Anchor Text Strategy: The Complete Guide (Including the 4th Type Nobody Talks About)

The anchor text types everyone knows (and the one they don't)

Every SEO guide lists the same anchor text types. Here they are, plus the one nobody teaches:

Type	Definition	Example (targeting "best car subwoofer")
Exact Match	Anchor exactly matches the target keyword	"best car subwoofer"
Phrase Match	Target keyword appears within a longer phrase	"find the best car subwoofer for your budget"
Partial Match	Parts of the keyword present, but NOT in a sequence that forms the exact match	"subwoofer reviews for car audio"
Branded	Your brand name	"GrindstoneSEO"
Naked URL	The raw URL as the anchor	"grindstoneseo.com/subwoofers"
Generic	Non-descriptive action text	"click here", "read more", "this article"
Natural Anchor Text (NAT)	Topically relevant but does NOT match the keyword in any form — not exact, not partial, not phrase	"upgrade your car audio system", "deep bass on a budget", "what to look for in car speakers"

The first six types are in every guide on the internet. Moz, Yoast, Semrush, Neil Patel — they all cover these. NAT is the one they don't.

The difference between NAT and the others isn't just semantic. It's strategic. Every other anchor type is defined by its relationship to your target keyword: it either matches it (exact), contains it (phrase), partially overlaps with it (partial), or avoids it entirely in a non-descriptive way (generic). NAT does something different: it's relevant to the TOPIC without touching the KEYWORD. Google understands the topical connection. You get the relevance signal without the penalty risk.

Why Natural Anchor Text works

Here's the core insight that took me years of testing to trust:

Google doesn't need you to tell it that the connection between the best car subwoofer review page on another site and your best car subwoofer review site needs to be conjoined by the words "best car subwoofer review." It just doesn't. It can make that association by itself just fine.

Google has been building entity association models since the Knowledge Graph launched in 2012. By 2026, it understands that a page about "upgrading your car audio" is topically related to a page about "best car subwoofers" without needing the anchor text to spell it out. The anchor provides contextual signal. It does NOT need to be a keyword match to provide that signal.

The Fred/Core sliding scale

When the Fred update rolled into Core, it established a sliding scale for how anchor text is scored based on the authority relationship between the linking site and the receiving site:

• Strong site → strong site: You can be more aggressive with keyword anchors. A DR 70 publication linking to a DR 50 site with exact match anchor text is normal editorial behavior. Low risk.
• Strong site → weak site: Moderate risk. A DR 60 site linking to a DR 15 site with exact match looks less natural. Use NAT or partial match.
• Weak site → weak site: High risk. Exact match anchors between two low-authority sites is the exact pattern Penguin was built to detect. NAT or branded anchors only.

This is why negative SEO works: exact/phrase/partial match anchors from low-quality sites to a site that isn't strong enough to absorb them creates what I call "toxic stew." The anchor text itself isn't toxic. The authority gap between the sites makes it toxic.

The future-proofing argument

Here's the practical reason NAT matters beyond theory: link authority decays. A site that has DR 50 and 10,000 organic visits today might have DR 50 and zero organic visits in 6-8 months. If you built exact match anchors from that site, those anchors are now exact match from a dead site — and that's a penalty signal.

If you used NAT anchors instead? You don't care. The anchors were never tied to your target keyword. Even if every linking site in your profile drops to zero authority tomorrow, your anchor profile still looks natural because it was never optimized to begin with. That's future-proofing.

NAT in practice: real examples

Theory is nice. Let me show you what NAT looks like in a real campaign.

Target keyword: "jobs that pay good money"

What most SEOs would build:

• "jobs that pay good money" (exact match)
• "high paying jobs" (partial match)
• "find jobs that pay good money near me" (phrase match)
• "click here" (generic)

What I actually built (NAT anchors):

• "for good money jobs"
• "start working for good money"
• "earn some real cash"
• "find the easiest way to earn money"
• "ready to work for real money"
• "a chance for employees to earn real money"
• "if you want to earn more money"
• "competing for higher wages"
• "earn real money"
• "work for larger sums of money"

Result: Zero exact match. Zero phrase match. The closest any anchor got to partial match was about 20%, once. The page ranked. Google made the topical association on its own. And when core updates hit? Nothing moved. Because there was nothing to penalize.

The hidden trap: every anchor is exact match to SOMETHING

One insight that most SEOs miss: every anchor text string is an exact match for some search query, somewhere in Google's index. "Earn some real cash" might be a NAT anchor for "jobs that pay good money," but it's also an exact match for "earn some real cash" as a search term.

This matters when your NAT anchors inadvertently become high-percentage matches for long-tail keywords. If you build 10 links with anchors containing "earn money" and Google sees those anchors as over-optimized for the query "earn money," you can get those long-tail rankings knocked. The mitigation: vary your NAT anchors as much as you vary everything else. Don't use the same NAT phrase twice. Every link, different anchor — even within NAT.

The anchor text ratio question

This is the PAA question everyone searches for: "What's the ideal anchor text ratio?" And the honest answer is: it depends. But here's a baseline that works for most campaigns:

Anchor Type	Baseline %	Notes
Branded	20-30%	Your brand name, brand + keyword variations
Naked URLs	20-30%	Full URLs, homepage URL, deep page URLs
Natural Anchor Text (NAT)	15-25%	Topically relevant, keyword-absent. The future-proofing layer.
Partial Match	10-20%	Parts of target keyword in different arrangements
Generic	5-15%	"Click here", "read more", "this resource"
Exact Match	5-10%	Use sparingly. Only from strong sites. Never from weak sites.

Three caveats that make generic ratios useless:

1. The existing profile matters more than the ideal. If your page already has 50 backlinks and 40% are exact match, adding another exact match is pouring gasoline. Pull the ratio back with branded, NAT, and generic anchors. Always analyze what's there before deciding what to add.
2. Competitive landscape sets the ceiling. Look at what the top 3 ranking pages' anchor profiles look like. If they're all at 5% exact match, going to 15% makes you the outlier. Match the competitive norm, then differentiate on NAT and content quality.
3. The Fred/Core sliding scale applies. Exact match from a DR 70 site is not the same risk as exact match from a DR 20 site. The "5-10% exact match" guideline assumes a mix of linking site authority. If all your links come from weak sites, even 5% exact match may be too much.

I wrote three deep-dives on this for my newsletter: Anchor Text Wanker Smecht covers the NAT philosophy, Internal Anchor Text: Don't Get Rekt covers the internal linking rules, and How To/How Not To With Live Examples covers advanced tactics with real campaign data.

Internal anchor text rules (post-2023)

This is the section most anchor text guides skip entirely, because most of them were written before Google changed the rules.

Until late 2022, internal exact-match anchor text was a loophole. You could interlink your own pages with keyword-rich anchors and get a ranking boost without the penalty risk that came with external exact match. SEOs exploited this aggressively. Then Google closed it down in late 2022/early 2023.

Today, internal body links follow the same penalty logic as external links. A page with 50 internal links all using exact-match anchor text for the same keyword looks just as manipulative as 50 external links doing the same thing. The difference: navigation and menu links are treated differently from body content links. Google recognizes DOM structure — a keyword in your nav menu isn't scored the same as a keyword-rich anchor in your blog content. If it were, every local business with "plumber in [city]" in their nav would self-penalize.

My internal linking rule: The more topically relevant the two pages are, the less I need the anchor to carry keyword signal. If I'm linking from a page about niche edits to a page about link building pricing, Google already understands the topical relationship. A branded or generic anchor works fine. The anchor doesn't need to do the heavy lifting when the content context already does it.

One tactic that still works: place at least one exact anchor back to the page you want to rank, but place it on the page itself — an internal self-reference. Old school SEO, and it still helps for incremental gains when you're competing on page 1 where every position matters.

What to do when your anchors are already over-optimized

If you've inherited a link profile where 40%+ of anchors are exact match, or you bought links from a provider who filled every order with the target keyword, here's the recovery path:

1. Try to get anchors changed first. Contact site owners or your link provider and request anchor text changes. This is the cleanest fix — no negative link loss. Some publishers will update anchors for free. Others won't respond. Start here anyway.
2. Disavow at the page level. If you can't get anchors changed on toxic links, disavow them — but do it at the page level, not the domain level. Use Link Research Tools (LRT) rather than Ahrefs for this analysis — Ahrefs captures roughly 28% of the links that LRT finds. You need the full picture to make disavow decisions.
3. Get pages recrawled after disavow. Submit the disavow file, wait 48 hours for Google to process it, then request indexing of the affected pages. The disavow only works once Google recrawls and re-evaluates.
4. Dilute with NAT and branded links. Build new links with NAT and branded anchors to shift the ratio. If your profile is 40% exact match with 100 links, you need ~50-80 new links with non-exact anchors to bring the ratio below 20%. That takes time and budget. Plan for it.

The mistakes that get sites penalized

• All exact match, all the time. This is the #1 anchor text mistake and it's been the #1 mistake since Penguin launched in 2012. If more than 10-15% of your anchors are exact match, you're in the danger zone. Above 20%, you're gambling.
• Ignoring the authority gap. A DR 70 site can link to you with more aggressive anchors than a DR 20 site. Building exact match links exclusively from low-authority sites is the pattern Penguin was designed to detect. Match your anchor aggressiveness to the authority of the linking site.
• Treating internal and external anchors as separate systems. Post-2023, they're scored by the same logic. If your internal links are 80% exact match and your external links are 10% exact match, you're still over-optimized — you just hid it in a different place.
• Using the same anchor twice. Even within NAT, repetition creates a pattern. Every link should have a unique anchor. If your provider is reusing the same 5 anchors across 20 placements, they're creating a footprint.
• Not analyzing before building. You can't set the right anchor ratio for new links without knowing what the existing ratio is. Pull your anchor profile in Ahrefs (or better, LRT) before placing a single new link. The data tells you what your profile needs, not a generic guide.
• Listening to advice from 2015. Anchor text strategy has fundamentally changed since Penguin went real-time in 2016, since Fred/Core in 2017, and since the internal loophole closed in 2023. If your strategy is based on a guide that recommends "30% exact match," update your information.

Frequently asked questions about anchor text strategy

What's the ideal anchor text ratio?

Baseline: 20-30% branded, 20-30% naked URLs, 15-25% NAT, 10-20% partial match, 5-15% generic, 5-10% exact match. But ratios depend on your existing profile, competitive landscape, and the authority of your linking sites. A page with 50 existing exact-match anchors needs branded and NAT dilution, not more keywords. Always analyze before building.

What is Natural Anchor Text (NAT)?

Anchors that are topically relevant to your target page but don't match your target keyword in any form. For "best car subwoofer," NAT would be: "upgrade your car audio system," "deep bass on a budget," "what to look for in car speakers." The anchors signal topic relevance without creating keyword match risk. I've been testing and refining NAT since 2019 — the Anchor Text Wanker Smecht newsletter covers the full philosophy.

Does anchor text still matter in 2026?

Yes — but the signal has shifted from keyword matching to topical context. Google no longer needs exact-match anchors to understand what your page is about. Anchor text contributes to the topical context around your page, not a keyword match instruction. NAT leverages this by providing topical signals without keyword risk.

What should you NOT do with anchor text?

Don't go all exact match. Don't ignore the authority gap between linking and receiving sites. Don't reuse the same anchor across multiple links. Don't assume internal links are exempt from penalty logic. And don't follow anchor text advice from guides written before 2023 — the rules have changed.

What is anchor text distribution?

The breakdown of anchor types across all links pointing to a page or domain. A natural distribution has a mix of branded, naked URL, NAT, partial, generic, and a small amount of exact match. An unnatural distribution — where one type (usually exact match) dominates — signals manipulation.

How do I fix an over-optimized anchor profile?

Four steps: try to get existing anchors changed, disavow toxic exact-match links at the page level (use LRT, not just Ahrefs), get pages recrawled, then dilute the profile by building new links with NAT and branded anchors. Recovery takes months. Prevention takes one conversation with your link provider about anchor text strategy.

---

The bottom line on anchor text strategy

The anchor text game changed when Google got smart enough to make entity associations without you spelling them out in the anchor. The old playbook — 30% exact match, vary your anchors, hope for the best — is the playbook that gets sites penalized in 2026.

The new playbook is NAT: build anchors that are topically relevant, keyword-absent, and future-proof. Let Google do the entity mapping it's already good at. Your job is to give it topical context, not keyword instructions. I've been testing this since 2019 through every core update, and the data doesn't lie. NAT profiles don't get hit. Over-optimized profiles do.

If you only take one thing from this guide: stop building anchors for the keywords you want to rank for. Build anchors for the topic. Google will figure out the rest.

Want anchor text handled right?

We build every link with anchor text strategy baked in. No bulk exact-match orders. No keyword stuffing. Every placement gets an anchor that fits the NAT framework — topically relevant, naturally written, and future-proofed. We push back when clients request aggressive anchors because protecting your site's long-term health is more important than filling an order. Start with a test order.

For the full picture on evaluating link building providers (including how they handle anchor text), read the outsource link building guide. For pricing, see the link building pricing guide. And for the three newsletter deep-dives that this article is built on — subscribe below.