The headline writes itself every twelve months: SEO is dead, killed this time by AI summaries, by zero-click results, by the latest core update. Each obituary is wrong, but each one points at something real. Search has changed. The way people get answers from the internet has changed. And the playbook that worked in 2018 — write a 2,000-word post stuffed with keywords and pile up links — stopped working a long time ago.
What is left after the dust settles is a discipline that looks more like editorial work than tactics. Search engines, for all their imperfections, are getting better at recognizing pages that actually help a reader and pages that are pretending to. The pretending pages get punished. The helping pages keep ranking, often for years, for terms the writer did not consciously target.
What actually changed
Three shifts matter more than the rest.
The first is that volume stopped being a moat. There was a long stretch where publishing more pages, faster than your competitors, was a viable strategy. That window has closed. A site with fifty excellent pages now consistently outranks a site with five hundred mediocre ones, because the algorithm has gotten better at noticing the difference and at trusting the source overall, not the individual page.
The second is that intent matching matters more than keyword matching. The old game was finding a search term, repeating it strategically, and earning the click. The new game is understanding what a person is actually trying to accomplish when they type that term, and giving them the most direct, useful answer to the underlying question. Sometimes that answer fits in two paragraphs. Sometimes it requires a 3,000-word guide. The right length is whatever serves the question.
The third is that AI overviews are absorbing a chunk of low-effort search traffic. If a question has a simple, factual answer, the search engine increasingly answers it itself. The pages that still get clicks are the ones that offer something the AI summary cannot: a strong point of view, a personal experience, a specific framework, a piece of original analysis. Generic pages get summarized away. Distinctive pages do not.
What still works, and probably always will
Strip the noise away and the durable parts of SEO are surprisingly old-fashioned.
- Write about things you actually know. Search engines have gotten very good at detecting expertise, partly through engagement signals and partly through linguistic cues. A page written by someone who has done the thing reads differently from one assembled from other people's pages.
- Answer the question fully and stop. Padding hurts. If the honest answer is 600 words, write 600 words. The dwell-time argument that long pages always win is a myth that has been disproven for years.
- Earn links by deserving them. Link building still works, but the version that works is the version where someone reads your piece and decides, without prompting, that it is worth pointing other people toward.
- Take care of the basics. Fast pages, clear structure, decent internal linking, sensible titles, no broken redirects. None of this is glamorous and none of it is optional.
- Update what you have. Refreshing existing posts that are slipping in rank tends to outperform writing new ones. The hardest work is already done.
The patience problem
The real reason most teams give up on SEO is not that it stopped working. It is that it works on a timeline that does not match how marketing budgets get reviewed. A new post might take six to twelve months to find its place in search results. By the time it does, the team that wrote it has often moved on or lost faith. Channels with faster feedback loops — paid ads, social — get the attention, and the slow-burn channel gets neglected.
This is, ironically, why SEO remains a viable advantage. Most companies cannot stay focused on it long enough to win. The ones that can — that publish patiently, edit ruthlessly, and treat their site like a library being built rather than a content machine being fed — eventually accumulate something competitors cannot replicate without the same years of work.
A starting point that is honest about the cost
If you are beginning, do not aim for fifty posts. Aim for ten that you would be proud to have your name on, on subjects where you actually have something to say. Publish them, share them, and then leave them alone for at least three months. Watch which ones move. Update the ones that almost work. Be willing to delete the ones that do not.
That is not a tactic. It is a posture. The tactic is downstream of it. SEO did not die. The shortcuts did. What is left is a long, useful, deeply unsexy discipline that still rewards the people who take it seriously.