Why core updates keep rewarding authority sites, and what that means for links

By Matija Konjić

Every core update produces the same theatre. Rankings shake, forums fill with traffic-drop screenshots, and a few weeks later the analyses agree on a familiar conclusion: established, authoritative sites came out fine or better, and thin sites with no reputation took the hit.

At some point a repeated outcome stops being news and becomes the system working as designed. This piece looks at why the pattern keeps repeating, what it says about how visibility is actually allocated now, and what a site should do about it, because the answer is refreshingly unexciting.

The pattern nobody should still be surprised by

Google ships several core updates a year, and the post-update winner lists have rhymed for years. Recognised brands, sites with real expertise on display, and domains with strong, clean link profiles hold or gain. Sites built on volume content, aggressive monetisation and borrowed authority slide, each update a little further.

Individual pages move both ways in every update, and any single site can buck the trend. The aggregate direction, though, has been one-way traffic for a long time.

What the recent updates actually showed

The freshest datapoint is the March 2026 core update, where Amsive’s winners-and-losers analysis found Google tilting visibility toward authoritative, brand-owned and official-source domains, category by category. Their read of it: a first-party, official-source correction, with the biggest gains going to names that already carried trust.

That matches the run of updates before it. The helpful content system folded into core updates and kept punishing unbacked volume. Spam updates sharpened the filter on manufactured links while leaving editorial ones untouched. Whatever the label on the release, the sorting principle underneath barely moves: who has demonstrable authority, and who merely publishes?

Two details from the recent cycle are worth pinning down. First, the volatility is real but the direction is not random: sites that gained in one update tended to keep or extend gains in the next, which is what you would expect if a stable underlying quality score were being re-weighted rather than reshuffled. Second, the losers skewed hard toward content whose scale could not be explained by the organisation behind it, the same pattern the scaled-content policies target. Both details point at reputation as the sorting key.

Why the algorithm keeps landing on authority

The consistency makes sense once you look at the problem from Google’s side. Content is now infinite and cheap to produce, so judging documents alone no longer separates reliable from plausible. Reputation scales where reading does not: signals about the source, who cites it, who searches for it, who quotes it, are far harder to fake in bulk than the content itself.

Every one of those themes is a proxy for the same question: does the world outside this website vouch for it? That is why the updates keep rewarding the same profile, and why chasing each update’s rumoured quirks misses the point entirely.

Where links fit in the update era

Links are the oldest of those external votes, and still the most transferable: authority literally moves from one domain to another through them. What has changed is the quality bar. Manufactured link patterns get filtered rather than rewarded, so the value has concentrated in exactly the links that also build brand: editorial placements on relevant, trusted sites, the standard behind what makes a backlink high quality.

Notice the loop in that table. Good placements produce authority, coverage produces brand search, brand search reinforces entity trust, and all three feed back into how the next update treats you. Links stopped being a trick that beats the algorithm and became infrastructure that aligns with it, which is the whole thesis of modern off-page SEO.

The same signals now feed AI answers

The update story would matter less if search were shrinking into irrelevance, but the authority signals travel. AI Overviews, ChatGPT and Perplexity synthesise answers and then cite a handful of sources, and the brands they name skew heavily toward the same well-cited, well-covered names the core updates promote. The overlap is not mysterious: these systems inherit the trust problem and reuse the trust signals, which is why getting cited in AI answers runs through the same work.

Practically, that means authority building now pays on two surfaces at once, classic rankings and AI citations, and the sites collecting update tailwinds are compounding on both.

Who keeps losing, and the common thread

The recurring losers are not small sites, as the folklore has it. They are unbacked sites: domains whose content vastly outweighs their reputation. The imbalance is measurable before any update makes it visible: pages per referring domain, authors nobody can verify, topics with no external footprint. Auditing for it annually is cheaper than discovering it in a winners-and-losers spreadsheet. Volume publishers with no named expertise, affiliates with nothing external vouching for them, and sites whose link profiles are a monoculture of directories and paid networks. Small sites with genuine authority in a narrow niche routinely ride updates upward.

The common thread is imbalance. When updates recalibrate trust, sites holding content they never earned the authority for give visibility back. The fix is rarely another content sprint. It is closing the authority gap the content already outran.

Reading an update like an operator

When a core update starts rolling out, the useful work is observation rather than reaction. Change nothing while it rolls, because rankings oscillate mid-flight and chasing the wobble wastes work on positions that were going to settle anyway. Note the dates, watch your named target pages daily, and wait for the confirmation that the rollout finished before drawing any conclusion.

Once it settles, read the movement in clusters rather than keywords. A drop across one topic cluster with the rest of the site steady points at a relevance or authority gap on that cluster, which is fixable work. A uniform site-wide slide alongside gains for bigger brands in your results usually means the authority bar moved, which is also fixable work, just slower. And check who replaced you before deciding anything: if the new occupants are visibly stronger brands with deeper link profiles, the update told you the price of the keyword, and the response is a resourcing decision rather than a mystery.

Keep a simple log of every update: date, affected clusters, net movement. After three or four updates the log becomes the most honest report of whether your authority trajectory is working, because it shows the trend across recalibrations rather than the noise inside one. Three green updates in a row is a strategy confirmation no consultant can sell you; three red ones is a clear instruction about where the next year of budget belongs.

If an update already hit you

First, diagnose before reacting. Separate a site-wide slide from a cluster-specific one, check whether the losses cluster around pages with weak external support, and compare your referring-domain profile against the pages that replaced you. In most post-update audits we see, the pages that fell were outgunned on authority long before the update, and the update simply enforced it.

Second, resist the two reflexes that make it worse: mass-deleting content and buying a burst of cheap links. Recovery from a core update tends to arrive at subsequent updates rather than gradually, so the window between updates is for building the missing signals, expert authorship where it is absent, editorial links into the demoted clusters, and internal structure that routes existing authority to them. Sites that spend the window that way walk into the next update with a materially different profile. Sites that spend it panicking usually do not.

The update-proof playbook

Nobody can promise update immunity, and the honest version of this playbook tilts the odds rather than eliminating the variance. The work is steady and unglamorous: editorial links from relevant, trusted sites at a consistent pace, digital PR that earns coverage and brand search, named experts behind the content, and internal structure that concentrates the resulting authority on the pages that matter. In practice that means a monthly rhythm a business can actually hold: a few quality placements, one PR angle in the market, and a quarterly review of which clusters still lack the authority to match their content. Depth of profile matters more than any single coup, because updates read the accumulated record rather than last month’s win. The other half of the discipline is refusing the shortcuts that update cycles punish: no bursts of cheap links after a dip, no content sprints into topics the site has no standing in, no chasing whatever signal this quarter’s folklore blames. Sites that hold the boring line for a year tend to discover the update calendar has quietly switched sides. Our link building statistics show the same picture from the data side: quality and relevance beat volume everywhere it is measured.

Sites built that way stop dreading the update calendar. Some quarters the rankings tick up on their own, because the recalibration keeps moving value toward exactly what they have been building. That is the quiet payoff of treating authority as the strategy rather than the workaround, and building it is what our link building service does all day.