In this module
What the HCU actually penalizes (and rewards)
The Helpful Content System (now part of Google's core ranking algorithm) is widely misunderstood. It's not a "penalty for AI content" — Google has repeatedly said AI content per se is not a policy violation. What it targets is content that exists primarily for search engines rather than people.
Specific patterns associated with HCU-impacted sites:
- No original perspective, research, or experience: Rewritten summaries of what's already ranking. No personal use, no testing data, no unique insight. Pure regurgitation.
- Excessive affiliate link density: Pages where the primary purpose is clearly to get a click on an affiliate link rather than to help the reader make an informed decision.
- Unsatisfying content: Pages that answer the surface-level question but leave readers with follow-on questions unanswered. Google measures whether users come back to the search results after visiting your page (a "pogo-stick" signal).
- Mass-produced thin content: A large percentage of a site's pages being low-effort, similar to each other, clearly produced at scale with no differentiation.
What the algorithm rewards:
- First-hand experience signals — screenshots, personal results, specific version details ("I tested this on Mac in January 2026")
- Nuanced recommendations — not just "X is the best" but "X is best for this use case, Y is better if you need this feature"
- Content that acknowledges trade-offs and limitations — including limitations of the products you're promoting
- Author credentials and/or a clear site topic focus
E-E-A-T for affiliate sites
Experience
Signal that you've actually used what you're reviewing. Screenshots of your account, mention specific pricing you saw, note what version/plan you tested.
Expertise
Demonstrate you understand the domain deeply. Your content should answer questions the average person wouldn't think to ask — the ones that show real knowledge.
Authoritativeness
Other sites should mention or link to you. Hard to control directly, but earned by high-quality content that actually helps people solve problems.
Trust
Clear affiliate disclosure, an About page that explains who you are, accurate information (no inflated claims), updated content with clear dates.
The most actionable thing you can do for E-E-A-T as a new site: write from your own experience. Link to your income reports. Include screenshots. Say when you last updated each article. These are signals that are hard for pure AI-generated sites to replicate.
The AI drafting + editorial workflow
The workflow that produces rankable content isn't "ask AI to write the article and publish it." It's a two-layer process where AI handles structure and research scaffolding, and a human handles experience, judgment, and editorial quality.
Before touching AI: list the products you've tested, make notes on what you actually think of each one, pull the pricing you verified (not the pricing AI might have stale data for), and write down the questions you had when evaluating the product. This is raw material AI cannot have.
Feed AI your keyword target and content type (roundup, comparison, review). Let it propose an outline. Revise the outline to match what you know from step 1. The outline is yours; AI just accelerated it.
Background information, feature definitions, use-case descriptions — sections that don't require your personal experience. AI writes a solid draft. You edit for accuracy and completeness.
Pros/cons from actual use, your specific workflow recommendation, the nuanced "but if you're doing X, then Y is actually better" insights. These sections must be you. This is where E-E-A-T is earned.
Read the full draft aloud. Cut sentences that say nothing. Replace vague claims ("great tool") with specific observations ("the bulk export function took under 30 seconds for a 500-row dataset in my test"). Add screenshots where possible.
Add the relevant structured data schema (see below), add internal links to and from related pages, verify all links work, publish.
Writing each content type well
Best-of Roundup pages
- Lead with a clear recommendation: "Best overall: [X], Best for beginners: [Y], Best free option: [Z]" — this answers the question before the reader has to scroll.
- Include a comparison table near the top with real pricing (current, not AI's cached data).
- Each product section should include: what it's best for, key features, pricing, one genuine limitation. Not just positives — limitations are the sign of an honest review.
- Update the page and change the published date when prices or product features change.
Comparison pages (A vs B)
- The reader wants a verdict. Give one clearly, early. "If you need [X feature], go with [A]. If [Y] matters more, [B] wins."
- A side-by-side feature table is expected — include it.
- Cover the "switching cost" question: which is easier to get started with, which has better import/export.
- Address price at their respective entry points, not just the highest tier.
Individual product reviews
- State the date you tested and the plan/tier you used — Google wants to see recency signals.
- Use the actual product interface for screenshots — these are uniquely yours and signal experience.
- Score against criteria the reader cares about, not generic "ease of use / features / price".
- Be honest about who should NOT use this product. Counterintuitively, this increases trust and conversion.
Schema markup for affiliate pages
Structured data doesn't directly cause rankings to improve, but it helps Google understand content type and can trigger rich result features. For affiliate pages, use these schemas:
Article / BlogPosting
Use on informational and how-to content. Include datePublished, dateModified, and author with @type: Person. The dateModified field is especially important — Google uses this to assess freshness.
Review
Use on individual product review pages. Include reviewRating with ratingValue and bestRating, author, and itemReviewed pointing to the product with its name and description. Note: Google no longer shows star ratings in search results for self-serving reviews (you reviewing your own affiliates), but the schema still helps with content understanding.
ItemList
Use on roundup / best-of pages. Wrap your list of products in an ItemList schema with each product as a ListItem. This can help trigger a list-style rich result in some queries.
Minimal example — review page
On-page SEO checklist
Run each new page through this checklist before publishing:
<head>. Points to the page's full URL (including trailing slash if site uses them).<head>. Validated with Google's Rich Results Test.datePublished and dateModified in schema. Published date visible to readers.Module 4 action steps
- Test the products in your first cluster before writing — take notes and screenshots
- Use the AI drafting workflow for your first 3 articles
- Add the relevant schema markup to each page type
- Run each page through the on-page SEO checklist before publishing