Elementor's div soup was silently killing a music marketing blog

A music marketing platform's blog traffic was stuck. AI citations were nowhere to be found.

The content was solid. The keyword strategy was on point. But the numbers weren't moving.

In fact, we were seeing a broad decline in impressions across the board. This wasn't just us, it was an industry-wide shift as Google AI Overviews rolled out. But while everyone else was panic-rewriting their content, I decided to look under the hood.

The diagnosis

I dug into the source code and found the real culprit: Elementor's deeply nested markup.

Developers call it "div soup." It's what happens when a page builder wraps every single text block, image, and heading in ten layers of useless <div> tags.

The page was bloated. It loaded slowly. And worst of all, search engine crawlers and AI models were choking on layers of empty wrapper elements before they could even find the actual content.

The fix

I'm a growth marketer, not a developer by training. But when the code is what's killing your traffic, you either fix it or keep wondering why nothing's working.

So, I went in and vibe-coded the pages for pure simplicity.

• Stripped the nested markup down to clean, semantic HTML
• Rebuilt parts of the WordPress theme from the inside out
• Custom-coded specific high-value pages directly, including our blog templates, case study pages, and a custom Spotify royalties calculator

"When your code is bloated, your content is invisible. I didn't rewrite the articles; I just removed the walls hiding them."

The results

Once the site was adapted to run on clean, fast code, the content that was already good finally started performing the way it should have from the beginning.

Rankings jumped across the board, and we saw a big increase in traffic numbers. But the most interesting shift was with AI models.

Before the rebuild, running the site through a Claude or ChatGPT audit was a nightmare. The AI would complain: "The author box is missing. There is no table of contents."

The truth? Those elements were there. The code was just so bloated with Elementor divs that the AI literally couldn't parse the page structure to find them.

Once the div soup was gone and semantic HTML was in place, everything clicked. The models could actually read the content, and the site started getting pulled into AI Overviews and citations almost immediately.