Google Updates
What Actually Moved After the March 2026 Core Update
Most of what got published about this update was speculation from a handful of screenshots. Here is what I measured across sites I actually have data on.
Key takeaways
- Sites with thin, templated content saw the clearest declines across the tracked set.
- Sites with original data or first-hand testing were largely stable or grew.
- No consistent technical signal (page speed, schema, etc.) explained the movement on its own.
The question
Core update write-ups tend to appear within 48 hours of a rollout, based on a handful of screenshots and a lot of speculation. I wanted to wait for real before-and-after data across sites I actually track.
How I tested it
I compared rankings across 12 sites I have direct access to, both owned properties and client accounts, in the two weeks before and after the rollout completed. I categorized each site's affected pages by content type and depth.
What I found
The clearest pattern across the 12 sites: pages with thin or templated content, the kind assembled from a formula rather than original reporting or testing, saw the most consistent declines. Pages backed by original data, first-hand testing, or clearly demonstrated experience held steady or grew.
I did not find a consistent technical signal, page speed, schema presence, or similar, that explained the movement on its own. The pattern tracked much more closely with content originality than with anything technical.
What this means in practice
If you saw a decline after this update, I'd look first at whether the affected pages demonstrate real, original work, rather than assuming a technical fix will resolve it. I checked common technical culprits first — see the AI crawler behavior analysis for a related look at technical crawl factors — before concluding originality was the stronger signal here.
Where this could be wrong
12 sites is a small sample for a Google-wide update, and my categorization of "thin versus original" content involved some subjective judgment on my part. Broader third-party analyses may find different patterns.
References
- Before and after ranking comparison, 12 owned and client properties. Internal tracking data, March 2026.
Frequently asked questions
Did this update target a specific content type?+
Not cleanly. The pattern that held across my tracked sites was originality of content, not a specific format or technical factor.
Should I have panicked if I saw a drop right after this update?+
A single update-timed drop is rarely enough to diagnose the cause. I'd look at content originality and depth first before assuming a technical penalty.
Get notes like this before they're published.
SubscribeMore from Research
Building a Repeatable Way to Measure AI Citation Rate
Most citation rate numbers people quote are one-off screenshots. This is the tracking methodology I built so the number means something over time.
Jul 2026
Why AI Search Doesn't Always Follow Google Rankings
I assumed the page ranking first would get cited first inside AI Overviews. After pulling 400 queries, that assumption barely held up.
Jul 2026