https://lffafeqydexbjcnuctew.supabase.co/storage/v1/object/public/notes-audio/notes/3778c3685191807e93dbc131fc2addc8.mp3

🎧 Listen to this note (7 min): Open player

<aside> 🚨

Want more notes like this? Subscribe to weekly strategy emails: seonotebook.com

</aside>

Howdy Notebooker!

Back in December, Google quietly updated one of the most useful pages they've ever published. Almost nobody noticed.

It's their "Creating helpful, reliable, people-first content" guide. The last refresh landed December 10, 2025.

Here's why I care about it more than most SEO content out there. It's the closest thing we have to a published checklist of what Google's ranking systems actually reward. Not a leak. Not a theory. Google's own words.

There are 32 self-assessment questions on that page. Things like whether your content gives original information, whether it's written by someone who clearly knows the topic, and whether you're refreshing dates to fake freshness.

The problem? Almost everyone reads it once, nods along, and never opens it again.

I got tired of grading pages against it by hand. It's slow, it's inconsistent, and two people on my team would score the same page differently.

So I built a bot that does it for me.

ChatGPT - Helpful Content Grader

What the grader actually does

I built it twice. Once as a Custom GPT and once as a Claude Project, so you can use whichever one you live in.

You paste in a page (the raw text, or a URL in the GPT version), and it scores that page against all 32 of Google's questions.

Every question gets a PASS, FAIL, PARTIAL, or N/A, with a one-line reason tied to the actual content. No vague "work on your E-E-A-T" hand-waving. It points at the exact thing that's missing.

Then it rolls everything up into a single letter grade, A through F, plus a subtotal for each section.

The four buckets it grades

Google's 32 questions fall into four groups. Here's how the grader splits them:

  1. Content & Quality (12 questions). Original info, depth, comprehensiveness, clean writing.