Link swaps that never reciprocate (and how to structure them)
We placed our side of a link swap weeks ago. Two dofollow links, live, on real pages that get real traffic. We’re owed eight links back.
We’ve received zero.
Not one, weeks later, from a counterparty who answers messages promptly and has never once disputed that the links are owed.
I’m publishing this while it’s still unresolved, because every article about link exchanges is written afterwards by someone presenting a tidy outcome. This one has no outcome yet. Carelessness would be an easier explanation, but what actually happened was a sequence of small reasonable decisions, each defensible alone, that added up to a position nobody would choose deliberately.
And it isn’t a one-off. I have a second swap outstanding with a different partner right now, same shape: our side placed, their listings and referral link still not delivered. Two for two on the same mistake, which is how I know it’s the structure rather than the counterparty.
Why swaps look so appealing
A paid placement means an invoice, a payment method, a budget line and, in a company of any size, somebody’s approval. A swap means none of that. Two people agree, both place a link, everybody gets an asset, no money moves.
It feels free. That’s the whole attraction, and also the problem, because the word “free” is quietly doing an enormous amount of work there and nobody audits it.
A cash deal has one unambiguous moment. Money arrives or it doesn’t, both sides know instantly, and the relationship reorganises around that fact. A swap has no such moment. It has an intention, and intentions decay.
The structural flaw
A swap creates an obligation with no due date and no enforcement.
Think about what you actually hold if the other side never delivers. No invoice to chase. No payment to reverse. No platform holding funds. No contract worth enforcing over a few hundred dollars of value. You have an email thread and a feeling of being owed something.
So the counterparty’s incentive to complete is entirely social. They’ll do it if they like you, if it’s easy, and if it stays near the top of their list. The moment it slips below their fifth priority it stops existing. Not from malice. It just falls off.
What went wrong on ours
We placed both of our links first, in full, before receiving any of theirs.
I did it because the conversation was warm, because the counterparty is a real company with a public reputation to protect, and because placing our side was a five-minute job I could do immediately while theirs needed coordination across several properties. All true. None of it mattered.
By placing first I converted our position from a negotiation into a request. Before placement I had something they wanted. After placement I had a favour to ask. Those are different bargaining positions and I handed myself the weaker one voluntarily.
The retroactive-credit trap
There’s a subtler version I also fell into.
One of our two links predated the swap entirely. An existing editorial link, put there months earlier because it genuinely belonged in that article. When the swap was agreed, it got counted as part of our consideration.
That sounds efficient. It’s a small disaster, for two reasons. We gave up something already given freely, converting a genuine editorial mention into a chip we’d already spent. And we inflated what our side appeared to be worth without doing anything new, which makes the eventual imbalance harder to argue about honestly.
If you’re counting an existing link as swap consideration, you aren’t contributing anything to the deal.
Nobody here is dishonest
It would be easy and satisfying to write this as though they cheated us.
As far as I can tell, no cheating occurred anywhere. They’re a real business with real priorities and this deal sits below all of them. The person who agreed it isn’t the person who has to execute across four properties, and what they still owe includes exact replacement copy and a specific target URL that requires somebody to sit down and decide.
That’s precisely what makes it worse. If they were dishonest I’d know where I stood and would pull my links today. Because they’re simply busy, I’m left in the uncomfortable middle, where escalating feels disproportionate and waiting feels like being taken for a ride.
What a swap link is actually worth
Most people treat swaps as though both sides are automatically equal. Put a cash number on each link instead.
If you’d have sold that placement for $250, giving it away in a swap costs you $250, and the link you receive is a fair trade only if you’d have paid something similar for it.
Do that arithmetic and the numbers are often lopsided in a way nobody said out loud. Eight links from four small properties can be worth considerably less than two links from one article that actually ranks. Quantity is the easiest thing to concede in a negotiation and the least meaningful thing to receive.
Count value, not links.
Swaps often arrive bundled
Partner deals in this space rarely arrive as pure link exchanges. They come bundled with a discount code for each other’s audience, a listing on a partners page, sometimes a referral arrangement.
That bundling makes the deal much harder to hold anyone to, because there’s no clean unit of completion. Is a live discount code worth two of the eight links? Nobody defined it, so when the links don’t arrive the other side can point at what they did deliver and you’re both arguing about an exchange rate that was never agreed.
If the deal has several components, write down what each is worth and which are conditional on the others.
Structure it properly
What you want is escrow: a neutral party holds both sides and releases together. That doesn’t exist for links and isn’t going to. The closest substitute is sequencing, and it costs nothing.
One for one, alternating, never more than one placement ahead. They place one, I place one, they place the next. If the deal is eight against two, that ratio still gets executed in alternating single steps rather than as two batches.
If somebody insists on receiving the whole batch before starting theirs, that’s your answer. A partner who won’t go one for one is telling you what the completion rate will be.
Start with the smallest, least valuable placement on both sides. The first exchange tests whether this counterparty executes at all, and you want that test cheap. Save your good pages for after they’ve proven they ship.
Written terms for a deal with no money
People skip paperwork on swaps precisely because no money is involved, which is backwards. The absence of money is the reason you need terms written down.
Put four things in an email: the exact URLs on both sides, the anchors, a date by which each side completes, and what happens if one side doesn’t, which in practice means the other side’s links come down.
That last clause changes the character of the whole thing. It converts a vague intention into a deal with a defined failure state, and gives you something to point at that isn’t an accusation.
I sent no such email. I had a warm chat instead.
Track swaps in the same ledger as paid deals
Record swap obligations in exactly the same file where you record placements you’ve sold.
A swap has two sides and both are trackable: what you owe them, what they owe you, dates against each. If it lives in an email thread, the only thing prompting you to check is memory, and memory is what fails here. I went several weeks before sitting down and actually counting that we were two links out and eight short.
It also does something psychological. It turns a friendly arrangement into a line item with a status, and a line item reading “outstanding” for a month is much easier to act on than a vague sense of being owed.
Mine now carries a status and a date for every side of every deal, paid or swapped. The two outstanding swaps are the oldest unresolved rows in it, which is uncomfortable to look at. That’s the point.
When to pull
So why haven’t I removed our two links yet?
Partly because the relationship has value beyond this deal and I’d rather have the links back than have the argument. Partly because ours sit in articles where they’re honestly reasonable editorial choices, so removing them is a small act of vandalism against my own pages.
But mostly because I haven’t yet sent the message that says here is the deadline and here is what happens after it. That message is the actual next step, and I’ve been avoiding it, which is the same failure of nerve that produced the situation.
If you take one thing from this, take that. The difficulty with swaps is rarely the structure. It’s that enforcing them feels rude, so nobody does, so they don’t complete.
The one that worked
One partner swap this year completed cleanly, both sides live, no chasing.
The difference was small and entirely structural: two links against two, decided in one message, both sides placed within about a day of each other. Small, symmetrical, fast.
The ones that fail are large, asymmetric and spread across a lot of properties. Eight against two coordinated across four sites is a project, and nobody assigned it an owner.
The terms sheet I use now is here.