The March 2026 Google core update finished rolling out in under 20 hours.
For context: the August 2025 spam update took almost four weeks. A typical broad core update takes seven to fourteen days. Twenty hours is not a normal rollout speed, and the reason it moved that fast tells you something important about what Google is targeting and how it found those targets.
This post covers what changed, who got hit, and what you need to fix.
Why it rolled out in under 20 hours
When Google deploys an update this fast, it almost always means the same thing: the targets were pre-computed.
Google’s SpamBrain system runs continuously in the background, scoring and flagging pages long before any enforcement action. The March 2026 update appears to have been an enforcement trigger: Google pulling the lever on a list of sites that SpamBrain had already identified and queued.
The update did not need to crawl and re-evaluate the web. The decisions were already made. The rollout was just the application.
This matters because it means there was no warning period, no gradual rollout to monitor. If your site was on the list, the drop happened immediately.
What actually got hit
The update has three distinct targets. Understanding which one applies to you determines what you need to fix.
1. Scaled content abuse
This is the primary target. Sites that used programmatic SEO to generate thousands of thin pages (pages that essentially repackage what is already on the SERPs without adding anything new) are seeing the largest drops.
The pattern looks like this: you build a template, you feed it a list of keywords, and you publish ten thousand near-identical pages. Each page is technically unique but functionally identical. Google has been warning about this for years. The March 2026 update is the enforcement.
If your site has a large number of pages that follow the same template with minimal variation in content, this is the section to pay attention to.
2. The Information Gain filter
This is the more nuanced, and more widely applicable, target.
Google has deployed what is being described as a Gemini-powered semantic filter that evaluates whether a page adds something genuinely new to the internet. The technical term is “information gain.” The practical question the filter asks is: if this page did not exist, would the internet be missing anything?
Pages that score poorly are not necessarily AI-generated, thin, or technically broken. They are pages that cover ground already covered by dozens of other pages, in the same way, without original data, unique perspective, or genuine expertise.
This is the filter that is hitting the most legitimate sites. A well-written, properly structured page can fail the information gain filter if it adds nothing beyond what is already ranking.
3. Parasite SEO devaluation
The third target is more specific: third-party content sections on high-authority domains. The model (sometimes called parasite SEO) works by publishing content on a trusted site’s subdomain or subfolder to inherit its authority. Google has been aware of this for some time and the March 2026 update applied a significant penalty to sites using leased subfolders and third-party review sections.
If your content lives on a domain you do not control, or if your rankings dropped despite your own content being solid, this may be the cause.
The AI Overviews signal
One thing that has surprised people watching the data: some sites losing traditional blue-link traffic are seeing an increase in AI Overview citations.
This creates an important distinction. Google is now running two separate evaluation processes: one for ranking in traditional results, and one for what the AI summarises. The criteria overlap but they are not identical.
Sites that are losing rankings but gaining AI Overview presence tend to have strong structured data, clear question-and-answer content patterns, and E-E-A-T signals (expertise, experience, authoritativeness, trustworthiness). These are sites that Google trusts to summarise accurately even if it no longer ranks them as highly in traditional results.
The implication is that optimising only for traditional rankings is no longer sufficient. A site that appears in AI Overviews for high-intent queries may capture the buyer even without a top-three organic ranking.
What to check on your site
The March 2026 update introduces signals that most SEO tools do not yet check for. Here is what actually matters now:
Information gain signals
Your page needs to demonstrate that it adds something to the internet. Google evaluates this through a combination of signals:
- Author presence: is there a named author or clear attribution on the page? Pages with no author signal read as anonymous filler.
- Date signals: is there a visible publication or update date? Fresh, dated content signals active maintenance.
- Original data: does the page contain statistics, numbers, findings, or research that did not come from somewhere else?
- Structural depth: does the page contain genuinely developed content: comparisons, lists with real detail, examples, counterarguments?
A page that scores zero on all four of these is a candidate for the information gain filter regardless of how technically clean it is.
AI Overview eligibility
To appear in Google AI Overviews, your pages need to be structured in a way that makes them easy to summarise. The two most actionable things you can do:
- Add FAQPage, HowTo, or QAPage schema to your key pages
- Structure your H2 and H3 headings as real questions your customers ask
Both of these serve double duty: they help with traditional structured data rich results and they increase the likelihood that Google’s AI will pull from your page when composing an overview.
The basics still apply
The March 2026 update did not change the fundamentals. Pages still need a canonical tag, a proper meta description, correct heading structure, image alt text, Open Graph tags, and clean schema markup. These are table stakes. The new signals are additive: they matter in addition to the basics, not instead of them.
A practical checklist
Run through this on every important page:
- 600 or more words of original, developed content
- Visible author name or clear attribution
- At least one date signal (published or updated)
- One or more original data points or statistics
- Three or more structured list items
- FAQPage, HowTo, or QAPage schema, or at least two H2/H3s phrased as questions
- Meta description between 120 and 155 characters
- Canonical tag present
- Robots meta not set to noindex
- Open Graph tags: title, description, image
- At least one internal link to a related page
This is not a comprehensive SEO checklist. It is specifically the set of signals most relevant to the March 2026 update.
What this means going forward
The March 2026 update is a signal, not an isolated event. Google has been moving toward quality evaluation based on genuine contribution for years. The difference is that the tooling has now caught up with the intention. The Gemini semantic filter is good enough to evaluate information gain at scale. SpamBrain is good enough to pre-identify and queue enforcement targets.
Future updates will be faster and more precise. The gap between building low-value content and being penalised for it is shrinking.
The sites that survive and grow in this environment share a common characteristic: they publish things that are genuinely worth reading, written by people who know what they are talking about, structured in a way that both humans and machines can understand.
That is not a new standard. It is the original standard, finally being enforced at scale.
Dablin’s SEO Audit now includes two checks built specifically for the March 2026 update: Information Gain score and AI Overview eligibility. It runs 15 checks on any page URL and generates AI fixes for every issue it finds. Free to start at dablin.co.
See if ChatGPT, Gemini, and Claude mention your brand. Results in 20 seconds.
Try Dablin →