AdvancedLinkTraining.com logo — a free link building course by Bill HartzerAdvanced Link TrainingA resource by Hartzer.com

PlacedHigh risk

Paid links and sponsored posts

Money changes hands, a link appears, and the only genuinely reliable outcome is that you now have a recurring bill.

Verdict: Against Google's published policy, sold on metrics that are frequently manufactured, and rented rather than owned - the downside is asymmetric and the upside decays.

What paid links and sponsored posts are

A paid link is any link where money, goods or services moved in exchange for the link itself passing ranking credit. A sponsored post is the packaged version: you pay a publisher, an article appears on their site with your brand in it, and somewhere in the body is a link pointing at a page you want to rank.

Google's position has not moved in over a decade and is published in plain language: buying or selling links that pass ranking credit is a link scheme. The policy is not that you may never pay for placement. Advertising is legal, normal and older than search engines. The policy is that a paid placement must be qualified so it does not pass credit - rel="sponsored" or rel="nofollow" on the link - and, separately from anything Google requires, disclosed to readers because advertising regulators say so.

So the honest framing is this. Paying a publisher to write about you and mark the link as sponsored is advertising, and it is fine. Paying a publisher to leave the link unmarked so that it passes credit is the scheme. Everyone selling into this market knows the difference, because the unmarked version is what commands the premium - and the size of that premium is the clearest evidence that everyone involved understands exactly what is being sold.

How it works, step by step

The supply side explains everything downstream. A publisher with an audience discovers a second revenue line alongside display advertising: selling placements to SEOs. It needs no sales team and no audience growth. Within a year the site has an inventory desk, and a decision has been made that its outbound links are no longer editorial. Between that publisher and the buyer sit brokers, whose function is aggregation - thousands of sites presented as a searchable catalogue with a metric and a price attached to each. The buyer never speaks to the publisher, and that frictionlessness is why the market grew.

The transaction follows a predictable sequence:

  1. The buyer selects a site from a catalogue, filtered by an authority metric, a claimed traffic figure, and a topical category.
  2. The buyer supplies a target URL and, almost always, the anchor text they want. The anchor is the real product. Nobody pays a premium for a brand-name link they could have got by asking.
  3. Content is produced - by the buyer, the broker, or a language model - to a length the publisher will accept.
  4. The publisher publishes, usually without a sponsorship label, sometimes under a fictional byline.
  5. A live URL is returned as proof of delivery, and often a clock starts on a placement that expires.

What it costs in time and effort

The effort is close to zero, which is the commercial appeal. Someone with no relationships and no understanding of the subject can acquire fifty links in a month. Every legitimate tactic in this course requires a skill; this one requires a card.

Price is driven by five things, and they tell you what the market believes it is selling. The host domain's claimed authority metric, which price tracks steeply - meaning the metric is the product, and anyone who can inflate a metric can inflate a price. Claimed traffic, usually a tool estimate rather than an analytics figure. Vertical, with finance, gambling, health and legal at multiples of a hobby niche. Whether the link is followed and undisclosed. And whether the deal is a one-off or a rental.

That last one is the cost nobody prices properly. Much of this inventory is rented: stop paying and the link comes down, so you have not bought an asset but a subscription that runs for as long as you want the ranking. Even placements sold as permanent decay. On my own profile - 1,301,839 links from 22,260 referring domains, none of them bought - the median referring domain stops linking after roughly 1,080 days, and 49.6% of all links ever recorded are gone. A purchased link is less durable than an earned one, because the publisher has no editorial reason to keep it and a commercial reason to resell the slot. Add perpetual monitoring, and a purchased profile costs more to maintain than an earned one.

When it works and when it does not

Let me be accurate rather than pious. This market exists at its current size because it sometimes moves rankings, particularly in weak competitive fields and on sites with thin profiles. Pretending otherwise damages the credibility of everything else on this page.

Here is what is also true. The most common outcome is not a penalty but silence - the links are discounted algorithmically and the buyer concludes they need to buy more expensive ones. Search engines have spent twenty years learning to ignore rather than punish, because punishing links you do not control creates an obvious attack vector. Most of the money in this market buys nothing, and the buyer never finds out which half.

The failure modes stack up:

  • The authority you are paying for is often manufactured. A domain with a good Trust Flow (Majestic's measure of how trustworthy the sites linking to a page are) may have got it from a network rather than from readers. There is a ten-minute check, below, and almost no buyer runs it.
  • You are buying from a site that sells to everyone, including your competitors.
  • The pattern is legible from outside. Not from any one link, but from the profile: exact-match commercial anchors, a referring domain set with no brand mentions in it, and a first tier where every referring domain sends precisely one link. Buying one at a time produces exactly 1.00 links per referring domain, and that is a footer statistic in a public tool.
  • The exposure is asymmetric. The upside is a ranking improvement you might have earned another way. The downside is a manual action requiring you to document and remove links you paid for - which means explaining the whole program to whoever signs the invoices.

Common mistakes

These are not tactical refinements. They are the reasons buyers lose money even by their own standards.

  • Believing the metric on the listing. Authority scores are computed from inbound links, and inbound links can be bought. A listing metric is a claim about a claim.
  • Buying exact-match commercial anchors. The anchor is what buyers want and also the loudest signal in the profile. An earned profile's top anchors are the brand, the bare URL, article titles and empty anchors from images. A bought profile's are the phrase somebody wanted to rank for.
  • Keeping no records. If you ever need to document removals, you need a register of every URL, date, cost, anchor and contact. Buyers who kept nothing spend weeks reconstructing it under pressure.
  • Assuming a nofollow makes it fine and never checking. Attributes get stripped in template migrations, and plenty of vendors promise a qualified link and deliver a followed one.

A worked example

Take a catalogue listing of the kind that lands in your inbox weekly: a plausible niche, a strong authority score, a price that feels like a bargain against the metric. Here is the qualification pass, offered as diagnosis rather than as a shopping guide.

  1. Read the metric pair, not the metric. Trust Flow beside Citation Flow (Majestic's measure of raw link volume). Trust well below citation means many links and little trust - volume purchased rather than earned.
  2. Scan the neighbours for a signature. In my own site's second tier sits a run of seven domains - hotfri.com, mantaw.com, zomatt.com, zlutag.com, dotpim.com, yelpad.com, ylutag.com - all with Trust Flow between 33 and 36 and Citation Flow between 48 and 50. Nonsense names, metrics identical to the point. Independent sites do not converge on the same two numbers. When a prospect's authority comes from a set like that, it is borrowed.
  3. Read the outbound pattern. Open six recent posts. If each carries a followed link to an unrelated commercial site, you are looking at inventory, and your link joins a queue.
  4. Check the site against a second client's profile. A network sells to everybody, so it appears in everybody's graph. Those seven domains turned up in two unrelated sites' link graphs, and that co-occurrence is what a network looks like from outside.

Running those five checks on a catalogue disqualifies most of it. That is the useful output: not a shopping list, but a demonstration that the inventory is largely manufactured.

How to measure it

Measurement here is risk accounting. Keep a placement record: live URL, linking domain, date, amount paid, anchor text, link attribute as delivered, vendor. Re-check quarterly for four things - is the link still on the page, still followed, still indexed, and has the page been redirected or folded into an archive. That check exists because links disappear silently. Across 645,202 lost links on my own profile, 97.4% were lost while the source page was still perfectly reachable. Purchased links behave the same way, only faster.

Then measure the thing vendors never report: referral traffic. If a year of placements produced no sessions, you did not buy publicity, and the only thing on offer was the ranking credit the policy says you should not be getting. Finally, hold a control set of earned links from the same period and compare survival, traffic and ranking movement. Most programs never do this, which is why the belief that it works survives contact with the data.

The verdict

Buying links is against Google's published guidelines, sold on metrics that are frequently manufactured, delivers an asset you usually rent rather than own, and produces its most common result - nothing - so quietly that buyers rarely learn from it. The risk is not that you will certainly be penalized; most buyers are not. It is that you build a growth channel with a switch in somebody else's hand, on an expense that must be renewed forever, while the reason nobody links to you voluntarily goes unaddressed.

If you have inherited a profile full of purchased links, do not mass-disavow. Read the guidance on manual actions first: the tool is a scalpel for a named problem, not a cleaning product. And if you are being pitched, the useful question is not how much. It is: tell me what this site is for.

Questions

Is buying a link always against Google's guidelines?

Paying for a link that passes ranking credit is. Paying for advertising or a sponsored placement is not, provided the link carries rel="sponsored" or rel="nofollow" so it does not pass credit, and the placement is disclosed to readers. The distinction is the attribute and the disclosure, not the payment.

What usually happens if you buy links?

Most often, nothing. The links are discounted algorithmically and the money is simply gone. Manual actions do happen, particularly to sites buying at obvious volume with commercial anchors, but silent devaluation is the far more common outcome - which is why buyers so often conclude they need to spend more rather than stop.

Are sponsored posts worth it for traffic rather than rankings?

Sometimes, and that is a legitimate reason to buy one. Judge it as advertising: does the publisher have an audience that matches yours, and did the placement send sessions? If a year of placements produced no referral traffic, you were never buying an audience, whatever the invoice said.

How do search engines detect purchased links?

Not usually one link at a time. Patterns give it away at profile level: exact-match commercial anchors dominating, referring domains that each send precisely one link, sites whose outbound links point to unrelated commercial verticals, and clusters of domains with near-identical authority metrics that could only come from a shared link source.

Should I disavow links I bought years ago?

Not reflexively. The disavow tool is for a specific, documented problem - typically a manual action naming links you placed. If you have no manual action and no evidence of harm, a large speculative disavow file is more likely to remove value than to restore it. Read the manual actions lesson first.