"Link building service", "link building agency", "link marketplace" and "freelance link builder" are used more or less interchangeably in sales material. They describe four different operating models with different economics, different failure modes and a price range of roughly fifteen to one.
Knowing which one you are buying is more useful than comparing the numbers on the proposals.
The four models
| Model | What you buy | Good at | Fails when |
|---|---|---|---|
| Marketplace | Inventory from a catalogue | Speed, transparency of price, one-off needs | You need vetting, strategy or accountability |
| Service (productised) | Fixed monthly output to a published spec | Predictability, auditability, budgeting | You need strategy owned end to end |
| Agency (retainer) | Bespoke programme with strategy attached | Complex situations, multi-channel, senior input | Scope drifts quietly and cannot be audited |
| Freelancer | One person's relationships and time | Deep niche access, low overhead, flexibility | Capacity, redundancy, infrastructure |
Marketplace
A catalogue of sites with prices. You choose, you pay, the link appears. Fast and honest about what it is — you are buying inventory, and the vetting is yours to do.
Reasonable for a one-off need or for a team with the expertise to vet. Bad as a programme, because nobody is deciding which links you should be acquiring, only which are available.
Productised service
Fixed scope, published price, defined deliverables. You know before signing how many placements you get, what each must satisfy, what happens when one breaks and what the blended unit price is.
Suits teams that already own content and technical SEO and want acquisition handled predictably. The limitation is real: a service executes a specification, it does not own your strategy.
Agency retainer
Bespoke, strategic, usually more senior input. Right when the situation is genuinely complicated — a damaged profile, a category with no clear SERP structure, multi-market work, or where link building is one part of a wider mandate.
The characteristic failure is opacity. Scope flexes without anyone deciding to flex it, and by month four nobody can reconstruct what was delivered.
Freelancer
A specialist with genuine relationships in your niche can outperform a mid-tier agency comfortably, because you are buying their relationships rather than an average. Below roughly ten placements a month this is frequently the best value available.
The trade-offs are predictable: no redundancy, limited capacity, and you will maintain the tracking sheet. The failure mode is hiring a generalist with no relationships in your category — then you are funding their education.
Which to pick
Marketplace — you have in-house expertise and want to buy specific placements you have vetted yourself.
Service — you own strategy internally and want acquisition delivered predictably against a spec you can audit.
Agency — the situation is complex, or you need someone to own the thinking as well as the doing.
Freelancer — under ten placements a month in a narrow vertical where one person's relationships are the asset.
What a service should include as standard
Regardless of model, these should not be extras.
A gap analysis before anything is purchased
Competitor referring domains minus yours, ranked. Without it, the monthly volume in your proposal came from the supplier's capacity rather than your situation.
A published quality specification
Written thresholds for traffic, relevance, publishing history, editorial identity and footprint. "We only work with high-quality sites" is a sentiment, not a specification.
A verification record per placement
Live URL, anchor, surrounding paragraph, rel attribute, referring-page traffic, index date. Written on publication day, into something you can open on any Tuesday.
A replacement warranty
Twelve months, with defined triggers and replacement rather than credit. Link decay is predictable; a supplier who does not address it has moved a known cost onto you silently.
Four things worth refusing to buy without
- Billability tied to live and indexed. Not to outreach sent, not to placements "secured". This single clause reallocates most of the execution risk.
- Disclosure of paid placements. Some placements involve payment; round-up inclusion typically runs $350–$750. Fine. Not knowing which is not fine.
- Ownership of the gap analysis and target list. If those stay with the supplier, switching means paying someone else to rebuild what you already funded — six to nine months of work.
- The rejection log. What a supplier refuses to buy describes their standard far better than what they deliver.
What things cost, for calibration
Independent 2026 pricing surveys put the average cost of a quality editorial link at roughly $500, up about 45% from $350 in 2022. Credible SaaS placements run $150–$500; premium publications $700–$1,500+. Guest posts average $220–$609, link insertions around $141, category round-up inclusion $350–$750.
Monthly programmes commonly quote about $6,000 for 16 links, $10,000 for 27 and $15,000 for 40. At seed stage, $1,500–$3,000 a month realistically buys five to ten quality links.
Below roughly $150 per placement the arithmetic stops closing. Nobody can prospect, score, pitch, write a publishable draft and compensate an editor at that price — so the inventory is coming from somewhere else, and that somewhere else is what you are actually buying.
The question that sorts most shortlists
Ask each supplier: "Show me the specification a placement has to satisfy before you will invoice me for it."
Some will send a document. Some will describe a process. Some will explain that quality is assessed case by case by experienced people, which is a way of saying there is no specification.
All three answers are informative, and you will know which model you are dealing with within one reply.