Set two budgets side by side: $60,000 a year for a digital PR program versus $60,000 a year for purchased links. One produces a shrinking number of citations on publications that were hard to get and remain hard to copy. The other produces a predictable count of referring domains that devalue on a schedule someone else controls. The comparison is not about ethics. It is about which line item appreciates.
The asset comparison
| Property | Digital PR link | Purchased link |
|---|---|---|
| Acquisition cost | $2,000–$10,000 per earned placement (amortized campaign cost) | $100–$1,500 per placement, invoiced instantly |
| Permanence | Persists; often cited for years | Devalues in waves; network death risk |
| Policy status | Clean by construction | Named violation unless disclosed and qualified (which strips value) |
| Secondary value | Brand citations, journalist relationships, referral traffic | None; disclosure would end the arrangement |
| Audience | Real readers on the linking page | Usually none |
The purchased link is cheaper per unit and more expensive per outcome. That sentence summarizes a decade of penalty cycles.
What the money buys in a digital PR campaign
Follow a real $8,000-per-month program: researchers produce a proprietary data study — original surveys, FOIA requests, public dataset analysis; a designer builds assets journalists can embed; a pitching desk matches the story to reporters who cover the beat. One strong campaign earns 15–60 placements, most of them permanent, several from publications no budget could buy into. The unit economics look slow until the second-order effects arrive: journalists who used the dataset come back for the next one, and the asset itself — a canonical data page — accumulates links for years without further spend.
The purchased-link ledger by contrast: 60 placements at $1,000, all live within a month, all devalued over the following 12–24 months as the networks holding them are classified. To hold rankings constant, the buyer must repurchase continuously — a subscription to a decaying asset, with the renewal price set by the seller.
Why markets still pay for the decaying asset
Three documented reasons, none of them mysteries:
- Procurement metrics. Agencies are hired on "40 referring domains per quarter." Count-based contracts select for count-based products. Digital PR cannot promise a count; purchased links are priced per count. The metric chooses the product.
- Timeline pressure. A funding announcement or product launch needs rankings this quarter. Earned coverage compounds over years; purchased links deliver the appearance of it in weeks. Buyers facing a hard date buy the appearance and inherit the decay.
- Attribution simplicity. "We bought 60 links for $60K" fits a spreadsheet. "We earned 22 links and a journalist relationship that will produce links for three years" does not. Finance departments systematically undervalue assets that lack invoices.
The compounding difference, modeled
Run a simple three-year model on the same $180,000. Purchased: 180 links at $1,000, assuming 50% survive each year — a generous survival rate for the category — yields roughly 20 surviving links at month 36, all carrying policy exposure. Digital PR: six campaigns at $30,000, each earning 20–40 permanent placements with partial year-over-year accumulation, yields a conservative 250–400 surviving links at month 36, zero exposure, plus the journalists' and the assets' continuing yield. The gap widens every year because one column compounds and the other subtracts.
Related stories: The Honest ROI Math of White-Hat Links: $10,000 In, Four Revenue Paths Out · Niche Edits: The Market for Smuggling Links Into Other People's Rankings.
What the second-order effects are worth
The ledger's hardest-to-price column is also the purchased market's absence. A data study that journalists cite positions the brand's researchers as sources for the next cycle: reporters facing deadlines call people they trust, and trust is built in the first campaign's delivery. Documented second-order effects across mature programs include repeat citations without new pitching, journalist-sourced inbound requests for data, speaking and podcast invitations that trace to coverage, and hiring lift from candidates who cite the campaign as their first exposure to the company. None of these accrues from network inventory, which by design leaves no relationship, no audience, and no asset behind. There is also a defensive dividend: a profile dominated by editorial citations from real publications is structurally resistant to the periodic devaluation waves, because there is nothing to classify — the profile is made of the signal itself. The purchased alternative builds a profile that is, in classification terms, entirely signal-shaped noise. When the wave comes, the earned-heavy site loses little because it never held any manufactured value. That resilience is worth pricing, and it never appears on an invoice.
Making the comparison in your own board deck
Organizations comparing the two budgets should force the comparison into columns that cannot be gamed. Column one: acquisition — placements delivered, with URL, publisher, and disclosure status. Column two: survival — the same list re-run quarterly, live links only. Column three: audience — third-party traffic to each linking page. Column four: policy status — earned, disclosed-qualified, or in-violation, stated per link. Purchased-heavy programs produce impressive column one and collapse down the page; earned programs start smaller and grow across columns two through four every quarter. Presenting all four columns also changes the vendor conversation: sellers of network inventory cannot fill column two without documenting their own decay, and agencies that resist the format have told you which column they were planning to skip. The four-column table, updated quarterly, is the entire governance framework for link spending — and it costs nothing to adopt before the first dollar moves.
How to protect yourself
- Fix the contract metric. Replace "number of referring domains" with survived-link velocity, linking-page traffic, and placement quality tiers — metrics the purchased market cannot fake cheaply.
- Model survival, not acquisition. Any link investment case should project month-36 survival, not month-1 delivery. Demand the vendor's own survival history.
- Budget for assets, not placements. A canonical data study or free tool outlives every campaign that seeded it; placements are the exhaust, not the engine.
- Insist on the audience test. Every placement should have human readers on the linking page. Purchased inventory fails this test almost by definition.
- Blend, but label. A disclosed, qualified sponsored placement is a legitimate media buy with an ad budget — never mix those invoices with the organic-link budget, or the reporting becomes a scheme ledger.
The two budgets buy different things: one buys appearances that decay on the seller's schedule; the other buys assets that compound on yours. Everything else in the debate is the sales literature of the first column.
For more context, read The Honest ROI Math of White-Hat Links: $10,000 In, Four Revenue Paths Out.
For more context, read link price ladder.
For more context, read google link spam update.