Industry averages make terrible budgets. The figure you need is cost per surviving, usable link at your own quality floor — and it is materially higher than any published number. Here is how to build the model.
Reference pricing current as of August 2026, drawn from published industry studies. Treat every figure as a band rather than a quote.
Most link building budgets are set by taking a published average, multiplying by a target link count, and rounding up.
That method fails in month two, reliably, for a reason that has nothing to do with the average being wrong. It fails because the average describes a transacted price for a single successful placement, and your budget has to cover attempts that fail, links that do not meet your standard, and placements that disappear.
Three multipliers the headline figure leaves out. Model those and you get a number you can plan against.
Start with the target, derived rather than chosen
Do not pick a round number. Derive it.
- List your three to five commercial target URLs. The pages with genuine business value, not your homepage
- For each, find the pages currently outranking you for the term you want
- Check their referring domains — at page level, not domain level. In Ahrefs set the mode to Exact URL. This is the number that matters and reading the domain-wide figure will inflate your target several times over
- Calculate the gap between their page-level count and yours
- Divide by your timeframe in months
That gives you a monthly requirement grounded in something. If the arithmetic produces a number you cannot afford at any of the routes below, that is useful information — it means the target is wrong rather than the budget, and the better move is picking less competitive terms and returning to the hard ones with a larger base.
Set the floor before you see any prices
Written down, in advance: minimum DR, minimum monthly traffic, acceptable niches, and maximum outbound linked domains on a partner site.
The order matters. People who set the floor after seeing prices set it wherever the prices are comfortable, which is not a standard — it is a rationalisation.
Two of these are worth more than DR and cost less to satisfy. Minimum traffic catches rebuilt expired domains that carry authority with no audience. Maximum outbound domains catches sites that sell to everyone, which is the check that most predicts whether a link holds its value.
The three multipliers
Multiplier one: pass rate. What proportion of prospects clear your floor.
If you filter genuinely hard, most will not. Published analyses of guest post opportunity quality report that only a small fraction meet a serious bar — with one analysis finding over 85% of sites accepting contributions fall below DR 40 with under 10,000 monthly visitors.
Assume you review ten and accept two. Your cost per accepted link includes the review time for all ten.
Multiplier two: conversion rate. Of the prospects that pass, what proportion become live links.
This varies enormously by tactic. Relationship links convert at half or better. Resource page pitches convert modestly. Broken link building converts poorly. Paid placements convert at close to 100%, which is the actual thing you are paying for.
Multiplier three: survival. What proportion is still live in a year.
Links disappear through redesigns, content pruning, expired domains, and quiet SEO cleanups. If you have no data, assume attrition and be pleasantly surprised. This is the multiplier that converts a one-off price into a running rate.
The model, worked
A DR 32 B2B site. Target: 40 new referring domains over twelve months, derived from a page-level gap on two commercial URLs.
Route A — paid placements only.
- Target: 40 usable links
- Pass rate at the quality floor: 1 in 5 prospects reviewed
- Placement cost at the quality band: $600 average, per the bands in the cost per link breakdown
- Review time: 5 minutes per prospect, 200 prospects = 17 hours at $60 = $1,000
- Assumed survival at 12 months: 80%, so 50 placements to hold 40
- Total: 50 × $600 + $1,000 = $31,000, or $775 per surviving link
Route B — mixed programme.
- Relationship sweep — 12 links, one afternoon, $300 in time
- Unlinked mentions and lost link recovery — 6 links, two afternoons, $600
- Exchange programme — 18 links over the year, roughly 1.5 hours each including review, $1,600 in time plus platform cost
- Paid placements for the gap — 8 links at $600 with the same 1-in-5 filtering, $4,800 plus $400 review time
- Survival across the mix: assume 85%, so build 47 to hold 40
- Total: roughly $8,500, or $212 per surviving link
The gap is not because the mixed route is cleverer. It is because the first eighteen links in Route B cost almost nothing, and Route A pays full price for links that were available for the asking.
Four line items most budgets omit
Review time. Vetting prospects that you reject is real cost. At a hard quality floor it is the largest hidden line, and leaving it out is how a budget that looked adequate runs out in month two.
Content production. If guest posting is in the mix, competent 1,500-word pieces run $150 to $400 each, or four to six hours of internal time.
Tooling. $100 to $500 monthly for a backlink tool, and you need one to derive the target and run the audits.
Monitoring and audit time. Roughly ninety minutes quarterly. Small, and it is the line that determines whether the survival number in your model is real or assumed.
The unit that makes the model usable
Every published price is cost per link, which is wrong because links are not permanent.
Cost per surviving link-year is the honest unit. A $400 placement lasting four years costs $100 a year. A $164 placement gone in eight months costs $246 a year — under half the sticker price and 2.5 times the running cost.
You cannot know which you bought at purchase time. You can measure it afterwards, and almost nobody does, which is why most reported cost-per-link figures quietly inflate every month after they are calculated.
The link building budget breakdown covers how that unit changes the comparison between acquisition routes, and it changes it considerably.
What the number should look like at your stage
Rough guidance, because a budget with no reference point is hard to sanity-check.
Under DR 20. A meaningful budget here is close to zero in cash and real in hours. The tactics that work — relationship links, mentions, lost link recovery, internal routing — cost time. Spending $3,000 a month at this stage buys execution of work you should be doing directly, and buys it badly.
DR 20–40. Tooling plus a small paid allocation. $200 to $500 monthly in tools, and paid placements only for specific gaps the free tactics cannot reach. Most sites in this band overspend, usually out of impatience rather than analysis.
DR 40–60. The band where sustained spend starts justifying itself. A mixed programme with a genuine engine, an asset in development, and selective paid placements. $1,000 to $4,000 monthly is defensible depending on competitiveness.
Above DR 60. Digital PR and high-tier placements become reachable and worth their cost. Budgets here are driven by competitive position rather than by anything on this page.
The test at every stage: if you cannot name what the last $1,000 bought in referring domains that are still live, the budget is not being managed. It is being spent. That distinction is why the backlink pricing 2026 breakdown treats survival as part of the price rather than as a reporting extra.
How to allocate what you have
Whatever the total, the sequence is the same and it is not the one most budgets follow.
First, spend nothing. Internal routing, the relationship sweep, unlinked mentions, lost link recovery. These produce links that cost only time and they are consistently skipped because they are unglamorous. On a site that has never done them, this alone frequently produces ten to twenty referring domains.
Second, build the engine. One tactic you can run every week without heroics, where effort does not scale linearly with output. This is the line item that determines whether the programme continues past month three.
Third, build one asset. Original data, a tool, a definitive reference. Highest cost, highest return, longest timeline — start it early so it ships around month three.
Fourth, buy the gap. Paid placements for what the above cannot reach. This is the last line, not the first, and budgeting in this order routinely halves the total.
Reviewing the model
Quarterly, replace assumptions with data:
- Actual pass rate against your floor
- Actual conversion by tactic
- Actual survival on links older than six months
- Actual hours spent, including review and audit
The first model you build is guesswork with a structure. The second one, built on a quarter of real numbers, is a budget — and it is usually lower than the first, because the free tactics outperform their assumptions and the paid ones underperform theirs.
That pattern is consistent enough to plan around. Build the model, run it for a quarter, then rebuild it on what actually happened.
