The March 2026 core update finished rolling out two weeks ago. We have now seen enough recovery and post-update data to talk about what actually changed and what site owners should do next.
This update was big. SISTRIX flagged it as the most volatile core update since the September 2023 helpful content rollout. Semrush Sensor hit 9.4 on a 10-point scale across the US, UK, and Australia. The pattern of winners and losers also tells a clearer story than usual.
What Google said
Google announced the core update on a Tuesday morning. The official rollout window was 17 days. The blog post was short and largely repeated the standard line: this is a routine improvement to the ranking systems and there is nothing specific that affected sites should change.
That part is technically correct and practically misleading. Core updates do not target specific sites, but they do shift the weighting of signals. The pattern of which signals matter more after a core update is what we can analyse.
What we are seeing in the data
Three patterns came through clearly in our analysis of around 6,400 sites that we track via Search Console snapshots and third-party rank tracking.
Original research and first-hand experience won
Sites with original research, case studies, and content built from real-world testing held or gained ground. Sites built from rephrased information that already exists on the web lost ground.
This is not new advice, but the magnitude is bigger this time. A site we work with that publishes one original benchmark study a month gained 38 percent organic traffic during the update. A competing site that summarises news without testing anything dropped 22 percent.
Author signals matter more
Posts with named authors who have other published work in the same niche performed better. Anonymous, “team” by-lines, and brand-only by-lines lost ground in How-To and informational queries.
Schema markup helped here. Sites with proper Author schema (with sameAs links pointing to LinkedIn, Twitter, or other publication profiles) saw smaller drops than sites without it.
Affiliate and review content was hit hard if it had no first-hand testing
The “best [tool]” listicle pattern was re-evaluated. Pages with original screenshots, video reviews, or measurable test data held position. Pages that read like “here are 10 tools we found on G2” dropped sharply.
This pattern lines up with what Google has been signalling for two years. The update enforced it more strictly.
Forum and community content gained
Reddit, Stack Overflow, and Quora appeared more often in informational queries, sometimes pushing publishers down a position. This is the continuation of the trend that started with the March 2024 update.
For publishers this is hard to fight directly. The lever you have is original research and unique data that forums cannot match.
Who recovered
Sites that lost traffic in the September 2023 helpful content update and made specific structural changes are starting to see partial recovery in this rollout.
The pattern of recovery suggests that Google now revisits previous demotion decisions during core updates. This is consistent with what John Mueller has said in office hours over the past year.
If you were hit in 2023 and have been making improvements (real authors, original research, removing AI-generated thin content), you may have moved up in this rollout. If you have not made structural changes, you probably did not.
What to do if your site dropped
The instinct is to panic and revert recent changes. Resist it. Here is the order we work through with clients.
Pull the data first. Get your Search Console performance report for the last 90 days. Compare positions and impressions for your top 50 queries before and after the update window. Identify which queries lost the most.
Look for query patterns. Are the queries that dropped all informational? Commercial? Product-specific? The pattern points to the issue.
Check for content depth gaps. Pull the URLs that lost the most. Compare them to the top 3 ranking results for the same query. Do the top 3 have things your page does not? Original data, deeper coverage, fresher information, real screenshots, expert quotes?
Check author signals. Is the URL bylined? Does the byline link to a real author page? Does the author page have a real bio, links to social profiles, and other published work?
Audit thin content. If you have published programmatic or AI-assisted content with low original value, this update probably devalued those pages. Decide whether to improve them or noindex them entirely.
Do not mass-delete content. Removing 30 percent of your site in a panic is the wrong move and usually makes things worse. Slow, deliberate improvements work better.
What we are doing on our own sites
For our SEO clients, this is the playbook for the next 60 days:
- Add original research to the top 10 informational pages
- Add proper author schema across the whole site
- Replace any AI-generated thin pages with full rewrites or noindex tags
- Review programmatic pages for first-hand value
Most sites we manage have either held position or recovered. The pattern is consistent: the more original signal you produce, the more durable the rankings.
Next core update
Google does not pre-announce core updates, but the pattern of the last three years suggests another core update will land in 6-10 weeks. Spam updates can land at any time. We will cover both as they happen.
TL;DR
The March 2026 core update rewarded original research, real authors, and first-hand testing. It punished thin AI content and unattributed listicles. If your site dropped, do not panic. Pull the data, identify the query patterns, fix the content depth gap, and add author signals. Forum content gained share against publishers. The Sept 2023 helpful content losers who made structural changes are starting to recover.