Guide
Link exchanges and reciprocal links: when a swap stops paying
Reciprocal links are not banned. They are budgeted, and almost nobody says what the budget is, because saying it honestly means admitting there is no published number. Here is the version we can defend: the limit is a share of your backlink profile rather than a count, a swap costs something a one-way link does not, and whether any individual swap earned anything is a question you can answer by reading the page.
What counts as a link exchange
The plain form is an agreement: you link to me, I link to you. The common form in software is a badge. A directory lists your product and asks you to put its badge on your homepage, and now two sites point at each other. A footer swap, a "partners" row, a mutual blogroll and a badge are the same transaction wearing different clothes.
None of that is inherently a trick. Two sites in the same field linking to each other because a reader benefits is ordinary web behaviour, older than search engines. What Google’s spam policies name is excessive link exchanges, and the whole argument lives in that word.
The number nobody has, including us
You will find articles giving a cap. Fifteen. Twenty. Ten percent. Ask where the figure came from and the trail ends immediately, because Google has never published one, and a threshold that cannot be sourced is not a threshold.
We know how those numbers get made because we made one. An earlier version of this product told makers that roughly fifteen badge slots was the budget and past that they were over. It sounded specific, it fit the interface, and it was reverse-engineered from a single product’s badge count at the time it was written. It was not derived from anything. We removed it, and the rule that replaced it is the one below, which has the advantage of being checkable.
The honest limit
A ratio, not a count. Reciprocal links should stay a minority of the sites linking to you, and specifically a minority of the dofollow ones, since a dofollow reciprocal link is the manipulative signal a nofollow one is not. Practitioner consensus puts that comfortable ceiling somewhere under a quarter. It is a rule of thumb, and it is labelled as one.
A share beats a count for a reason that has nothing to do with search: it scales with you. Six swaps against eight total backlinks is a profile made of trades. Six swaps against two hundred is a rounding error. A fixed number cannot tell those apart, which is why every fixed number eventually gives bad advice to somebody.
What a swap actually costs
A one-way link costs you the work of earning it. A swap costs that plus a permanent slot on your homepage, and homepage slots are the scarcest inventory you own. Every outbound link there divides the same finite authority, so the tenth badge takes a share of what the first nine were passing along, and the row of logos gets longer while each logo is worth less.
The slot is also the part you cannot take back cheaply. Removing a badge can cost the listing, because some directories recheck and delist quietly. So a swap is closer to a standing contract than a purchase: you are agreeing to keep paying rent on your homepage for as long as you want the link.
When a swap is still the right trade
At the very beginning, and where there is no free lane. A product with no backlinks has no ratio to spoil, and a swap is the fastest dofollow link a completely unknown site can get. That is a real advantage and the advice that says never trade links ignores it.
The test afterwards is simple and it is about supply, not about a cap. If a directory will list you without a badge, take that lane: it gives you the same link without spending a slot. Trade a slot only where the badge is the only way in. That rule needs no threshold, and it caps your swaps naturally, because the directories that demand a badge are a finite set.
The Domain Rating trap
Swaps make metrics move, which is what makes them so easy to over-trade. Ahrefs’ Domain Rating is computed from inbound links; it has no view of what your homepage links out. So every badge you accept counts fully toward the number, while Google may be passing nothing back through the link you gave in return.
Say that plainly to yourself before reading a rising chart as proof: a higher DR after a swap tells you a link pointing at you exists. That was never in doubt. It says nothing about whether the trade earned anything.
How to tell whether a swap earned anything
Stop reading the agreement and read the page. Three things have to be true at once, and each one fails silently on its own:
- The anchor is there and it is dofollow. A rel="nofollow", ugc or sponsored passes nothing. Check a page’s link.
- The page is indexable. A dofollow link on a page carrying a noindex is the most convincing failure in this whole category: it looks like a win in every tool that only reads rel. Check for noindex.
- The crawler is allowed to fetch it. A path disallowed in the directory’s robots.txt is never read, so the link on it is never seen. Check robots.txt.
Most makers have never read their own swaps this way, and the ones that failed look exactly like the ones that worked: the badge is up, the listing is live, and something in the markup quietly says no. Read one of your own listings, no account needed.
One more thing that only shows up if you look: some directories do not link to you directly at all. The button points at a counter they control, which redirects onward. That can be a perfectly good link, or it can be one carrying a header that strips it, and the anchor looks identical either way.
What this page cannot tell you
The ratio above needs a denominator: every site linking to you, not just the ones you know about. Reading a full referring-domain profile means an index like Ahrefs’ or Semrush’s, and the cheap sources under-report small sites badly enough to be worse than no number.
So ShipDR computes the share against the dofollow links it has verified for you, which is a smaller, honest denominator that grows only when you go and earn one-way links. It is stated as that, never as a measurement of your whole profile. If a tool quotes you a precise reciprocal percentage of your entire backlink graph, ask where the graph came from.
Questions people actually ask
Are reciprocal links against Google guidelines?
Not in themselves. Google’s spam policies name "excessive link exchanges" as a link scheme, and the load-bearing word is excessive. Two sites in the same field linking to each other because it helps a reader is ordinary web behaviour and always has been. What the policy is aimed at is volume and intent: links traded for the purpose of manipulating rankings, at a scale no editorial reason explains. Nobody outside Google can tell you where that line sits, which is exactly why any specific number you read is invented.
How many link exchanges is too many?
There is no published number, and anybody quoting one is guessing. The useful question is a share rather than a count: what fraction of the sites linking to you also carry a link from you. Practitioner consensus puts the comfortable ceiling somewhere under a quarter of your dofollow referring domains, which is a rule of thumb rather than a measurement. A share also scales with you, so it keeps making sense as you grow, while a count stops being meaningful the moment your profile changes size.
Do badge swaps count as reciprocal links?
Yes. A badge on your homepage pointing at a directory that lists you is a link exchange with a graphic on it. It costs a permanent slot on your most linked page, it is visible to anyone auditing either site, and at scale it reads exactly like what it is. That is not a reason to avoid them, it is a reason to spend them where a directory offers no other lane.
Are reciprocal links worth it for a brand new site?
Often yes, and this is the one case where the usual advice inverts. A product with no backlinks has no ratio to ruin, and a swap is the fastest dofollow link a nobody can get. The mistake is treating the cold start as the strategy: the first handful are the cheapest links you will ever get, and each one after that adds less while the share it costs you grows.
Does a reciprocal link still pass value?
Usually something, and it is not the binary people expect. Google is not obliged to discount a link because one points back, and a genuinely useful mutual link on a real page can carry weight. What changes with volume is the pattern, not the individual link. So the honest answer is that one swap is a link, forty swaps are a signature.
Why do link exchanges raise my Domain Rating so easily?
Because Ahrefs’ Domain Rating reads inbound links only. It has no view of what you link out, so every swap counts fully toward the number while Google may pass nothing back through it. This is worth saying plainly: a rising DR is not evidence that a swap worked. It is evidence that a link exists pointing at you, which was never the thing in doubt.
How do I tell whether a swap earned anything at all?
Read the listing page, not the agreement. Three things have to be true together: the anchor is there and it is dofollow, the page is not carrying a noindex, and the crawler is allowed to fetch it in the first place. Any one of the three failing makes the slot on your homepage a donation. The free checkers here read each of them, and ShipDR re-reads all three on a schedule, because a listing that passed last month can quietly fail this one.
A swap you stop checking is a slot you stopped charging rent for.
Badges get quietly delisted, rel attributes get tightened when a directory rebuilds its template, and a listing page can pick up a noindex without anyone telling you. ShipDR reads every listing you track on a schedule, tells you the day an answer changes, and shows which of your homepage slots are still earning.
- Every listing rechecked on a schedule, not once
- Live, dofollow and indexable, read from the page itself
- Told the day a backlink drops or flips
Adding a website and charting its Domain Rating is free, no card. The backlink engine is $10 once, per website, with no subscription.