Content Marketing for Lead Generation Broke in 2025. Here's What Replaced It.

What Changed: The Click Stopped Being the Unit of Value

Ahrefs has measured twice how Google's AI Overviews affect clicks. Its first study, published in April 2025, found that an AI Overview cut the click-through rate for the top-ranking page by 34.5%. When Ahrefs re-ran the analysis on 300,000 keywords using December 2025 data, the drop had grown to 58%.

Pew Research Center reached the same conclusion with a different method. It tracked the real browsing behavior of 900 US adults across 68,879 Google searches. Users clicked a regular link in the search results 8% of the time when an AI Overview appeared, compared with 15% when it didn't. Only 1% clicked a link inside the AI Overview itself.

For your business, this means your rankings can hold steady while organic traffic drops sharply, and your rankings report won't explain why.

Why Fewer Clicks Can Still Mean More Quality Leads

That doesn't mean content has stopped generating leads. The visitors who still arrive, especially from AI tools, tend to be worth far more than the ones you lost.

Semrush studied more than 500 high-value digital marketing and SEO topics in June 2025. It found that visitors from AI search converted at about 4.4 times the rate of traditional organic visitors. Ahrefs saw the same pattern in its own traffic around the same time: AI referrals made up just 0.5% of sessions but drove 12.1% of signups.

An AI engine does the comparison work before the click happens. By the time someone reaches your page from ChatGPT or Perplexity, they've read a summary, narrowed their options and chosen to learn more about you specifically. A regular search click usually starts the research. An AI-referred click usually comes near the end of it.

The catch is volume. Conductor's 2026 AEO/GEO Benchmarks Report analyzed 3.3 billion sessions from May to September 2025. It found AI referrals made up only 1.08% of total traffic on average.

One warning before these numbers go into a board deck: the studies define "conversion" differently. Some count form fills, some count signups, and some count purchases. Mixing them produces a number that looks authoritative but means nothing.

Most of the Buying Decision Happens Where You Can't See It

The traffic shift matters even more once you look at how B2B buyers actually buy.

6sense's 2025 Buyer Experience Report, a survey of more than 4,000 B2B buyers, found that buyers first contact a vendor about 61% of the way through their buyer's journey. That's earlier than the 69% of past years, but most of the work still happens before a vendor hears from them. By then, 94% have already ranked their shortlist, and roughly four out of five buy from the vendor they favored going in.

Buyers also want less contact with sales, not more. A Gartner survey of 646 B2B buyers, published in March 2026, found that 67% prefer a buying experience with no sales rep involved. That's up from 61% in Gartner's previous survey.

So when a buyer finally hands over an email address, the decision is mostly made. The form fill marks when they chose to identify themselves, often weeks or months after your content did the persuading. AI assistants extend that invisible phase. Research that once spanned ten browser tabs now happens inside one conversation, which often sends no referral data to your site at all.

If buyers form their preferences while anonymous, the brand awareness you build during their research matters more than any single conversion on a landing page.

What AI Engines Actually Cite

A closer look at AI citation studies point to three consistent patterns.

Position. Kevin Indig analyzed 1.2 million ChatGPT answers and found that 44.2% of citations came from the first 30% of a page. If the key point of a piece of content is buried under a long introduction, the model may never use it.

Evidence. The research paper that introduced "generative engine optimization," presented at KDD 2024, tested nine ways of rewriting content across 10,000 queries. In those controlled tests, adding statistics, quotations and cited sources increased a page's visibility in AI answers by 30% to 40%. Clearer writing helped by 15% to 30%. A more authoritative tone did little, and keyword stuffing didn't help at all.

Focus. Search engines rewarded comprehensive pages, which pushed everyone toward 3,000-word ultimate guides. ChatGPT doesn't. In a study of 815,000 query-page pairs, Indig and AirOps found that focused content pieces covering two or three related angles outperformed exhaustive guides. The citation sweet spot was between 500 and 2,000 words.

Here's how that changes the playbook:

For AI engines, valuable content is content where every section makes sense on its own, because they lift sections out of pages and use them independently.

The Part Almost Nobody Does: Publish Original Data

Search "AI referral conversion rate" and you'll find article after article recycling the same handful of studies. When an AI engine can cite the original research directly, those roundups have little to add.

The strongest content marketing ideas start with a number nobody else has. You don't need a research department to get one. You need to measure what you already run:

• Your performance data, anonymized. Strip out customer information and report conversion rates by vertical, traffic source and content format
• Your AI visibility tracking. Measure citation share monthly across ChatGPT, Perplexity, Copilot, Grok, AI Overviews and AI Mode, long enough to show a trend
• Your market research. A survey of 400 buyers in your category produces a statistic others will cite
• Your failures. Negative findings get cited often because so few companies publish them

Here's a great example of what that looks like in practice. Between March 24 and September 24, 2026, we tracked AI citations for seven review and comparison sites, including one of our own, using Ahrefs data. Perplexity responses citing these sites grew at all seven, by 37% to 229%. ChatGPT responses citing them fell at six of the seven, with a median drop of 48%, and three sites lost 85% or more.

Methodology: Citation counts from Ahrefs Brand Radar, comparing ChatGPT and Perplexity responses that cited each site between March 24 and September 24, 2026. Sites were selected as direct competitors in the review and comparison category.

We saw something similar at the page level on one of our own review sites. Over [three-month window, add dates], Google's AI features report showed our antivirus section earning 44% of the site's AI impressions. Nearly all of it came from 257 product-versus-product comparison pages, averaging about 12 impressions each. Our auto warranty section uses the same format across 82 pages, yet those pages averaged about 2. Same site, same template, roughly five times the difference per page. The format alone doesn't earn the citation.

No roundup can offer findings like these. That's what separates content marketing leaders from everyone else: content other people have to cite to make their own argument.

Build the Strategy Before You Commission a Single Article

The Content Marketing Institute's 2026 B2B research, a survey of more than 1,000 B2B marketers, shows the gap between having a content strategy plan and having one that works. Ninety-seven percent said they have a content strategy, but only 12% said their marketing was highly effective over the past year, a documented strategy no longer sets you apart. The thinking inside it does. Among marketers who rated their programs effective, 65% credited content relevance and quality, more than any other factor, including team skills, sales alignment, and tools.

Relevance comes from researching the needs of your audience, not just your topic. Keyword volume tells you what people type. It doesn't tell you what keeps them up at night. Build buyer personas and your storytelling techniques for your target audience from sales calls, support tickets, churn interviews, and review sites, then map each persona's pain points to the stage of the buyer’s journey and the customer journey where they come up, ensuring you incorporate relevant content.

Different types of content do different jobs at each stage. The right type of content at the wrong stage is wasted.

Most teams flood the top of the funnel with awareness blog posts and underinvest in validation, where one strong white paper or detailed case study can rescue a stalled deal. In that same Gartner survey, buyers who felt confident in their decision were twice as likely to report a high-quality deal. Content that builds that confidence, and helps your internal champion win the argument inside their own company, is some of the highest-value work a content team can do.

Measure the Whole Path, or You Will Kill Your Best Work

Most marketing dashboards were built to measure marketing campaigns on a web where search engines sent the traffic. Track these instead to see what actually helps boost lead generation:

• Citation share by platform, monthly. Visibility in AI answers moves before referral traffic does
• AI referrals as their own GA4 segment. This traffic is often misattributed as direct
• Conversion rate by source, not blended. A page can lose 40% of its sessions and still gain revenue if the remaining visitors are higher quality
• Assisted influence, not just last touch. Last-touch attribution credits the moment someone left their contact details and ignores the content that created the preference
• Time to shortlist. If buyers choose a favorite before contacting you, what matters is whether you made the list

What to Stop Doing

Publishing for volume. AI tools made content cheap to produce, so volume no longer sets you apart. Instead, a great way to ensure success is one page that becomes the reference on its topic, which beats four thin posts a month.

Treating AI visibility as an SEO subtask. Ahrefs found that only 12% of the links cited by ChatGPT, Gemini and Copilot appear in Google's top 10 results for the same prompt. They are related channels, not the same one.

Gating your best material by default. If a form stands between an AI crawler and your strongest evidence, AI engines can't cite it. Publish the findings, and gate the tools and templates instead.

Frequently asked questions

Is content marketing still effective for lead generation?

Yes, but it works differently now. Content generates leads by earning citations in AI answers and by converting fewer, higher-intent visitors. Semrush found that AI search visitors converted at about 4.4 times the rate of traditional organic visitors.

What kind of content gets cited by AI engines?

Focused pages that answer the question in the opening lines, attribute claims to named and linked sources, and publish original data. Kevin Indig's analysis of 1.2 million ChatGPT answers found 44.2% of citations came from the first 30% of the page.

What are the benefits of content marketing compared with paid acquisition?

Content compounds and paid doesn't. A page that earns citations keeps working after the spend stops, and it reaches buyers during the anonymous research phase where paid channels have limited reach. The trade-off is time, so judging a content program on a single quarter usually leads to the wrong decision.

Should I still do traditional SEO?

Yes. AI engines pull from search indexes, so ranking is still the entry ticket. Ahrefs found that 76% of AI Overview citations come from pages in Google's top 10. We've seen this firsthand. One of our sites has 51 payroll and PEO pages, full coverage of the category. They rank on pages five through nine of Google, and they earned almost no AI Overview impressions. Coverage without authority doesn't get cited.

How often should content be updated?

Quarterly for anything that includes statistics. When Ahrefs analyzed 17 million citations from AI assistants, the pages they cited were 25.7% fresher on average than pages ranking in Google's organic results. An outdated figure is worse than no figure, because it undermines the credibility of everything around it.

Need more qualified leads? We deliver them.

Get Started