Google July 2026 Core Update: AI Content Farms and Thin Affiliates Are Getting Surgically Removed

I had a client panic-call me yesterday. Their site — a well-established affiliate review blog they'd been running since 2019 — had lost 60% of its organic traffic overnight. Not over a weekend. Overnight. One day they were ranking comfortably for a stack of commercial keywords. The next morning, they weren't. Sound familiar?

That's the July 2026 Core Update doing exactly what Google said it would do. This one isn't a broad, ambiguous "quality" shuffle like we saw with older core updates. SEOs across the community are calling it surgical — Google has clearly deployed new classifiers that are specifically trained to identify content produced at scale using AI without meaningful human editorial oversight, and thin affiliate pages that add little beyond a commission link and a table.

If you've got skin in this game, here's everything I know right now about what's getting hit, why, and what you can actually do.

What's Actually Getting Hit

The pattern emerging from the SEO community is pretty consistent. The sites taking the hardest hits share a few common traits:

The defining characteristic of what's getting hit isn't just "AI content" in a blanket sense. Google clearly has enough signal now to distinguish between AI content that's been properly edited, fact-checked, and enriched with genuine expertise versus content that went straight from the LLM to the CMS with nobody in between. The latter is getting nuked.

Thin affiliate sites are catching it too — specifically those review pages where the "review" is basically a product description scraped from Amazon plus a star rating widget and a Buy Now button. If your page doesn't add anything a user couldn't get directly from the manufacturer's site, Google's new classifiers appear to be flagging it as commodity content.

The New Classifiers: What We Think We Know

Google doesn't publish its algorithm internals, obviously. But based on what's happening in practice and some reading between the lines of recent Google documentation and Search Central talks, here's what seems to be working against affected sites:

1. Content Velocity vs. Editorial Depth Ratio

Sites publishing 50–200 articles per month with no visible editorial team are getting flagged. Google has always said content velocity shouldn't matter if quality is high, but I think they now have classifier signals that can detect when that quality isn't actually there — even if individual pages look decent in isolation. The pattern across the whole site is the tell.

2. Author Entity Signals

Sites with anonymous authors or fake persona bylines are dropping hard. Meanwhile, sites with real named authors who have verifiable entities — LinkedIn profiles, social accounts, published works, speaker credits — are holding or gaining. Google's entity graph has gotten sharp enough that it can evaluate whether "Sarah Johnson, Senior Content Strategist" actually exists anywhere on the web.

3. The Affiliate-to-Content Ratio

Pages with excessive affiliate links relative to original content are being demoted. This isn't new policy — Google has been warning about it for years — but the enforcement mechanism has apparently become much more aggressive. One affiliate link in a 2,000-word honest review is fine. Fifteen affiliate links in a 600-word summary of products you've never touched is not.

Who Survived — And Why

Here's what I find genuinely interesting about this update: it's not a blanket hammer on AI content. I'm seeing sites that use AI extensively in their content workflow that are completely unaffected — or gaining. The difference is editorial oversight and genuine value add.

One site I track closely — a B2B SaaS review publication — uses AI for first drafts on every piece they publish. But they have a team of actual practitioners who review, edit, add real-world context, and often rewrite entire sections. Their traffic is up 18% week over week during this update. Google can't tell (and shouldn't care) whether a human wrote the first draft or an AI did. What it cares about is whether the final product adds something real.

Sites with strong first-hand experience signals are also doing well. Reviews that include photos of the actual product, specific usage scenarios that reflect genuine testing, complaints as well as praise, comparisons that go beyond manufacturer specs — that content is holding up. Fluff that hedges everything and commits to nothing is not.

Your Recovery Checklist

If you got hit, here's where I'd start. This is not a "wait it out" situation — core update recoveries that happen passively tend to be partial at best. The sites that recover fully are the ones that make substantive changes between now and the next core update.

What Not to Do Right Now

I want to spend a minute on the wrong moves, because I'm seeing panicked responses in SEO communities that are going to make things worse for the people making them.

• Don't mass-delete pages without a plan. Pruning works when done strategically. Randomly deleting anything that lost traffic will remove pages that might have recovered naturally.
• Don't swap one AI tool for another. Rewriting your AI content with a different AI model doesn't solve the underlying problem. The issue is editorial oversight, not which LLM you used.
• Don't buy links to try to offset the quality hit. A link profile overhaul won't counteract a content quality demotion. You're solving the wrong problem.
• Don't submit reconsideration requests. This is a core update, not a manual penalty. Reconsideration requests don't apply and won't help.
• Don't wait for Google to "clarify." They won't. Core updates come with no detailed explanations. The SEO community's analysis of what's moving is the best signal you have.

The Bigger Picture Here

I've been doing this long enough to remember when spinning articles was the cutting edge of content automation, and when people thought the game was over when Penguin penalized link networks. Every time Google tightens a signal, someone declares the sky is falling and then the people who were producing genuinely useful content all along end up just fine.

The July 2026 update isn't the end of AI-assisted content. It's the end of AI-assisted content as a replacement for thinking. The sites that used AI to write faster while still investing real expertise in their content are coming out of this update stronger than they went in. The sites that used AI to avoid having to know anything about their topic are the ones paying the price.

That's actually a pretty reasonable outcome. I'd argue it's the outcome Google has been working toward for years. The tools that got too good at gaming the system — article spinners, private blog networks, doorway pages — have all eventually been neutralized. AI slop content was always going to face the same fate. It just took until July 2026 for the classifiers to catch up with the behavior.

Run an Audit Before You Guess

Before you start making sweeping changes, it's worth knowing exactly what technical and content issues your site currently has. One of the mistakes I see repeatedly after core updates is people making changes based on gut instinct when the actual issues are measurable.

A proper SEO audit can surface thin pages, content duplication, pages with no internal links pointing to them, slow-loading pages that might be degrading user experience signals — all of which contribute to Google's overall quality assessment of your site. It's not just about the individual page that lost rankings. It's about the signals your whole domain is sending.

Look, this update is genuinely painful if you're caught in it. I'm not going to pretend the traffic losses aren't real or that recovery is quick. But the path forward is the same one it's always been after core updates: make your content better, demonstrate real expertise, and stop trying to rank content that doesn't deserve to rank. It takes time, and it requires actual work. That's kind of the point.