavoyannis created this practical SOP (standard operating procedure) for marketers wanting to integrate AI into their content workflows.

This guide walks through her exact process for using AI as a tool—not a replacement—for creating content that ranks.

The post covers everything from finding rankable keywords and identifying search intent, to adding original info that AI can't generate on its own.

Despina's SOPs have performed especially well via Google Discover traffic…

Google Discover users seem to engage deeply with practical, step-by-step SOPs that help them complete real tasks.

What Is llms.txt, and Should You Care About It?: Ryan wrote this explainer in just a couple of hours using ChatGPT. Now it’s our fourth most-viewed blog of 2025.

Not bad ;)

The post breaks down llms.txt, a proposed standard designed to help large language models understand and access structured content from websites.

While some developers are experimenting with it, no major LLM provider currently supports it as part of their crawler protocol.

Ryan's take: It's a solution in search of a problem.

Search engines already crawl and understand content using existing infrastructure like robots.txt and sitemap.xml—LLMs use much of the same setup.

In 2025, Ryan built a seven-part AI content process and has since created 6+ wholesale AI blogs, which, together, have driven ~63,000 views to date.

You can watch how he does it on the Ahrefs Podcast below.

It just goes to show that AI generated content really can perform well. Use Ryan’s process to tackle trending and emerging topics before your competitors do.

Google Says “Links Matter Less”—We Looked at 1,000,000 SERPs to See if It's True: Google has been telling SEOs that links are less important than they used to be, and many in the industry have accepted this as fact.

But is it actually true?

Patrick Stox, Ahrefs' Technical SEO expert, decided to settle the debate with data.

He analyzed the top 1 million keywords by search volume, calculating Spearman correlations between SERP rankings and 20 different SEO metrics including domain rating, backlinks, referring domains, and internal links.

And, guess what? Links still matter—a lot.

While the correlation dropped slightly since a similar 2019 study (from 0.27 to 0.21 for total backlinks), links remained one of the strongest ranking signals, especially for high-volume, competitive queries.

The research also showed that links matter even more for informational content, local queries, and branded searches.

So while Google may be hedging on the importance of links publicly, the data tells a different story.

Similarly to the SparkToro team, most of our top content in 2025 was research-based.

In fact, 56% of our most-viewed posts (across every channel) were research blogs.

Our top 5 articles also tended to differ slightly from source to source.

Here are our top performers across organic search, AI search, direct, and social media.

Top blogs by search traffic

	Blog	Views	Type
1	AI Content Creation: My Process for High-Quality, SEO-Friendly Articles	20.6K	SEO blog
2	AI Overviews Reduce Clicks by 34.5%	19.5K	Research
3	What Is llms.txt, and Should You Care About It?	15K	AI generated
4	How to Rank for a Keyword (8 Steps)	12.8K	SEO blog
5	Google Says “Links Matter Less”—We Looked at 1,000,000 SERPs to See if It’s True	11.4K	Research

Top blogs by AI traffic

	Blog	Views	Type
1	AI Overviews Reduce Clicks by 34.5%	3.3K	Research
2	An Analysis of AI Overview Brand Visibility Factors (75K Brands Studied)	1.8K	Research
3	74% of New Webpages Include AI Content (Study of 900k Pages)	~1K	Research
4	What Is llms.txt, and Should You Care About It?	~900	AI generated
5	The Best AI Content Optimization Tools for 2025	~800	Best list

Top blogs by direct traffic

	Blog	Views	Type
1	AI Overviews Reduce Clicks by 34.5%	52.3K	Research
2	An Analysis of AI Overview Brand Visibility Factors (75K Brands Studied)	17.8K	Research
3	63% of Websites Receive AI Traffic (New Study of 150k Sites)	9.7K	Research
4	The Great Decoupling (or Why Your Clicks Are Down While Impressions Are Up)	9.3K	Thought leadership
5	The Best AI Content Optimization Tools for 2025	8.6K	Best list

	Blog	Views	Type
1	AI Overviews Reduce Clicks by 34.5%	7.4K	Research
2	The 50 Fastest-Growing Companies (January 2025)	4.5K	SEO blog / Research
3	An Analysis of AI Overview Brand Visibility Factors (75K Brands Studied)	4.3K	Research
4	Google Says “Links Matter Less”—We Looked at 1M Pages	4K	Research
5	100 Most Expensive Keywords for Google Ads in 2025	3.9K	SEO blog / Research

Final thoughts

Here’s what we learned from our top blogs of 2025…

Test claims people are skeptical about: When official narratives don't match what people see IRL, data-driven research will get you attention. That’s why Patrick’s research was so widely viewed—it provided hard data that challenged Google's narrative.
Pick debates people care about: The best research answers questions your audience is actively debating, not ones you think they should care about.
Get in early on trending topics: Many of these posts worked because they jumped on trending topics early. Some didn't even have search demand when we started writing about them—we just kept our ear to the ground on social media and in communities, and wrote about the topics driving discussion. Then, over time, those topics naturally gained search volume. That’s why 4 of our top 5 articles were also our most-viewed articles in search.
Don't be afraid of AI-assisted content: One of our top posts was whipped up in an afternoon with ChatGPT, and still pulled in huge amounts of traffic. For timely topics or straightforward explainers, AI can get you 80% of the way there—just add your edits, pepper in some original examples, and ship it.
Give people SOPs they can steal: Not everything needs to be a data study. Step-by-step processes people can actually follow and adapt for their own work consistently perform well.
Use data nobody else has access to: Three of our top five posts used Ahrefs' first-party data to answer questions no one else could. If you have proprietary data or access to information that other people don't—use it. Whether that's customer surveys, internal sales metrics, support ticket trends, user behavior analytics, financial data, or even aggregated insights from client work—unique data gives you a competitive moat that's impossible to replicate.

When it comes to AI, the only certainty is uncertainty. What worked in 2025 might not work in 2026. But that's exactly why we'll keep sharing what we learn!

Post byLouise Linehan

Louise is a Content Marketer at Ahrefs. Over the past ten years, she has held senior content positions at SaaS brands: Pi Datametrics, BuzzSumo, and Cision. During this time, she has published hundreds of blog posts, spearheaded industry-leading research, and developed expert-led webinar programs.

Louise first learned the ropes of SEO during her time at enterprise SEO company, Pi. Within weeks of joining, she had published a piece on cannibalization that caught the attention of the biggest name in SEO at the time: Rand Fishkin. Since then, she has written on topics ranging from manipulative link building, to content syndication, and search demand forecasting. In her spare time, she has deepened her practical understanding of SEO by building websites for two local businesses.

You can read Louise's words in industry publications like Search Engine Watch, The Drum, The Telegraph, and Spin Sucks.