Backlink types

25 Types of Backlinks: Value, Risk, and How Each One Dies

A backlink is a link from one site to another, and the type describes how it came to exist: somebody cited you, you pitched an article, you paid, you traded. That is a useful way to organize a strategy and a poor way to judge a single link, because every category below spans the full quality range. The best link you get this year and the worst one might both be guest posts.

So this page does two things. It groups all 25 types by how they are acquired, with what each is worth and what it risks. And it adds a column nobody else publishes: how each type eventually disappears, because a link that dies in eight months and one that lasts a decade look identical in an export until something rechecks them.

18 min read
Published August 18, 2026
Updated August 18, 2026
The LinkSwapy team

The short answer

Type describes acquisition, not quality. Every category here spans the full range.
Four groups are reliably worth it, three depend entirely on execution, two are not worth your time.
Guest posts and niche edits generate the most argument because execution decides them.
Durability varies more between types than value does, and almost nobody tracks it.
Collecting one of every type is not link diversity. It is a checklist.

Google's link spam policy

Backlink acquisition categories branching from a webpage while emphasizing that quality depends on the specific page — Backlink type describes how a link was acquired. The page, editorial decision, placement, and audience determine its value.

Contents

What this covers

The catalog is the middle section and it is long. If you only want the judgment, the tiers and the verdict are near the bottom.

Start here

Why the label predicts less than you think

Nearly every article on this subject sorts backlink types into good and bad, and the sorting falls apart the moment you apply it to a real opportunity. Short definitions for every term used below are in the link building glossary.

The best link you get this year and the worst one might both be guest posts.

Take guest posts. A contributed piece in a publication your customers actually read, written because you had something to say, is among the better links you will earn all year. A structurally identical link from a site publishing forty posts a week from anyone who pays is worth nothing and carries policy risk. Same type. Opposite outcomes. The category told you nothing.

That holds for most of the list. Directory links span industry bodies people genuinely use and submission farms that have been worthless since roughly 2012. Partner links span a real integration and a wall of traded placements. Even editorial links, the type everyone agrees is the good one, include citations on pages nobody has visited since publication.

Which is why we would not build a strategy around collecting one of each. The type is a prior, useful for deciding where to spend a week of effort. The specific page is the decision. Use the catalog below to set expectations, then judge the actual URL in front of you.

Two guest article backlinks with the same acquisition type but different outcomes based on relevance and editorial control — The same backlink type can produce a useful editorial citation or an unrelated placement built around the link.

The catalog

25 types of backlinks, grouped by how you get them

Acquisition is the split that matters operationally, because it maps onto what a team actually does on a Tuesday. Value and risk are our editorial readings, not scores anyone publishes. The durability note is the mechanism, not a statistic.

Earned without asking

Somebody decided to reference you. These are the links you cannot schedule, which is exactly why they carry the most weight and why every other group on this list exists.

01Editorial backlink
High valueLow risk
A writer references your page inside their content because it supported the point they were making. No arrangement, no ask.
How it dies: Slowly. Usually a site migration that drops old URLs, or a content prune years later. The most durable type on this list.
02Digital PR backlink
High valueLow risk
Coverage of a story, dataset or survey you produced, in a publication with an actual newsroom.
How it dies: The original article is stable. Syndicated copies vanish without warning, and paywall migrations can take the link out of reach.
03Expert quote backlink
High valueLow risk
A journalist or blogger quotes you and credits the source, often through a query service or a direct relationship.
How it dies: Roundups get refreshed annually and old contributors get cut. News pieces stay put.
04Image or infographic backlink
Medium valueLow risk
Someone uses your chart, diagram or original visual and credits it with a link.
How it dies: When the host post gets updated and the image is swapped for a newer one. The credit line goes with it, silently.

Built through outreach

You found the opportunity and asked. Legitimate work, and the group where execution quality varies most, because the same email can produce an excellent placement or a worthless one.

05Guest post backlink
VariesMedium risk
You write an article for another publication and it links back to you, usually in the body or the author bio.
How it dies: New editor, contributor purge. Guest content is the first thing pruned when a publication tightens up, because nobody on staff is attached to it.
06Niche edit backlink
VariesHigh risk
A link inserted into an article that was already published, usually arranged with the site owner.
How it dies: Fast, relatively speaking. It was added to somebody else's finished piece, so the next content refresh removes it and nobody notices.
07Resource page backlink
Medium valueLow risk
Your page is added to a curated list of tools or references on a topic.
How it dies: On the periodic prune. Resource pages rot faster than any other page type, so their owners audit them more often.
08Broken link backlink
Medium valueLow risk
You find a dead outbound link on a relevant page and offer your resource as the replacement.
How it dies: Rarely. You solved a problem the page owner already had, which is the most durable reason a link exists.
09Unlinked mention backlink
High valueLow risk
A site already names your company, product or research. You ask them to make it a link.
How it dies: Like any editorial link, because that is what it became. The mention was already theirs.

Relationship backlinks

Links that exist because of something you did together. These behave differently from every other group, because they live and die with the relationship rather than with the content.

10Interview backlink
Medium valueLow risk
A publication or blog interviews you and links to your site in the intro or the bio.
How it dies: Rarely. Interviews are dated content, so nobody feels the need to update them.
11Podcast backlink
Medium valueLow risk
A link from an episode page, show notes or a guest bio.
How it dies: Wholesale, when the show migrates hosting platforms and the entire back catalog changes URLs.
12Case study backlink
Medium valueLow risk
A vendor, client or partner writes up work involving you and links back.
How it dies: When the relationship ends, or when the client churns and the page comes down with them.
13Partner backlink
Medium valueMedium risk
Integration partners, suppliers, associations and technology partners referencing each other.
How it dies: When the partnership lapses, and nobody sends a note about it. This is the most commonly missed removal we see people describe.

Listings and citations

Structured entries rather than editorial writing. Modest ceilings, and the group where a small number of good ones beats a long list every time.

14Directory backlink
Low valueMedium risk
A listing in a business, industry or professional directory.
How it dies: When the annual listing lapses, or when the directory itself is abandoned, which happens more often than people expect.
15Local citation backlink
Medium valueLow risk
A listing on a local platform, chamber of commerce or regional business index, usually with name, address and phone.
How it dies: Slowly if the listing is claimed. Unclaimed entries get overwritten on the next data aggregator refresh.
16Profile backlink
Low valueLow risk
A link in a business, author or community profile on a platform you signed up to.
How it dies: Sitewide, when the platform changes its outbound link policy and nofollows every profile at once.

Community backlinks

Links from places people talk. Worth having as a byproduct of genuine participation, and not worth building as an activity.

17Forum backlink
Low valueMedium risk
A link inside a forum answer or signature, where the discussion is genuinely relevant.
How it dies: When threads are archived, or the whole forum moves platform and the URL structure changes.
18Blog comment backlink
Low valueMedium risk
A link left in a comment on somebody else's post.
How it dies: Easily. Comment purges, plugin migrations and spam sweeps take them out in batches. Almost always nofollow anyway.

Commercial arrangements

Money or compensation changed hands. Legitimate when qualified with the right attribute, and a policy violation when not, which is the distinction this group turns on.

19Affiliate backlink
VariesLow risk
A link from a partner who earns commission on the traffic or sales it sends.
How it dies: When the program ends or the affiliate stops promoting. Should carry a sponsored or nofollow attribute throughout.
20Sponsored backlink
VariesLow risk
A placement you paid for and disclosed, marked with rel=sponsored.
How it dies: At the end of the term, by contract. That predictability is the honest version of paying for placement.
21Paid backlink for ranking purposes
VariesVery high risk
A link bought specifically to pass ranking signals, without a qualifying attribute.
How it dies: The day you stop paying, which is the entire business model. Named directly in Google's spam policies.

Traded backlinks

You gave a link to get a link. Ordinary between real partners, and a recognizable pattern once it becomes a program. This cluster has three articles on it, so the entries here stay short.

22Reciprocal backlink
VariesMedium risk
Two sites linking to each other, whether by arrangement or coincidence.
How it dies: When your partner cleans house, or when they notice your side came down first.
23ABC or three-way backlink
VariesHigh risk
You link to one site and a third site links back, so no two participants point at each other.
How it dies: Quietly, because nobody clearly owns the obligation once the arrangement involves three parties.

Manufactured networks

Sites that exist to link. Both entries are here for completeness rather than consideration, and we would not place a client on either.

24PBN backlink
VariesVery high risk
A link from a private blog network: sites bought or built specifically to link to properties their owner controls.
How it dies: All at once, when the network is deindexed or sold on. The correlated failure is the point people miss.
25Link farm backlink
Low valueVery high risk
A page or network whose purpose is holding large numbers of outbound links to unrelated sites.
How it dies: Whenever. It was passing nothing while it was alive, so the removal costs you little beyond what you paid.

The value and risk readings are our editorial judgment against Google's published policies, not scores anyone publishes, and reasonable practitioners disagree on several rows. What is not judgment is the durability column: those are mechanisms, and you can verify each one by watching a link of that type for a year.

The missing dimension

Nobody publishes how each type dies

Every guide to backlink types rates them on value and most rate them on risk. Almost none mention how long they last, which is odd, because a link that disappears in eight months and one that lasts a decade are not the same asset even if they scored identically on the day you got them.

A link that dies in eight months and one that lasts a decade look identical in an export.

The mechanisms are predictable once you look for them. Links inside content that was written for its own sake survive, because nobody has a reason to revisit them and the page has a reason to stay. Links inserted into content that already existed are fragile, because the next refresh sweeps them out and the person doing the refresh has no attachment to them. Links tied to a commercial or partnership arrangement end when the arrangement ends. Links on curated lists die on the prune, and curated lists get pruned because they rot.

That reframes a few of the standard rankings. Broken link building looks unglamorous and produces some of the most durable links you can get, because you solved a problem the page owner already had. Niche edits look efficient and are the most fragile of the built types for exactly the opposite reason.

We are not going to give you survival percentages by type, because we cannot source them and the ones circulating in this industry have no dataset behind them. What we will say is that the mechanism is checkable: pick ten links of one type, recheck them in twelve months, and you will have a number for your own niche that beats any published benchmark.

Four backlink loss mechanisms: site migration, content refresh, a partnership ending, and an outbound link policy change — Backlinks disappear through predictable structural changes, and each acquisition type has a different failure pattern.

Where the tool fits

The durability argument only pays off if something is checking. Rechecking placements on a schedule is what turns a list of links you once earned into a list of links you still have, and the gap between those two is usually larger than teams expect. See how backlink monitoring works.

What actually predicts value

Three questions that beat the category

If you collapse the 25 types down to what genuinely separates a good link from a bad one, you get three questions. A fourth, where on the page it sits, is independent of all of them. They apply to every row in the catalog and they take about a minute each.

Is the page relevant, at page level rather than site level. Would this link exist if backlinks counted for nothing. And does anybody actually read the page it sits on. That is the whole test, and it correctly sorts a guest post from a guest post, which no category system can do.

The third question is the one people skip, because domain metrics feel like a substitute for it and they are not. A page can sit on a domain with strong third party scores and have had eleven visitors since publication. The link is a link, but it is doing almost nothing, and no amount of category theory will tell you that. Opening the page will.

Relevance, judged on the specific page rather than the domain
Whether the link would exist with no search engines involved
Whether the page has a real audience, not just a domain rating

Three questions for evaluating a backlink: reader usefulness, independent editorial choice, and whether the page has a real audience — These three questions separate strong and weak links inside the same category more reliably than the type label.

The tiers

Reliable, conditional, and not worth it

Same 25 types, sorted by how much the outcome depends on execution rather than on the category, which is where in-house teams usually need the argument settled. The middle row is where nearly every argument about link building actually happens.

Tier	Types	Why they sit here
Reliably worth it	Editorial, digital PR, expert quotes, unlinked mentions, broken link, interviews, local citations	Somebody chose to reference you, or fixed a problem they had. Hard to get, hard to get wrong
Depends on execution	Guest posts, niche edits, resource pages, partner links, case studies, podcasts, directories, affiliate, sponsored, reciprocal	The type spans the full range. The specific site and page decide, not the label
Rarely worth the effort	Profile links, forum links, blog comments, image credits at volume	Usually nofollow, usually unread, and recognizable as an activity when built at scale
We would not place these	PBN, link farm, three-way networks, paid links passing ranking signals	Named in Google's spam policies, and they fail together rather than individually

Notice how large the second row is. Ten of the 25 sit there, which is the strongest argument against treating this as a taxonomy of good and bad links rather than a taxonomy of how links are acquired.

Matching effort to outcome

Which types suit which goal

Types are a poor quality signal and a decent planning signal, which is the one job they do well, and the one a monthly report has to represent. If you know what you are trying to move, the list of sensible activities gets short fast.

What you are trying to do	Where to spend	What to skip
Build authority in a competitive niche	Digital PR, original research, editorial citations	Directories, profiles, volume anything
Get referral traffic that converts	Partner links, integration docs, podcasts, case studies	Anything on a page with no audience
Improve local visibility	Local citations, chambers, regional trade bodies	General purpose submission directories
Earn links to a data asset	Digital PR, expert quotes, image and chart credits	Guest posts about the data
Promote a SaaS product	Integration partners, comparison pages, technology directories	Bulk guest posting
Build links to a long guide	Resource pages, broken link building, unlinked mentions	Niche edits into unrelated articles
Raise brand visibility	Interviews, expert commentary, podcasts, PR	Anything nobody reads

Every row's second column is the same idea stated differently: skip the activity whose output does not match the outcome you named. Most wasted link building budget is a mismatch rather than a bad link.

Two things people conflate

Type is not the same as attribute

The type says how a link came to exist. The rel attribute says how the HTML marks it, and the two are independent axes that get mashed together constantly.

A guest post link can be a plain follow link or carry rel=sponsored. A partner link can be either. Google asks for sponsored or nofollow wherever payment or compensation is involved, and for ugc on user generated content like comments and forum posts, which is why the community group in the catalog is mostly nofollow by default rather than by anyone's decision.

The practical consequence is that you cannot infer one from the other. A follow attribute does not make a paid link compliant, and a nofollow does not make a good editorial link worthless. Both facts about a link matter, and you have to check each separately, which is why dofollow, nofollow, sponsored and ugc get their own page rather than a paragraph here.

And the third axis

Type is not the same as placement

Placement is where on the page the link sits: inside a paragraph, in an author bio, in a sidebar, in a footer, on every page of the site at once.

Acquisition, placement, attribute. Type is the one people ask about and the least informative of the three.

This is the axis that quietly does the most work, and it cuts across every type in the catalog. The same guest post can put your link in the third paragraph of the article or in a bio block under it, and those are not equivalent. A partner link in a sentence explaining the integration behaves differently from the same partner link in a footer rendered across four thousand pages.

So when you are judging an opportunity, you are actually judging three things at once: how the link was acquired, where it will sit, and what attribute it will carry. The type is the one people ask about and the least informative of the three.

Backlink type, HTML attribute, and page placement shown as three separate evaluation axes — Type explains acquisition, attribute qualifies the relationship, and placement identifies where the link appears.

One process, any type

How to evaluate a backlink of any type

Seven checks that work on every row in the catalog, and the free backlink status checker covers the mechanical half. The order matters: the first two disqualify more opportunities than the other five combined, and both take under a minute.

1
Open the page the link will sit on
Not the homepage, not the domain metrics. The page. Everything that determines what a link is worth belongs to that URL, and thirty seconds of looking at it beats any amount of reasoning about what category the link falls into.
2
Check relevance at page level, not site level
A marketing blog is not automatically relevant to a link building tool. Ask whether somebody reading that specific page would want what you are offering. If you have to construct the argument, the answer is no.
3
Work out whether anyone reads it
Does the page rank for anything. Does the site have a history. Do the comments, shares or update dates suggest a person maintains it. A page with no audience passes very little regardless of what the domain metrics say about it.
4
Ask whether the link would exist without SEO
The single most useful question in link building, applied to one placement. If the honest answer is that the link only makes sense because backlinks count, you already know what type it really is, whatever the pitch called it.
5
Read the anchor request
Nobody dictating anchor text is doing editorial work. A supplied exact match commercial phrase tells you more about the arrangement than the site's domain rating does, and it is the fastest disqualifier on this list.
6
Check what else that page links out to
Count the outbound links and look at where they go. A page with forty unrelated outbound links is a different object from a page with three, and the category label will not tell you which one you are looking at.
7
Ask how this link is likely to die
Then decide whether you would still do the work knowing that. An insertion into somebody else's old article and a citation inside a piece written this week can look identical in an export and have completely different life expectancies.

The verdict

So which types of backlinks actually matter?

The ones where somebody chose to reference you, on a page people read, for a reason that would survive the disappearance of search engines. None of them is replaceable by internal linking. That describes about seven of the 25, and every one of them is harder to get than the eighteen that remain. The taxonomy is worth knowing so you can plan, and it is a poor substitute for opening the page and looking.

Editorial citations and press coverageLow
Outreach links on pages with real audiencesLow
Guest posts and niche edits, execution dependentMedium
Traded links once they become a programHigh
PBNs, link farms, unqualified paid linksVery high

Do not collect one of every type. Collect links from pages that would have linked to you anyway.

Questions

Questions people actually ask about this

Still unsure about something? Talk to our team and we will give you a straight answer.

How many types of backlinks are there?

There is no official number, and anyone presenting one as settled has picked a taxonomy and rounded it. You can split backlinks by how they were acquired, where they sit on the page, or which rel attribute they carry, and the three cuts produce different lists. This page uses acquisition, which is the split that matters when you are deciding what to spend a week on.

What is the best type of backlink?

An editorial link, inside the body of a relevant page that people actually read, placed because the writer wanted to reference you. That is the answer for almost every site, and it is unsatisfying because it is also the hardest to obtain on demand. The types that scale are the types that carry the most execution risk, and that trade off is the whole subject.

Are guest post backlinks still good for SEO?

They are the clearest example of a type where execution decides everything. A contributed article on a publication your customers read, written because you had something to say, is a good link. The same structural link from a site that publishes forty posts a week from anyone who pays is not, and Google's policy names large-scale article campaigns with keyword-rich anchors specifically.

Are directory backlinks still worth building?

A small number of them, yes. Industry bodies, genuine trade directories and local listings that people use to find suppliers still do something, mostly for discovery and local visibility rather than for authority. General purpose submission directories that exist to sell listings have been worth nothing for well over a decade, and paying for them is the single most common waste in a beginner budget.

Are profile and forum backlinks worth the effort?

Not as a link building activity. Most are nofollow, most sit on pages nobody visits, and building them at volume is a recognizable pattern rather than a strategy. Where they earn their place is as a byproduct: if you are genuinely active in a community, the profile link is fine to have and it is not the reason you are there.

Which types of backlinks last the longest?

Links inside the body of an article that was written for its own sake, because nobody has a reason to remove them and the page has a reason to stay. The fragile end is anything inserted into content that already existed, anything on a curated list, and anything tied to a commercial arrangement, because all three have somebody with a reason to revisit them. We do not publish survival percentages for this since we cannot source them, but the mechanisms are predictable.

Does the backlink type matter more than the quality of the link?

No, and treating the label as the decision is how teams end up with a spreadsheet full of technically diverse links that moved nothing. A relevant partner link from a page with a real audience beats an editorial mention buried on a domain nobody visits. Use the type as a starting prior, then judge the specific page.

Keep reading

Knowing the type is easy. Knowing it is still there is the work.

Recheck every placement on a schedule, catch the day a link goes nofollow or disappears, and report on what you still have rather than what you once earned.

Start free trial See pricing