The question gets asked roughly every eighteen months and has been asked since 2012. It is being asked again now because answer engines changed how a share of research happens, and because the quality bar on acquisition rose sharply enough that a lot of what used to work stopped.
Both of those are real. Neither means what the headline version claims.
The short answer
Yes, for pages that compete in a SERP where competitors have more referring domains than you. No, for several situations that are more common than the industry admits.
The rest of this is the detail on both halves.
What genuinely changed
The quality floor rose, hard
Scaled placement on sites nobody reads moved from "useless" to "actively harmful". That is the single biggest change in the last few years, and it invalidated a substantial share of what the industry was selling.
Companies still running that playbook are not standing still — they are accumulating something they will pay to remove.
Mentions started counting alongside links
Answer engines synthesise from sources rather than counting votes. Whether your product appears in a generated category comparison depends on whether you are named, in context, in sources the model draws on.
A link is one way of being named. It is no longer the only way that counts — which raises the value of coverage that passes no equity at all.
The cheap end got more expensive
Market average per quality editorial link went from roughly $350 in 2022 to about $500 in 2026. Meanwhile the inventory below $150 became more clearly identifiable and more clearly risky.
What did not change
Three structural facts, none of which depend on how search results are rendered.
Publishers own the comparison surface. Search any software category and count vendor pages in the top ten — frequently two or three. The rest are round-ups you cannot write your way onto. Appearing in them is a placement problem, and it produces demand whether the click comes from a blue link or a generated summary.
Nobody links to a product page voluntarily. People link to evidence, tools and explanations. The pages that make money remain structurally starved of the thing that makes pages rank, and something has to bridge that.
Your competitors have eighty referring domains and you have nineteen. That gap is countable, it did not appear by accident, and no amount of rewriting closes it.
The honest framing is not "do links still work". It is "does the specific page I need to move sit in a SERP where link count is the differentiator". Sometimes yes, sometimes no, and it is checkable in ten minutes.
The ten-minute check
Take your target term. Pull the referring domain count for the four URLs above you, and for your own page.
| What you find | What it means |
|---|---|
| They have 80, you have 19 | Link problem. Relevant. |
| Numbers comparable, you still lose | Content, intent match or technical. Not relevant. |
| Top four are all publisher round-ups | Different problem — pursue inclusion, not acquisition |
| You have more and still lose | Almost certainly the page. Fix it first. |
Three situations where our answer is no
1. The page will not convert the traffic. Ranking a page that converts at 0.3% is an expensive way to learn the conversion rate. Fix the page; it is cheaper and faster.
2. The SERP wants something you are not. If the top four results are publisher round-ups, acquiring links to your product page is fighting the format. Pursue inclusion instead — different workstream, different budget.
3. Your budget cannot close the gap. Run the arithmetic: gap ÷ monthly output. If the answer is over twenty months, buying placements is the wrong instrument at that budget. Build something citable instead.
We decline roughly one enquiry in five for one of these three reasons. It is the least commercial thing we do.
What the "it's dead" argument gets right
Worth conceding properly rather than dismissing.
A large share of the link building industry sells something that does not work, and has done for years. Volume placement on unread sites, DR-filtered inventory, marketplaces with no vetting. If your experience of link building is that, then within your experience it genuinely is dead — and being told to try harder at it is not useful advice.
The distinction that matters is not between old and new. It is between acquisition that creates a reason for someone to link, and acquisition that buys a slot. The second has been dying steadily for a decade and is now most of the way there.
What the "nothing changed" argument gets wrong
Equally worth saying. Anyone claiming their 2019 playbook still applies unchanged is either not measuring or not looking.
The specific things that stopped working: scaled guest posting, infographic outreach at volume, generic directory submission, and anchor-text-led acquisition. Those were meaningful parts of most programmes and they are gone.
Where it is going
Our working view, stated as a view rather than a fact.
The unit of value is shifting from the link toward the mention — being named, in context, in a source that both search engines and language models treat as reliable. That shift rewards exactly the tactics that were always the best ones: original data, expert commentary, genuine trade contribution, category placement.
It penalises exactly the tactics that were always marginal. Which is a convenient conclusion for us to reach, and also what the evidence appears to show.
The practical implication: a programme built on "would I want this placement if it passed zero equity" is robust to whatever comes next. One built on authority scores is not.