Thin Content vs. Expert Writing: How Claude Tells the Difference

Content quality (8% weight) is harder to score than technical health but more consequential. Bad content signals a business that doesn't understand what it does — or doesn't care enough to explain it well.

What makes content "thin"

Thin content is content that exists to occupy space without providing value. It's recognizable by:

• Generic claims ("we deliver excellence in everything we do")
• No specifics (no prices, no timelines, no process descriptions)
• No evidence (no case studies, no client names, no numbers)
• Vague calls to action ("learn more" with no destination)
• AI-generated paragraphs that flow smoothly but say nothing

In 2026, with every business able to generate unlimited content with AI, thin content has become the baseline. The differentiator is expertise — writing that could only come from someone who has actually done the thing.

What we look for

Specificity. "Our plumbers have 10+ years of experience" is thin. "Our licensed master plumbers have installed over 2,400 water heaters in King County since 2014" is specific. Specificity is expertise.

Evidence. Claims should be backed by something. Case studies, testimonials with full names, before/after photos, numbered results. Anything that says "we're telling you this because we can prove it."

Depth where depth matters. Product pages for complex products should explain how the product works, what problem it solves, who it's for, and what differentiates it from alternatives. Surface-level descriptions of complex offerings signal that the author doesn't understand them.

Voice consistency. Does the blog sound like the same company as the homepage? Inconsistent voice is a common sign of either team turnover or outsourced content that hasn't been edited.

Original research or methodology. The highest-scoring content is proprietary — results from studies the company ran, methodologies they developed, data only they have. This content can't be replicated by a competitor and can't be generated by AI.

The AI content problem

Our AI is specifically calibrated to assess content that was itself AI-generated. The patterns are recognizable: perfect paragraph structure, balanced considerations, a "however" or "that said" every few paragraphs, and a complete absence of first-person expertise or genuine opinions.

We don't automatically penalize AI-assisted content — AI-drafted, human-edited content at a high standard scores well. But content that reads as entirely machine-generated without editorial judgment typically scores in the 50-65 range for content quality.

Improving your content score

1. Add one case study per quarter — real client, real problem, real result, real numbers

2. Rewrite your homepage hero to include one specific claim your competitors can't copy

3. Add a FAQ that answers the questions your sales team actually gets

4. Write one blog post per month from your own expertise, not from research

5. Put a named author on every blog post