Agency-mediated links average around $1,459 each — roughly three to four times what the same placement costs direct. Whether that markup is worth paying depends on what sits inside it, and most retainers do not break it out.
Pricing current as of August 2026, drawn from published industry studies. Treat every figure as a band rather than a quote.
An agency proposal arrives at $5,000 a month for ten links. That is $500 a link, which sounds reasonable against published averages.
Then you discover the publishers were paid $164 each, and roughly $3,300 of your monthly spend is service fee. Whether that is outrageous or entirely fair depends on what the service fee bought, and the proposal almost never says.
What agencies actually charge
Per-link pricing. Agency-mediated links have been reported averaging around $1,459 — roughly three to four times direct placement cost.
Retainers. Typically $3,000 to $10,000 monthly, producing somewhere in the region of five to fifteen links a month depending on tier and quality floor.
In-house comparison. A dedicated person plus tooling runs $3,000 to $8,000 monthly in salary and software, before any placement fees at all.
The three converge more than they look. What differs is what you get for the overhead — and the underlying placement costs those figures sit on top of are covered in the how much do backlinks cost breakdown.
What the markup is actually buying
Break the fee into its parts and it becomes assessable.
Placement cost. What the agency pays publishers. Somewhere between $141 and $459 for most tiers, higher for quality placements.
Prospecting and qualification. Finding candidates and vetting them. Genuinely time-consuming, genuinely skilled, and the part most worth paying for if it is done properly.
Relationships. Established agencies hold publisher relationships that produce better rates, faster turnaround, and access to sites that do not respond to cold approaches. This is real value and it is the hardest thing for you to replicate quickly.
Content production. If guest posts are involved, someone writes them. At competent quality that is $150 to $400 a piece.
Project management and reporting. Overhead. Necessary, not valuable in itself.
Margin. Also necessary. Agencies that do not make money stop existing mid-campaign.
The question worth asking: which of these am I paying for, and how much of each? An agency that will break this out is one worth talking to. One that treats the composition as commercially sensitive is usually hiding that placement cost is a small fraction of the fee.
Nine questions before signing
- What is the quality floor, in numbers? Minimum DR and minimum monthly traffic. “High quality sites” is not an answer
- Do you pay for placements? If yes, that is paid link acquisition and the guidelines risk is yours, not theirs. Some agencies are vague here deliberately
- Can I see the sites before placement? If the answer is no, you are outsourcing brand association decisions to someone who does not carry the consequence
- What is the link, exactly? Do-follow or nofollow, body or author bio, one or two, permanent or time-limited
- Who chooses anchor text? And what distribution are they targeting per URL? An agency without a per-target-URL anchor policy is not managing the thing that matters most
- Do you monitor links after placement? And do you report survival rate, or only links built?
- What is the monthly pace, and is it proportional to my current profile? A flat house target applied to a DR 15 site is the wrong number
- Do you work with competitors, and do accounts share partners? Partner pool overlap across an agency’s clients is a real risk and almost nobody asks
- What happens to the relationships when we stop? You keep the links; they keep the publisher relationships. Both parties should understand that going in
Question six is the one that separates most agencies. Reporting acquisition without survival produces a figure that decays after it is sent, which the link building budget breakdown covers as the difference between cost per link and cost per surviving link-year.
What should stop the conversation
Guaranteed numbers or guaranteed rankings. Nobody controls either. A guarantee means either the links are worthless or the guarantee is.
No visibility into placements before they happen. Your domain, your association, your risk.
A fixed link count regardless of your profile size. Ten a month is aggressive for a fourteen-domain site and trivial for a four-hundred-domain one. A house number means nobody looked.
Reporting backlinks rather than referring domains. The flattering metric. Fifty links from one site is one relationship.
No mention of anchor distribution. This is the loudest signal in a link profile and an agency that has not raised it is not managing it.
Extremely low pricing. Sub-$100 per link at volume means the links come from somewhere that sells to everyone. The arithmetic does not work otherwise.
Refusal to name the tactics. “Proprietary methodology” almost always means paid placements on a network.
A worked proposal, decomposed
A real-shaped proposal: $5,000 a month, ten links, twelve-month minimum. Headline figure $500 a link, which reads well against published averages.
Decomposed on plausible assumptions:
- Placement fees — ten placements at roughly $200 average: $2,000
- Content production — if guest posts, six pieces at $250: $1,500
- Prospecting and qualification — perhaps 15 hours at internal cost: $600
- Account management and reporting — 5 hours: $250
- Margin and overhead — the remainder: $650
Nothing in that is unreasonable. A 13% margin is thin for an agency, and the prospecting line is real work you would otherwise do yourself.
What the decomposition tells you is what to negotiate. If you can write the content, the $1,500 line is negotiable. If you have a partner pool already, the prospecting line is. If you want fewer, better links, the placement line moves up and the content line moves down.
What should worry you is a proposal where this decomposition produces a margin line of $3,000 — which happens when placement fees are $100 and no content is produced because the links are insertions into existing posts. That is a $1,000 cost sold at $5,000, and the only way to find out is to ask for the breakdown.
Ask. The answer, or the refusal, is the most informative thing in the whole process.
Agency, in-house, or tools
Agency suits you if: you have budget but not time, you need results faster than hiring allows, you value publisher relationships you cannot build quickly, or you want the work to continue without internal management.
In-house suits you if: link building is continuous rather than a project, you want the relationships and knowledge retained, you have someone who can own it, and your monthly spend is approaching a salary anyway. At $5,000 a month you are already paying for a part-time person without getting one.
Tools and self-managed suits you if: your site is under roughly DR 40, you have not exhausted the free tactics, and your constraint is process rather than capacity. Most sites in this position are paying agencies to do work that would be more effective done directly, because the highest-return early tactics — relationship links, unlinked mentions, lost link reclamation, internal routing — are ones an agency cannot do better than you can.
The honest sequencing: exhaust what only you can do, build a process for what scales, and bring in an agency for what genuinely requires relationships or budget you do not have.
The arrangement worth considering
A pattern that works and rarely gets proposed: pay an agency for the part you cannot do, and run the rest internally.
Digital PR and high-tier placements genuinely need relationships and budget. Relationship links, mention sweeps, lost link recovery, internal routing, and steady exchange do not — they need consistency, which is a process problem rather than an access problem.
Splitting on that line usually costs less than a full retainer and produces more, because the internal half is the half where an agency has no advantage and charges you anyway.
Contract terms worth getting right
Four clauses that matter more than the monthly rate.
Minimum term. Twelve months is standard and it is negotiable. Link building genuinely takes time to show results, so a three-month contract is unrealistic — but so is committing twelve months to an agency you have not seen work. Six months with a break clause is a reasonable middle.
What happens to unplaced links. If you contract for ten a month and they deliver seven, does the shortfall roll over, get credited, or disappear? Rollover is standard and frequently unstated.
Quality dispute process. What happens when they place a link on a site you would have rejected. Agree in advance whether that counts toward the monthly number, because arguing about it after the fact damages a working relationship over something predictable.
Data and reporting ownership. You should have the placement log — partner domains, target URLs, anchors, dates — in a format you keep. Agencies that treat the placement list as proprietary are leaving you unable to audit what you paid for, and unable to monitor it after the engagement ends.
That last one matters more than it sounds. Without the log you cannot check survival, you cannot run an anchor audit, and you cannot tell a new agency what was already built.
What to hold them to
Whatever the arrangement, the reporting should contain five things:
- Referring domains — new this month, lost this month, net
- Survival rate on links older than six months
- Anchor distribution per commercial target URL, with direction
- Referral sessions from placed links, which is the only directly attributable outcome available
- Position movement on the specific targeted URLs, not sitewide traffic
Expressed as cost per surviving link-year rather than cost per link, those five are the honest unit for any acquisition route — which the link building budget breakdown applies across in-house, agency, and self-managed programmes alike.
Set these expectations at the start rather than at the first quarterly review. An agency that already reports on these is a better sign than any case study, and one that resists is telling you what its numbers look like under scrutiny.
