The March core update was a referendum on aggregators

Lily Ray's Amsive analysis dropped yesterday and the headline finding is being passed around as "Google hates UGC now." That's not what the data says. It's a lazier reading than the data deserves, and it leads to exactly the wrong conclusion for anyone trying to figure out what to do about it.

Here's what actually happened between 27 March and 8 April. YouTube lost 567 visibility points — roughly 30% larger than Wikipedia's drop in December. Reddit dropped 64. Instagram 48. X 46. TripAdvisor down 45. Yelp 33. Expedia 33. Indeed 18. WebMD 9. Mayo Clinic 6.

And on the other side: Hilton, Hotels.com, Trivago, NPS.gov, airport websites, USAJobs.gov (+16%), Disney Careers (+59%), CVS Health Careers (+45%), GoodRx (+55%), NIH.gov.

The pattern across travel, jobs, health, finance, and entertainment is the same one. Sites that aggregate, list, or talk about other people's stuff lost ground. Sites that own the thing being talked about gained it.

That's not Google deciding it dislikes user-generated content. It's Google deciding that when there are two pages on the same topic — one from the platform people use to discuss the thing, and one from the company that actually owns the thing — the second one should rank higher. Those are very different statements with very different implications.

The aggregator tax is finally being collected

For most of the last decade, the aggregator was the smartest position to occupy in search. You didn't need to make anything. You needed to organise other people's stuff, slap a clean UI on top, and let SEO do the rest. TripAdvisor, Yelp, Indeed, WebMD, and a hundred smaller versions of the same playbook all built billion-pound businesses on this trick. The original product was somewhere else; the visibility lived with them.

The aggregator's job has been absorbed into the SERP. What's left is the original.

Google was complicit in this. Aggregators tended to have more inbound links, more page-level engagement signals, and more topical depth than any single first-party site could produce. So they ranked. The hotel chain wrote one page about its property in Edinburgh; TripAdvisor had 4,000 reviews and a forum thread. Of course TripAdvisor won.

What's shifted is that AI Overviews and AI Mode now do the aggregation themselves. Google doesn't need to send the user to TripAdvisor to read 4,000 reviews. It can summarise the reviews directly in the result. The aggregator's value-add — "we've collected what other people said about this thing" — is now being performed by the model in the SERP.

So when the question becomes "what should rank underneath the AI summary?" the answer increasingly becomes "the source of truth for the thing itself." The hotel's own site. The employer's careers page. The government domain. The brand. Because that's the layer the AI summary couldn't replicate from another aggregator.

The aggregator's job has been absorbed into the SERP. What's left is the original.

YouTube losing 567 points is the most interesting number in the dataset

Everyone is going to focus on Reddit and Yelp because those are the obvious villains. The YouTube number is the one that should make you stop.

YouTube isn't an aggregator in the same sense as Reddit. It's a platform, sure, but the content on it is genuinely original — millions of creators making things that exist nowhere else. Yet it lost more visibility than any other domain in the dataset, and 30% more than Wikipedia did in December.

Ray notes the drop "returned YouTube's visibility to its level before the early March surge, not to a new low." So this is partially a correction. Fine. But it still matters that Google chose to correct *down* rather than let the surge stand. Google is signalling that even genuinely original platform content sits structurally below first-party brand content in the new hierarchy.

The implication for anyone running a small business: hosting your video content on YouTube and embedding it on your site is now a strictly weaker SEO move than hosting it natively, captioning it natively, and letting Google index your own page. YouTube as a discovery surface is still useful. YouTube as a place where your content lives instead of on your site is an increasingly bad bet.

The bounce-back data is what the cynics will use to dismiss this

Reddit and Indeed both recovered shortly after the update window closed. People are going to seize on this and say the whole pattern is noise — that Google did one thing during the rollout and another after, and there's nothing reliable to extract.

That's a misread, but a tempting one.

Core update windows are when Google's intent shows clearest. The bounce-backs happen when other parts of the system rebalance — fresh links land, query patterns shift, the model recalibrates. The signal of intent is the direction Google leaned during the rollout itself. Indeed and Reddit recovering doesn't mean Google didn't try to demote them. It means the demotion didn't fully stick under post-rollout conditions. Those are different things.

If you're a first-party brand site, the read is: Google is actively trying to lift you over aggregators, and the lift is partial and inconsistent, but the direction is unmistakable. If you're an aggregator, the read is: even when you bounce back, the ceiling is lower than it was, and the next update will lean the same direction.

The German data is the tell

The bit that should remove the last doubt is the SISTRIX analysis of German results. Same pattern. Online shops and utility sites lost ground. Official websites and brands held up. This isn't a US algorithm quirk. It's a global tilt in how Google evaluates the topic-to-source relationship.

When the same pattern appears across two markets, two methodologies (SISTRIX visibility and DataForSEO categorisation in Amsive's work; SISTRIX direct in Germany), and aligns with Cyrus Shepard's separate analysis of 400+ sites looking at year-on-year traffic — that's three independent studies with three different methods landing on the same finding. The SEO industry usually demands less evidence before declaring a trend.

What this actually means if you run a small or mid-sized business

Most of the small business sites I look at sit somewhere between "first-party brand" and "thin aggregator." They sell or deliver something real, but their content strategy was built on aggregator logic — long lists, comparison pages, "best X for Y" content designed to capture intent rather than to actually do the thing.

The implication isn't that you need to delete the comparison content. The implication is that the comparison content was never your moat. Your moat is the thing you actually do. Your service. Your product. The pages that prove you can deliver it. The pages that let someone book, buy, contact, or progress.

In my experience, the businesses that look most like first-party brand sites in Google's eyes share four traits. They have pages that allow task completion (book a call, get a quote, start a trial, place an order). They have proprietary content that nobody else can produce — case studies, original photography, owner bios, methodology pages. They have a tightly defined topical focus rather than a sprawling content map. And they have enough brand demand that some percentage of traffic searches for them directly.

If you have all four, the March update probably helped you, and you might not have noticed because your visibility moved up against competitors who got demoted rather than against your historic baseline. If you have one or two, you're vulnerable to the next update leaning further in this direction. If you have none, the long-tail informational traffic you're holding onto is borrowed time.

The fix isn't a checklist. It's a strategic decision about what your site is for. Most small business sites are still trying to be both a brochure and a content marketing engine, and ending up doing neither well. The March data suggests Google is increasingly happy to reward the brochure half — the part that proves you exist, deliver, and own the thing — while pulling support from the content marketing half unless it's tightly bound to genuine expertise and brand demand.

That's a real shift in what "good SEO" means for the average UK service business. Most of the strategy decks being shown to clients right now haven't caught up with it.

The thing nobody is saying out loud

Google has spent twenty years complaining that the open web is full of low-quality aggregator content gaming its system. The March core update is the clearest signal yet that Google is willing to do something structural about it, even at the cost of demoting domains it has historically protected.

That's a meaningful editorial position from a search engine. It's also one that benefits Google enormously, because the AI summary layer needs first-party sources to cite — not aggregators citing aggregators citing the original. The cleaner the signal from "thing exists" to "thing is described," the better AI Overviews and AI Mode can ground their answers.

Read that way, the March update isn't really about aggregators. It's about preparing the index for a world where Google itself is the aggregator, and everything underneath needs to be the source.