What Google's link spam policy actually says
Google does not ban reciprocal links. Its link spam policy targets excessivelink exchanges — the link-to-me-and-I'll-link-to-you arrangement done for ranking — and partner pages that exist exclusively for cross-linking. Two words in that policy carry all the weight:
- "Excessive" — the occasional mutual link between genuinely related sites is ordinary web behaviour, not spam. No number is published, which means the safe interpretation is that arranged pairs should stay a small minority of your profile.
- "Exclusively" — a partners page that exists to reciprocate links is called out by name. A partners page that describes real integrations, with links as a by-product, is not the same page.
The practical test is intent made visible: would this link exist if it passed no ranking value? A profile where the answer is mostly yes has nothing to fear from its reciprocal pairs.
When reciprocal links are fine
- Genuine partnerships and integrations — you link to a product you build on, and its makers link back to a customer story.
- Tools and services you actually use writing about you, and you citing them in return.
- Industry associations, local business groups, and event pages where mutual linking reflects a real relationship.
- A small number of relevant swaps with founders in your space — modest volume on topically related pages reads as ordinary.
The common thread is relevance and restraint. Every healthy backlink profile contains some mutual pairs; none is dominated by them.
When link swapping becomes a risk
- Scale — templated outreach proposing swaps turns a handful of pairs into a pattern an algorithm can rank you by.
- Irrelevance — a fitness blog and a crypto tool trading links looks arranged at any volume, because no natural path explains the pair.
- Dedicated partner pages — a page of outbound links whose only function is reciprocation is the exact artifact the policy names.
- Link sharing groups — Slack, Discord, and Facebook swap groups produce overlapping, mutual-heavy link graphs across every participant, a collective footprint no single member can see from their own profile.
How to audit your reciprocal footprint
Three steps, with any backlink tool you already use:
- Export your referring domains, and separately list the external domains your site links out to.
- Intersect the two lists — every domain appearing on both sides is a mutual pair.
- Judge the pattern, not the count: pairs explained by real relationships are fine; pairs clustered around one swap group or an unrelated niche are the ones to prune before building anything new.
Loops: the value of swapping, without the pair
The three-way exchange — A links to B, B to C, C back to A — is the old workaround, and its mechanic is sound: nobody links back to the site linking to them, so no reciprocal pair exists. What kills most ABC deals is process. They are negotiated ad hoc, nobody screens relevance, and nobody re-checks the links a month later.
LaunchPact's Backlink Network runs the same shape with the process enforced. Claimed founder domains are quality-scored into authority tiers 1–5 and matched in loops of 3–5 sites — same or adjacent niche, within two tiers. The matcher refuses any two sites that already link to each other in either direction, so it cannot create a reciprocal pair even by accident. Each placement is a one-sentence do-follow insertion, relevance-screened, then approved and placed by hand by the owner of the hosting site, and crawl-verified at day 7, 30, and 90. Credits are earned by giving links — never bought — which is what keeps sites that exist only to sell links out of the pool.
1:1 swap vs loop exchange
| Factor | Loop exchange | 1:1 reciprocal swap |
|---|---|---|
| Reciprocal pair created | Never — the matcher blocks existing links in either direction | Always — that is the deal |
| Detectable footprint | No mutual edges to intersect | Both halves of every swap are public |
| Relevance | Niche clusters + screening on every sentence | Whatever the two parties agree to |
| Verification | Crawled at day 7/30/90, credit clawback on removal | Nobody re-checks |
| Scaling behaviour | More gives, more credits — pairs never accumulate | Every additional swap deepens the pattern |
Related guides
Frequently asked questions
Two sites linking to each other — site A links to site B, and site B links back to site A. The pair can arise naturally (partners mentioning each other, tools you use citing you back) or from an arranged link swap. Search engines treat the natural and arranged versions very differently at scale.
Not on their own. Any site with real relationships accumulates some mutual links, and Google does not penalise the pattern in ordinary volume. What Google's link spam policy calls out is excessive link exchange — swapping arranged purely to manipulate ranking, and partner pages that exist only for cross-linking. The question to ask is not whether you have reciprocal links but whether mutual pairs dominate your profile.
A relevant link from a real site passes value in both directions of a swap — that part still works, which is exactly why the tactic persists. What has changed is detection: mutual pairs are the easiest link scheme to spot algorithmically, because both edges of the exchange are public. Swaps between topically related sites in modest volume keep working; templated swap campaigns are a footprint.
Google's link spam policy describes it as link-to-me-and-I'll-link-to-you schemes and partner pages made exclusively for the sake of cross-linking. No published number separates acceptable from excessive, which is the practical problem: the safe reading is that arranged mutual links should stay a small minority of your profile and always sit on topically relevant pages.
An arrangement where site A links to site B, site B links to site C, and site C links back to site A — no mutual pair, so the reciprocal footprint disappears. The mechanic is sound; the usual implementations are not, because ad-hoc three-way swap deals skip relevance vetting and nobody re-checks whether the links stay live.
The shape is the same — loops of 3–5 founder sites where nobody links back to the site linking to them — but the process is enforced rather than negotiated. The matcher only pairs same or adjacent niches within two authority tiers, refuses any two sites that already link to each other in either direction, screens every proposed sentence for relevance, and requires the receiving site's owner to approve and place the link by hand. Crawls then verify each placement at day 7, 30, and 90, with credit clawback if a link is removed.
Link sharing groups concentrate the two things that make reciprocal linking dangerous: no relevance vetting, and a shared pool of participants who all link to each other. Profiles built from the same group develop overlapping, mutual-heavy link graphs — a collective footprint no individual member can see. If you use one, track every swap and check whether the same domains keep reappearing on both sides of your profile.
Export your referring domains from any backlink tool, export the external domains you link out to, and intersect the two lists. Each domain on both sides is a mutual pair. There is no official safe ratio, but if arranged pairs are more than a small fraction of your referring domains — or clustered in one niche-irrelevant group — prune the weakest before building anything new.
No — avoiding them is the design. Credits are earned by placing one-sentence do-follow insertions for other founders and spent requesting links from different founders in the loop, so no mutual pair ever forms. Giving and requesting are free (3 domains, 10 concurrent placements, 12 gives per domain monthly); Pro ($19/mo) adds 1.25× credit earn, matching priority, and a 24/month give cap, with a 14-day trial on your first domain claim.
Exchange links without the reciprocal footprint
Claim your domain, give one sentence, and request owner-approved links from loops — never 1:1 swaps.
Claim your domain free