Study: Do Backlinks Matter in AI Search?
Backlinks have been the backbone of Google SEO for over two decades. But do they matter when ChatGPT, Claude, or Gemini decide which businesses to recommend? We analyzed 8,000+ AI responses and cross-referenced them with backlink profiles to find out.
The End of the Link Graph Era
Google's original breakthrough was PageRank, an algorithm that treated every backlink as a vote of confidence. More links from better sites meant higher rankings. This single innovation shaped two decades of digital marketing strategy, spawning an entire industry of link building services, guest posting agencies, and outreach tools.
But AI search engines do not work like Google. ChatGPT does not crawl the web and count links. Claude does not maintain a link graph. Gemini, while connected to Google's infrastructure, processes queries through a fundamentally different mechanism than traditional search.
This raises an urgent question for every business invested in SEO: if backlinks do not directly drive AI search rankings, what does? And should businesses continue investing in link building at all?
Study Design and Methodology
We designed this study to measure the relationship between traditional backlink metrics and AI citation frequency. Here is how we structured the research:
We identified 500 businesses across 25 industries, ranging from local service providers to global enterprises. For each business, we collected comprehensive backlink data using Ahrefs, including total referring domains, domain rating, and backlink quality scores.
We then ran 8,312 prompts across ChatGPT-4o, Claude 3.5 Sonnet, and Gemini 1.5 Pro. Prompts were designed to elicit business recommendations in each industry and geographic market. We tracked which businesses were mentioned, recommended, or cited in each response.
Finally, we performed correlation analysis between backlink metrics and AI citation frequency, controlling for confounding variables such as brand age, review volume, and content quality.
Key Finding: Backlinks Show Weak Direct Correlation
The headline finding is striking: the correlation between total referring domains and AI citation frequency was only 0.12. In statistical terms, this means backlinks explain approximately 1.4% of the variance in whether a business gets cited by AI.
For context, a correlation of 0.12 is considered negligible in social science research. Compare this to the correlation between backlinks and Google search rankings, which various studies have estimated at 0.30-0.45, a moderate to strong relationship.
Domain Rating, Ahrefs' aggregate authority metric, showed a slightly higher correlation at 0.19, but this is still weak. The relationship between traditional link authority and AI citation is far weaker than the SEO industry commonly assumes.
The Indirect Effect: What Backlinks Actually Signal
While the direct correlation is weak, we identified an important indirect effect. Backlinks correlate with other factors that do influence AI citations:
- Brand mention frequency (0.47 correlation with AI citations): Businesses with more backlinks tend to have more brand mentions across the web, and brand mentions are a strong AI citation signal.
- Content depth and quality (0.38 correlation): Sites that earn natural backlinks tend to have higher quality content, which independently influences AI citations.
- Domain age and trust (0.29 correlation): Older domains with extensive backlink histories tend to be more established brands, which LLMs recognize through training data.
The takeaway is that backlinks do not directly cause AI citations, but the activities that generate natural backlinks, such as creating valuable content, building brand awareness, and establishing industry authority, do influence AI visibility.
What Replaced Backlinks in AI Search
If backlinks are not the primary signal, what is? Our analysis identified five factors with substantially stronger correlations to AI citation frequency.
1. Entity Recognition (0.61 correlation)
The strongest predictor of AI citation was entity recognition, meaning whether the business is recognized as a distinct entity in the LLM's knowledge base. Businesses that appeared in knowledge graphs, had Wikipedia pages, or were consistently mentioned across multiple authoritative sources showed the highest citation rates.
2. Brand Mention Frequency (0.47 correlation)
The total number of times a brand is mentioned across authoritative web sources, regardless of whether those mentions include links, showed a strong correlation with AI citations. This is a fundamental shift from SEO, where an unlinked mention has historically been considered worthless.
3. Review Profile Strength (0.43 correlation)
Businesses with strong review profiles across multiple platforms, measured by review volume, average rating, and platform diversity, were significantly more likely to be recommended by AI.
4. Content Structure Quality (0.38 correlation)
The structural quality of a business's content, including clear headings, comprehensive coverage, expert attribution, and schema markup, showed a moderate-to-strong correlation with AI citations, independent of backlink metrics.
5. Topical Authority Depth (0.35 correlation)
Businesses that demonstrated deep expertise in a specific topic, measured by the number of comprehensive pages covering related subtopics, performed better in AI citations than businesses with broad but shallow content.
Industry-Specific Findings
The relationship between backlinks and AI citations varied by industry:
B2B SaaS: Backlinks showed the strongest correlation (0.24) in the B2B SaaS sector, likely because the software review ecosystem (G2, Capterra) creates a link-rich environment that overlaps with AI-trusted review data.
Local services: Backlinks showed almost zero correlation (0.04) for local service businesses. AI recommendations for local services were driven almost entirely by review data and local directory consistency.
Professional services: A moderate correlation (0.18) existed for professional services firms, but further analysis showed this was entirely mediated by the brand mention effect. Firms with more backlinks had more brand mentions, and it was the mentions, not the links, driving AI citations.
E-commerce: A weak correlation (0.09) existed for e-commerce brands. Product review data, brand recognition, and structured product information were far more predictive than backlink profiles.
The Rise of Unlinked Mentions
Perhaps the most paradigm-shifting finding in this study is the value of unlinked brand mentions. In traditional SEO, an unlinked mention has been considered essentially worthless because Google's algorithm primarily evaluates hyperlinks, not text references.
In AI search, unlinked mentions are nearly as valuable as linked ones. When a respected industry publication mentions your brand by name in an article, even without linking to your website, that mention becomes part of the training data and real-time browsing context that LLMs use to evaluate your authority.
This finding should fundamentally shift how businesses approach PR and digital marketing. A mention in a major industry publication is valuable for AI visibility regardless of whether it includes a link. Guest posts on authoritative blogs create AI authority even if the links are nofollowed. Podcast appearances where your brand is mentioned generate AI signals that backlinks cannot.
What This Means for Your SEO Strategy
We are not suggesting businesses should abandon link building entirely. Backlinks still matter for Google rankings, and Google traffic remains a significant channel for most businesses. However, the strategic allocation of resources should shift.
Shift #1: From Link Acquisition to Brand Mention Acquisition
Instead of focusing exclusively on getting links, focus on getting mentioned. This means PR campaigns, expert commentary, podcast appearances, conference speaking, and industry partnerships that generate brand mentions across authoritative platforms.
Shift #2: From Link Quality to Source Authority
Not all mentions are equal. A mention on a platform that LLMs trust, such as established news outlets, industry-specific publications, and high-authority review sites, is worth more than a mention on a low-traffic blog, even if the blog would have provided a valuable backlink for traditional SEO.
Shift #3: From Anchor Text to Entity Association
In traditional SEO, the anchor text of a backlink signals relevance. In AI search, what matters is the context in which your brand is mentioned. Being mentioned alongside industry keywords, competitor names, and topic-relevant discussions creates entity associations that LLMs use when generating recommendations.
Shift #4: From Link Volume to Mention Diversity
In Google SEO, more links from unique domains generally meant better rankings. In AI search, the diversity of platforms where your brand is mentioned matters more than the volume of mentions on any single platform. Being mentioned across news, reviews, directories, social media, and industry publications creates a more robust authority signal than concentrated mentions from a single source type.
Practical Recommendations
Based on this research, we recommend the following adjustments to digital marketing strategy:
- Audit your brand mentions: Use tools like Brand24 or Mention to understand where your brand is already mentioned. Identify gaps in authoritative source types.
- Invest in PR: Earned media coverage creates AI authority signals that paid link building cannot replicate. Focus on publications that LLMs trust.
- Build review diversity: Ensure your business has active profiles on at least three relevant review platforms with consistent high ratings.
- Create citable content: Publish original research, industry data, and expert analysis that other sources want to reference, generating both links and unlinked mentions.
- Optimize for entities: Ensure your brand is consistently described across all platforms with the same name, category, and key attributes to strengthen entity recognition.
Conclusion
Backlinks are not dead, but their role has fundamentally changed. In the AI search era, links are one small signal among many, and far from the most important one. The businesses that will win in AI search are those that build broad, authoritative brand presence across the platforms that LLMs trust, regardless of whether that presence comes in the form of links, mentions, reviews, or structured data.
Frequently Asked Questions
Build AI Authority Beyond Backlinks
Magna helps businesses earn AI citations through entity authority and brand mention strategies. Schedule a free intro call.
Schedule Intro Call →