Link building has a graveyard, and it is time to visit it. That is the premise of a Search Engine Journal recap of a webinar with Michael Johnson, CEO and founder of Resolve, hosted by Loren Baker on October 21. In The Link Building Graveyard: What To Bury and What Still Wins, Johnson explained which tactics he would leave behind and which he would fund heading into 2027, and his test was practical rather than moral: a worthwhile link should benefit the brand, the publisher, and the publisher's audience. A tactic that fails that test belongs in the ground.
Johnson's cut list covers the familiar manipulations. Private blog networks go in the ground, because they exist only to pass ranking signals and offer nothing to a reader. Link-swap schemes go too, because reciprocal links arranged for their own sake are a transaction, not an endorsement. Forced commercial anchors go, because anchor text engineered to carry a keyword is a signal of manipulation, not relevance. And generic requests to be added to roundups go, because a publisher asked to add a link for no reason has no reason to say yes.
The common thread is that each buried tactic treats the link as the product. The publisher gets nothing, the audience gets nothing, and only the link buyer benefits. That model worked when search engines could not tell the difference. It fails now, because Google's systems, and increasingly AI systems, are good at recognising links that exist for their own sake. A link report can look healthy while the outreach behind it gives publishers little reason to care, and that is exactly the situation Johnson wants marketers to stop accepting.
What survives is the work that gives a publisher something they cannot get elsewhere. Johnson favoured original research for digital PR, walking through campaigns built around workplace AI concerns and travel costs to show how a dataset becomes a timely headline. He demonstrated an interactive news-desert map built from public data and paired with an explanatory guide, a format that gives journalists both a story and a visual. The pattern is consistent: a unique dataset, a clear finding, and an asset a journalist can use.
Johnson also argued against letting domain metrics make the decision. A high Domain Authority or Domain Rating does not establish topical expertise, a useful audience, or a sensible reason for a publisher to link. He recommended checking topical relevance first, then looking beyond an estimated traffic total to the keywords a site actually ranks for. A placement on a relevant site with a real audience beats a placement on a high-metric site with none, and it is also the placement an AI system is more likely to treat as a genuine endorsement.
Johnson organised the work into a sequence: audit the existing programme, establish a baseline for relevant AI buying prompts, identify worthwhile publishers and resource pages, create the assets those audiences need, and pitch them. The inclusion of AI buying prompts in the baseline is the new part. A link building programme in 2027 is measured not only on rankings but on whether the brand appears in the answers buyers actually ask AI assistants.
He also counted branded mentions as an AI visibility win, which reframes what a placement is for. A mention without a link can still put a brand inside an AI answer, and for AI visibility that may be enough. And he warned that AI-generated prospect lists still need human qualification, because a list built by a model is a starting point, not a campaign. The graveyard, in the end, is not about tactics. It is about intent: build something worth linking to, pitch it to someone it genuinely helps, and let the link be the consequence, not the goal.
Sources: Search Engine Journal, November 2026.
Linkbaits helps you create content that earns citations and backlinks. Start with the free tools and build your first linkable asset this week.
Start free trial