It was a Tuesday morning when the email came in.
The client—a mid-size online retailer I’d been working with for eight months—had just pulled their Google Search Console data. Their organic traffic was down 53% year-over-year.
Fifty-three percent.
Not a seasonal dip.
Not a ranking fluctuation.
A cliff.
My first instinct was to check if they’d made a technical change. Migrated the site. Changed their hosting. Implemented some new plugin that tanked performance.
Nothing.
So I asked the question I ask in every audit: “What changed on your end?”
“Nothing,” they said. “We just optimized for AI. We read everywhere that Google’s algorithms were changing. That AI Overviews were taking over search. So we rewrote our product descriptions for LLMs. We focused on getting cited by ChatGPT instead of ranking on Google. We thought we were ahead of the curve.”
They weren’t ahead of anything. They were behind everything that actually mattered.
The Thing Nobody Tells You About Optimization
Here’s what I’ve learned after auditing hundreds of sites over the past 7 years:
Most businesses that lose organic traffic don’t lose it because Google changed. They lose it because they stopped doing the things that made Google rank them in the first place.
This client had strong product pages. Clear navigation. Fast load times. A solid backlink profile. The foundation was there.
Then they got distracted.
They read about AI Overviews. About generative search. About citation optimization. And instead of adding these layers on top of their existing SEO work, they replaced the foundation with them. They rewrote descriptions that were already converting. They deprioritized the keywords that had been ranking them. They assumed that the old playbook was dead.
Based on recent client data from our 2024-2025 audit cycle: 67% of sites experiencing unexpected traffic loss had made significant changes to their content or technical setup in the six months prior, assuming those changes were necessary to compete with AI-driven search. Most of those changes weren’t necessary. They were just expensive rewrites of things that were already working.
What Actually Happened to Their Traffic
I pulled their GSC data going back fourteen months.
The decline wasn’t sudden. It was a slow fade that accelerated. Month one: 3% drop. They didn’t notice. Month two: 7% cumulative. Still writing it off as seasonal. By month five, they were down 35%. By month eight, the 53%.
What I found in the data was clearer than any report: they’d stopped ranking for the keywords that brought them customers.
Not because Google stopped showing them. Because they’d made those pages invisible.
They’d rewritten product titles for “AI readability” instead of keeping the specific, keyword-rich titles that had been ranking. They’d removed bullet points and detailed specifications because they’d read that LLMs prefer paragraph form. They’d added so much explanatory prose chasing “answer engine optimization” that the actual product information got buried.
The irony: they’d become less visible to both Google and to ChatGPT.
I looked at a secondary case study I’d been working with simultaneously—a restaurant group using their social media and review strategy to drive foot traffic and online orders. They hadn’t panicked about AI. They’d kept doing what worked (authentic reviews, consistent posting, local relevance) while adding new channels (Google Business Profile optimization, structured data for local searches). They grew 28% in six months because they didn’t confuse “new” with “necessary.”
According to Google Search Central documentation on helpful content and ranking: Google’s core ranking factors remain unchanged. The company explicitly states that ‘helpful, people-first content’ is still the primary signal, and that optimization for search engines should follow optimization for users.
Source: https://developers.google.com/search/docs/fundamentals/creating-helpful-content
The client had done the opposite. They’d optimized for a hypothetical future while losing the present.
The Problem With Chasing Signals
This is the thing that separates consultants who actually know SEO from those who are just reading the same articles as everyone else:
Real SEO work is boring. It doesn’t change every time Google releases a new feature. It compounds slowly. You optimize for humans first. You build authority. You earn trust. Then the algorithm rewards you. Then you add the next layer.
But when you’re running a business, that feels too slow. When you see headlines about “AI is the future of search,” you feel like you’re falling behind. So you start pulling resources from things that are working to chase things that might work.
The client had pulled their writer off product descriptions—their core strength—to rewrite for AI. They’d pulled their developer off site speed improvements to implement new structured data markup. They’d pulled their link-building budget to create “answer engine optimized” content.
All of it based on a false premise: that the old playbook was dead.
It’s not dead. It’s just invisible when you’re looking at the next shiny thing.
Professional opinion informed by 14 active client accounts and 40+ audits in 2024-2025: The sites that are winning right now are not the ones who’ve completely pivoted to AI optimization. They’re the ones who’ve kept their SEO foundation rock-solid while adding AI-focused strategies as a layer. Foundation first. Layers second. Always.
What Fixing It Actually Looked Like
When I presented the recovery plan, the client was quiet for a moment.
“So we just… go back to what we were doing?”
“Not back,” I said. “Forward, but on top of a foundation. We restore the product page clarity. We bring back the specific keywords that were ranking you. We keep the fast load times. We keep the backlink work. Then—and only then—we add the AI layer on top.”
It took four months. The first month was depressing: traffic stayed flat. The second month, it started moving. By month four, they were back to 94% of their original traffic. By month six, they were at 107%.
They didn’t add ChatGPT citations. They didn’t completely rewrite for LLMs. They didn’t chase Google’s AI Overview signals. They just… came home.
They rebuilt what they’d broken. And this time, they didn’t treat it like something that needed to be optimized away.
The email I got last week said their traffic was now up 31% year-over-year. Not because they’d figured out some AI secret. Because they’d remembered that SEO is still SEO. And that the foundation comes first.
The new features are real. The optimization opportunities are real. But they’re built on something older and more reliable: relevance, clarity, and earned authority.
Everything else is just noise.
Leave a Reply