Unlinked Brand Mentions Beat Backlinks in 2026. Here's Why My Link-Building Agency Pivoted.

TL;DR

• Ahrefs studied 75,000 brands. Branded web mentions correlated with AI Overview visibility at 0.664. Backlinks correlated at 0.218. Mentions were roughly three times stronger.
• The May 2024 Google Search API leak confirmed Google tracks brand entities, site authority, and click behaviour as first-class ranking inputs. Rand Fishkin's summary was blunt: Build a notable, popular, well-recognized brand in your space.
• I run a link-building agency. This post argues that for most mid-market brands in 2026, the next 10,000 dollars is better spent earning mentions than buying links.
• The two are not enemies. Mentions often turn into links. The trick is knowing which one to chase first.
• At the bottom of this post you will find the exact email template I use to convert unlinked mentions into do-follow links, plus the conversion rate I see.

Why I'm Writing This (And Why It's Awkward)

I sell link building. My clients pay me to get their name, with a hyperlink, onto other websites. It has been a good business.

But here is what has been bothering me for the last 12 months. When I audit a client's AI visibility, the brands showing up in ChatGPT, Perplexity, and Google AI Overviews are not always the ones with the strongest backlink profiles. Some of them have mediocre link graphs. What they have is mentions everywhere. Podcasts, Reddit threads, press roundups, expert quotes, YouTube transcripts.

I kept explaining this away. Correlation is not causation. Links still work. AI will catch up.

Then two things happened. Ahrefs published a study of 75,000 brands. Then someone leaked 2,500 pages of Google's internal Search API documentation. Both data sources pointed in the same direction. Mentions, not links, now drive the heaviest off-page signal weight in the systems that matter most.

So my agency pivoted. We still build links. But I now tell clients something I would not have said two years ago: if you are choosing between another guest post and getting quoted in a podcast, take the podcast. This post explains why, with the data, and then shows you exactly how to do it.

What Are Unlinked Brand Mentions, Really?

An unlinked brand mention is any public reference to your brand, product, or founder that does not include a clickable hyperlink.

Examples I see every week:

• A journalist writes "According to SEO Engico, most agencies undercharge for audits" and never links.
• Someone on Reddit types "I used Be Cool Refrigeration and they were fast".
• A podcast host says your name in an episode, which ends up in a YouTube transcript, a Spotify description, and a dozen clipping sites.
• A directory lists you in a top-10 roundup but forgets the link.

For 15 years most SEOs treated these as almost links. Nice to have. Not the main event.

That framing is now outdated. Google's systems and every major LLM ingest the open web constantly, and they score trust based on how often your name co-occurs with certain topics, not just whether there is an <a href> pointing at you.

The Data That Broke My Link-Building Brain

Let's look at what actually changed. I am going to stick to primary sources only.

The Ahrefs 75,000-Brand Study

In May 2025, Ahrefs published an analysis of 75,000 brands and measured which factors correlated with appearing in Google AI Overviews. The top correlations were not what most SEOs expected. The full study is on the Ahrefs blog.

Here is what they found:

• Branded web mentions: 0.664 correlation with AI Overview visibility
• Number of backlinks: 0.218 correlation
• URL Rating (Ahrefs' page authority score): also weak, below backlinks

In plain English, brand mentions across the web correlated with AI Overview appearance roughly three times more strongly than raw backlinks did. Ahrefs then extended the study in December 2025 and looked at ChatGPT and Google AI Mode as well. That follow-up found that YouTube mentions had the single highest correlation (around 0.737) with AI visibility across platforms.

Ahrefs are careful to say correlation is not causation. I agree. But when three separate studies from the same dataset keep putting mentions at the top and links further down, you stop arguing with the signal.

The May 2024 Google Search API Leak

On 27 May 2024, Rand Fishkin published the first major analysis of the leaked Google Search API documents. Mike King followed the same day with a much deeper engineering breakdown at iPullRank. Search Engine Land's Danny Goodwin summarised it for the broader industry.

The leak exposed 2,596 modules and 14,014 attributes from Google's Content API Warehouse. A few things jumped out:

• Google maintains a siteAuthority feature used in its Q* ranking system. For years Google said no such thing existed.
• Google tracks homepage trust levels, classifying domains as NOT_TRUSTED, PARTIALLY_TRUSTED, or FULLY_TRUSTED.
• A system called NavBoost uses click-through data from Chrome to distinguish good clicks from bad clicks, and this directly feeds ranking.
• Author and entity recognition features track whether a named entity is the author of a document, particularly in news and science.

Fishkin's one-line takeaway was this: Build a notable, popular, well-recognized brand in your space, outside of Google search. You cannot gaming your way into the NavBoost tiers or the siteAuthority score. You build the underlying brand, and the signals follow.

That sentence is the entire thesis of this post. If you want to go deeper on how Google is now weighing brand narrative inside AI answers, I wrote a full piece on defensive SEO and how to shape your AI brand narrative.

Search Engine Journal Corroboration

Search Engine Journal's November 2025 coverage of the Ahrefs dataset quoted Tim Soulo, Ahrefs' CMO, calling citations "the new backlinks." When the CMO of a company that built its fortune on a backlink index starts using that language, something has shifted.

Linked Backlinks vs Unlinked Brand Mentions: Head to Head

Here is a blunt comparison based on what I see across client accounts in 2026. Numbers are my agency's averages, not lab data.

The one column where links still win outright is human click-through. A mention with no link does not send you referral traffic directly. A link does.

But in 2026 that matters less than it used to, because so much of the journey starts in AI interfaces where the user never visits the source page anyway. If your brand is mentioned in 40 places ChatGPT trusts, you get cited in ChatGPT. That is traffic of a different shape.

I still think white hat link building is worth doing. I just think the ratio of spend has to change. More on that in a moment.

Introducing Mention Velocity (The Metric I Track Now)

This is my framework, developed for agency clients over the last 18 months. I call it Mention Velocity. It measures how fast your brand is picking up new unlinked mentions per month, relative to your linked backlinks.

The formula is simple:

Mention Velocity = (New unlinked mentions in month) / (New linked backlinks in month)

If your Mention Velocity is below 1, you are earning more links than mentions. That used to be a healthy signal. In 2026 it often means you are over-invested in outreach and under-invested in brand.

If your Mention Velocity is between 1 and 3, you are roughly balanced. Good.

If your Mention Velocity is above 3, you have serious brand gravity. You are being talked about more than you are being linked to. This is where the biggest AI visibility wins live.

I benchmark clients every quarter now. I also cross-reference Mention Velocity against the approach I describe in my guide to citation strategy across ChatGPT, Perplexity, and Google AI Mode, because different platforms weight different mention sources.

8 Tactics That Earn Unlinked Mentions (In Order of ROI)

Here are the eight tactics my team actually runs, ranked by the ROI I see across about 40 active client accounts. I am not including anything I have not personally executed.

1. Podcast Guesting (Highest ROI)

One 45-minute podcast episode produces: a show notes page that sometimes links, a Spotify description, a YouTube auto-transcript, dozens of clipping sites, and search indexation of your name alongside the host's topic. Even without a link, you get mentioned in searchable text across five or six properties.

I use Podchaser and Listen Notes to find shows. I pitch the host, not a producer. My close rate is around 14 percent on cold outreach.

2. HARO and Modern Equivalents

HARO still works. Three emails a day. You reply to relevant queries with a real, useful answer. Journalists quote you in their articles.

Some of those quotes come with links. Many do not. In 2022 I would have bounced when no link was offered. Now I take the mention happily, and I use the follow-up template at the bottom of this post to convert maybe one in four into links later.

If you want a deeper breakdown of how we run this at scale, read my post on HARO link building as a service.

3. Original Data Studies

One original research piece generates more mentions than 50 guest posts. Journalists need data to write stories. If you publish defensible numbers, they cite you.

My rule: one 500-response survey per client per year, minimum. The mentions it produces keep compounding for 18 months.

4. Reddit and Forum Participation

Reddit threads now rank in Google AI Overviews constantly. If your brand gets mentioned in relevant subreddits by real users, that single mention can end up quoted inside AI answers for weeks.

I do not recommend manipulating Reddit. The community smells it from a mile away. I recommend answering questions under your real account, and making your product good enough that customers recommend you organically.

5. Industry Report Inclusion

Most industries have annual reports from analyst firms, trade bodies, or media brands. Getting listed as a vendor, case study, or expert source inside a PDF produces long-tail mentions every time someone quotes that report.

This is a slow play. Budget six months.

6. Expert Quotes in Roundup Articles

The entire "x experts weigh in on y" format survives because editors love it. Pitch three unique quotes per month to publications you would actually want to be in. My digital PR strategy post explains the pitch format I use.

7. YouTube Mentions (Via Partners and Creators)

The Ahrefs December 2025 follow-up found YouTube mentions had the single highest correlation with AI visibility, at around 0.737. Getting mentioned in a YouTube video, even once, gets your name indexed in the transcript and picked up by both Google and LLM training crawlers.

Strategy: identify five mid-size creators in your vertical. Send them a product, ask for honest feedback. One in four will mention you on camera.

8. Wikipedia and Wikidata Entity Pages

This is the hardest tactic and the one I recommend last. If your brand meets Wikipedia's notability threshold, an entity page there is gold. But the threshold is high, and paid Wikipedia editing violates their terms. Earn it the slow way with genuine press coverage, or skip it.

How to Monitor Mentions (Tools I Actually Use)

You cannot grow what you do not measure.

Here is the stack I run for every client. I am not affiliated with any of these.

1. Google Alerts. Free. I set up alerts for the brand name, founder name, and two or three product names. It misses things, but it is the cheapest early-warning system.
2. Ahrefs Brand Mentions. Paid. Covers the web index well and integrates with linked backlink tracking so you can see which mentions already link and which do not. My default for client reporting.
3. Mention.com or Brand24. Paid. Better for social and forum coverage than Ahrefs. I use one of these in parallel.
4. Manual Reddit and X searches. Weekly. Still catches threads the automated tools miss.
5. Google Search Console. Since Google rolled out the branded vs non-branded queries filter, I track branded query volume as a leading indicator of mention-driven brand awareness.

Once you have a mention inventory, you sort it into three buckets. Linked (great, done), unlinked but high-authority (go after the link), unlinked but low-authority (monitor, do nothing).

The Email Template I Use to Convert Unlinked Mentions Into Links

This is the part most articles skip. They tell you mentions matter, then leave you there. Here is exactly what I send.

Subject line: Quick tweak to your [topic] piece

Hi [Name],

I saw your piece on [specific topic, not generic] and loved the angle about [a specific detail you actually read]. Thanks for mentioning SEO Engico in paragraph six.

Would you mind linking the mention to [URL]? It would help readers who want to see the original data we referenced, and it takes about 10 seconds to add.

No pressure either way. Either way, nice piece.

Jhonty

That is the whole email. Three paragraphs. No pitch, no follow-up sequence, no emojis.

My current conversion rate on this template, tracked across 312 sends in the last 18 months, is 27 percent. Roughly one in four unlinked mentions converts to a do-follow link within 14 days.

The mistakes that tank the conversion rate:

• Asking for a specific anchor text. Never do that. Let them choose.
• Pitching a second piece of content in the same email. Kills it.
• Using a no-reply agency email instead of a real human one.
• Waiting more than four weeks from the original publication date. After that, editors have moved on.

How AI Search Platforms Weight Mentions Differently

Not every mention matters equally. Where the mention lives changes its value depending on which AI system you want to show up in.

Based on the Ahrefs branded web mentions study:

• Google AI Overviews: Strong correlation with mentions (0.65). Google is, in their words, "more biased toward big brands." Build breadth.
• ChatGPT: Much weaker correlation (0.15). ChatGPT draws heavily from its training data, which skews toward Reddit, Wikipedia, YouTube, Stack Overflow, and news sites. Target those.
• Perplexity: Middle ground at 0.30. It live-fetches, so fresh mentions matter more.

Practical translation: if your target is Google AI Overviews, run mentions broadly. If your target is ChatGPT, concentrate on a small number of specific property types. I go deeper on this in my post about LLM optimization and getting your brand into AI answers.

Where Links Still Win (The Honest Part)

I am not arguing you should stop building links. I would be out of business in a month.

Links still do three things mentions cannot:

1. They send direct referral traffic. A good PR mention on a popular site with a link can drive hundreds of visits. An unlinked mention sends zero.
2. They transfer PageRank. Despite everything the leak confirmed about click signals, PageRank-style link weight is still in the ranking system. Mike King's iPullRank analysis makes that explicit.
3. They are faster to verify. A link either exists or it doesn't. A mention's impact is statistical and delayed.

The budget question I now ask every new client is this: if we have 10,000 GBP per month for off-page, how do we split it? My default answer in 2026 is 40 percent on links, 40 percent on mention-earning activity (PR, HARO, podcasts, data studies), and 20 percent on mention conversion (the email template above, plus link reclamation tooling). Two years ago it was 80/20 in favour of links. The shift is real.

If you want to see how we structure ongoing campaigns around this split, read my link building agency framework post, or jump straight to the link building page for how my team runs it.

A Short Case Study From My Own Agency

One of my clients, a UK SaaS tool in the legal tech space, spent 18 months on a traditional link-building programme from late 2023 to mid 2025. They built 94 do-follow links. Their Google rankings moved, but their AI Overview appearances were stuck in the single digits per week.

In June 2025 we switched approach. We kept building links at half the previous pace, and poured the other half of the budget into:

• Eight podcast appearances for the founder
• One original data study of 400 UK lawyers, pitched to trade press
• Daily HARO and PR platform pitching
• A dedicated Reddit presence in two subreddits

By February 2026, their AI Overview appearances had grown from 4 per week to 61 per week. Their branded search volume in Google Search Console's new branded filter tripled. Total unlinked mentions grew from 47 per month to 290 per month. Mention Velocity went from 0.5 to 3.2.

Organic traffic up 38 percent. Inbound leads up 71 percent. Link-building spend actually dropped.

I wish I could pretend this is a complicated story. It is not. We moved money to the signal that was doing the heavier lifting. That is the whole insight.

The 30-Day Mention Acceleration Plan

If you want to run this yourself, here is the plan I give every client.

1. Day 1: Audit current mentions using Ahrefs Brand Mentions or Mention.com. Sort into linked, unlinked high-authority, unlinked low-authority.
2. Days 2 to 3: Set up Google Alerts for brand, founder, product names. Add tracking to your monthly SEO report.
3. Days 4 to 7: Email every unlinked high-authority mention using the template above.
4. Day 8 to 14: Set up HARO and pitch at least 10 queries. Identify five podcasts in your vertical.
5. Day 15 to 21: Record two podcasts. Launch a small data study (even 200 responses works).
6. Day 22 to 30: Publish the data study with a press-ready summary. Pitch to trade press. Track every mention.

By day 30 you should see 20 to 80 new unlinked mentions depending on industry competitiveness, and roughly one in four of the eligible ones converting to links via the reclamation template.

This plan layers on top of standard technical work. It does not replace the fundamentals I cover in my E-E-A-T implementation guide, or broader link building strategies. Do both.

Frequently Asked Questions

Are unlinked brand mentions a confirmed Google ranking factor in 2026?

Not confirmed by Google as a named ranking factor. However, Ahrefs' 75,000-brand study showed a 0.664 correlation between branded web mentions and Google AI Overview visibility, compared with 0.218 for backlinks. The May 2024 Google Search API leak also revealed brand-level signals like siteAuthority and entity recognition that Google previously denied existed. The honest answer is that mentions behave like a ranking factor even if Google does not label them as one.

Do unlinked mentions help with Google search rankings, or only with AI?

Both, indirectly. The clearest correlation is with AI visibility. But mentions also drive branded search volume, which feeds Google's NavBoost click system, and they build the entity-level signals that feed siteAuthority. If your brand is being discussed, Google's systems pick up the pattern.

How is a brand mention different from a citation?

The terms overlap. In local SEO, a citation usually means a business listing on a directory with NAP data (name, address, phone). A brand mention is broader and includes any reference to your brand in editorial content, podcasts, Reddit, YouTube, or press coverage. For AI search, mentions in the second sense are what matter most.

How long does it take for a new mention to affect rankings?

For Google's core ranking systems, expect indirect effects over 8 to 16 weeks as branded search volume and entity signals build. For AI Overviews and ChatGPT citations, the lag is usually shorter, around 4 to 8 weeks, because these systems re-index or retrain more aggressively.

Should I stop doing link building in 2026?

No. Links still drive referral traffic, PageRank, and rapid signal transfer. What should change is the ratio. My default is now 40 percent links, 40 percent mention-earning, 20 percent mention-to-link conversion. Two years ago it was closer to 80/20 in favour of links.

Can I pay for brand mentions like I pay for backlinks?

You can, and you should not. Paid undisclosed mentions violate FTC guidance in the US and equivalent consumer-protection rules elsewhere. More practically, AI systems seem to downweight the patterns that paid placements create. Earn the mentions.

What is the difference between Digital PR and mention-focused SEO?

Digital PR is a tactic. Mention-focused SEO is the strategy. Digital PR campaigns are one of the best ways to generate mentions, which is why I wrote a full digital PR and link building strategy post. But mentions can also come from podcasts, forums, research, and product excellence. PR is a subset of mention-building.

How do I track unlinked mentions if I do not have a paid tool?

Use Google Alerts plus a manual monthly Reddit, X, and YouTube search for your brand name. Create a simple spreadsheet with columns for source, date, URL, and whether it links. It is tedious but free, and it works up to maybe 50 mentions per month. Beyond that volume, invest in Ahrefs, Mention.com, or Brand24.

Closing Thought From Someone Who Sells Links for a Living

If you read one line from this post, read this. Google and the major LLMs have spent the last five years telling us in increasingly specific language that brand is the moat. The Ahrefs data is simply the quantitative version of that story. The Google leak was the engineering version.

I sell link building, and I still believe in it. But pretending links are the centre of the universe in 2026 is dishonest, and I would rather lose a bit of my own margin than lie to the people paying me to make them visible.

If you want help auditing your current mention footprint and building a 90-day plan that mixes links, mentions, and AI visibility work, that is what my team does every day. Have a look at our approach to brand mentions as a service, or reach out through the digital PR page if you want a more campaign-led conversation. And if you want to see how this thinking fits into broader AI visibility work, our LLM visibility page is the best place to start.

The honest version of SEO in 2026 is this. Build a brand people say out loud. The algorithms are listening.