---
title: "HCU Recovery Checklist: 27 Fixes That Worked"
slug: hcu-recovery-checklist
excerpt: "The Helpful Content Update classifier operates at the site level, which means recovery is slow and requires fixing more than a handful of pages. A 27-point checklist covering content, E-E-A-T, and pruning — and how to apply fixes in bulk."
author: Sood
author_title: Founder & CEO
author_bio: "Founder of RankWizAI. Building tools that connect GSC data to content decisions and help teams prove what fixed their rankings."
author_url: /authors/sood
author_same_as: ["https://www.linkedin.com/in/sudhirprakash/", "https://x.com/sudhirprakash"]
published_at: 2026-06-22 09:00:00
last_reviewed: 2026-06-22
meta_title: "HCU Recovery Checklist: 27 Fixes That Worked"
meta_description: "A 27-point HCU recovery checklist covering content, E-E-A-T, and pruning — and how to apply the fixes in bulk and prove they worked."
category: algorithm-recovery
reading_time_minutes: 13
featured: false
keywords:
  - hcu recovery checklist
  - helpful content update recovery 2026
  - unhelpful content fix
  - hcu recovery examples
related_posts:
  - google-algorithm-update-traffic-drop
  - may-2026-core-update-recovery
  - why-did-my-blog-traffic-drop-2026
  - content-decay-guide
---

HCU recovery is slow because Google's classifier penalizes at the site level — fixing a handful of pages on a domain with hundreds of unhelpful ones produces marginal results. This 27-point checklist covers the content, E-E-A-T, and pruning changes that have driven measurable ranking recovery across the March and May 2026 core updates, with the highest-impact fixes first.

The [Helpful Content Update](/glossary/helpful-content-update) (HCU) is no longer a standalone update. Google integrated it into the core ranking system in 2024, which means the classifier that identifies "content primarily created for search engines rather than people" now runs continuously as part of core updates — including the March and May 2026 core updates that are driving current recovery demand.

HCU recovery is distinct from recovery from other core updates in one critical way: the penalty operates at the site level. A high proportion of unhelpful pages on a domain suppresses rankings for the good pages too. Fixing ten pages on a site with two hundred unhelpful pages produces marginal results. Meaningful recovery requires addressing the site's overall quality signal — which is why this is a checklist of twenty-seven items rather than a list of five quick fixes.

Use this list in conjunction with [RankWizAI's recovery diagnostic](/recover). The diagnostic identifies which of these checklist items apply to your specific pages using the same rules the product uses in production: `ThinContentRule`, `EeatRecommendationRule`, `NoindexRule`, and the freshness pipeline. That gives you a sorted list of which items to address first, rather than working through the full checklist sequentially.

## Why HCU Recovery Is Slow (Classifier Mechanics, Core-Update Cadence)

The HCU classifier scores your site's overall content quality as a site-level signal, not page-by-page. When the classifier determines that a significant portion of your content was written for search engines rather than people, that signal suppresses the ranking potential of your better pages — even if those pages would rank well on their own merits.

This architecture explains why HCU recovery is slower than recovery from page-level issues:

1. The classifier re-evaluates your site as part of core updates, not on every crawl. Making quality improvements today may not be re-evaluated by the classifier until the next core update cycle.
2. The signal is aggregate. You need to shift the overall quality profile of your domain, not just fix the pages that lost traffic.
3. Google's quality evaluators (human raters who verify the classifier's outputs) assess sites holistically — they look at the full site experience, not individual pages.

The March and May 2026 core updates both incorporated HCU-style classifier scoring. Sites that lost rankings in both updates share a consistent pattern: the unhelpful content share on their domain is above approximately twenty percent. Recovery requires getting that share down meaningfully.

## Diagnose: Which Pages Dragged Your Whole Site Down

Before working through the checklist, identify your worst-performing pages. These are the pages dragging down the site-level quality signal, not just the pages that lost the most traffic.

Run a GSC page-level export for the ninety days before the HCU-aligned update hit you. Sort by clicks per page, descending. The bottom quartile of your pages by traffic — the pages with the fewest clicks but that still exist in your index — are your first audit targets. Many of these will be pages that have never ranked, never received meaningful organic traffic, and exist only to pad the site's topical coverage.

Compare this list against your indexed page count. If your site has 300 indexed pages and 150 of them received fewer than 10 clicks in the past 90 days, you have a thin content problem at scale.

For more detail on the [thin content](/glossary/thin-content) identification methodology, the [E-E-A-T content optimization guide](/blog/eeat-content-optimization) covers the specific GSC signals that indicate a page is underperforming relative to its inclusion in the index.

## The Checklist — Content Fixes (12 Items)

These twelve fixes address the page-level content signals that the HCU classifier and `ThinContentRule` evaluate.

**1. Add first-person experience evidence.** For every page that covers a topic where personal experience is possible, add a section documenting what you actually did, what happened, and what you measured. "When I ran this analysis on my own site, impressions dropped 40% during the May 21–22 window" is better than "sites typically see impression drops during update rollouts." Specific, checkable claims satisfy the Experience dimension of [E-E-A-T](/glossary/e-e-a-t).

**2. Replace "research + summarize" content with "do + report" content.** Content that synthesizes information from other sources without adding original analysis is the archetype of HCU-targeted writing. For each affected page, add at least one section where you report on something you did or tested, not something you researched.

**3. Add specific numbers to vague claims.** "Traffic typically drops during algorithm updates" is vague. "In the May 2026 core update, the median click drop across sites in our dataset was 23% in the first 48 hours" is specific. Specific claims are verifiable. Verifiable claims satisfy the Trustworthiness dimension of E-E-A-T.

**4. Expand thin H2 sections to 200+ words.** Sections under 150 words are often flagged as thin by content scoring tools and are visible to quality evaluators as padding. Each H2 section in an article should deliver meaningful value independently — if someone read only that section, they should gain something useful.

**5. Remove or consolidate duplicate-intent pages.** If your site has three articles all targeting "how to improve blog SEO," the classifier sees three thin pages where one comprehensive page should exist. Consolidate by 301-redirecting the two thinner versions to the most comprehensive one.

**6. Update stale statistics, examples, and references.** Content that references 2022 data in a 2026 article signals neglect to both readers and quality evaluators. Go through your top-traffic pages and update every statistic, case study reference, and tool recommendation that has aged out.

**7. Add a concrete "what to do" recommendation in each article.** HCU evaluators look for content that leaves the reader with a clear action, not just information. Add an explicit "what this means for your site" or "the specific step to take next" section to every informational article that currently ends without a recommendation.

**8. Fix misleading or over-promised titles.** If your H1 says "complete guide to X" and the article is 800 words, that gap between title claim and content depth is a quality signal. Either deliver what the title promises or retitle to match what you actually cover.

**9. Remove content that no longer reflects accurate information.** Outdated how-to guides for tools that have changed UIs, articles covering products that no longer exist, or advice that conflicts with current best practices are actively harmful to your HCU signal. Remove or update them.

**10. Add a genuine "who this is for" introduction.** The HCU classifier looks for evidence that content is written for a specific audience with real needs. A clear statement of who the article serves and what specific problem it solves signals intentionality.

**11. Ground recommendations in data from your own experience, not industry convention.** "SEO best practices say to update content every six months" is a generic claim. "We run a freshness analysis monthly on connected sites and find that posts over 180 days old with declining impressions respond well to targeted updates" is grounded. The more your recommendations derive from your own observed data, the more the Experience signal is satisfied.

**12. Add a meaningful FAQ section.** FAQ sections address the related questions your target readers have — questions that appear in People Also Ask boxes and in the "related searches" data in GSC. Answering three to five specific related questions at the end of each article increases content comprehensiveness and signals thoroughness.

## The Checklist — E-E-A-T & Author Signals (8 Items)

The `EeatRecommendationRule` in RankWizAI's engine surfaces pages with weak author entity signals. These eight items address the signals that rule evaluates.

**13. Create a proper author page with biographical detail.** The author page should include: what the author does professionally, their relevant experience, publications or projects they've worked on, and links to their external professional presence. A one-sentence bio under a post is not sufficient.

**14. Add Person schema to every author page.** Minimum required fields: `@type: Person`, `name`, `jobTitle`, `description`, `sameAs` (array of verifiable profile URLs). Add `knowsAbout` with topic strings matching your site's content themes.

**15. Connect author pages to articles via schema.** Every article should have an `author` property in its `Article` schema that references the author page via `@id`. This creates a machine-readable entity graph that connects content to a verified author.

**16. Add `sameAs` links to verifiable external profiles.** LinkedIn, GitHub for technical content, academic profiles, bylines on other publications — these are the signals that make authorship verifiable. Profile URLs that resolve to a real professional presence with relevant credentials satisfy the Expertise and Authoritativeness dimensions.

**17. Retrofit bylines onto all articles, not just new ones.** If your archive has articles with no visible author attribution, add author bylines retroactively. Anonymous authorship is a consistent pattern in sites that underperformed in the March and May 2026 updates.

**18. Add a "last reviewed" or "updated" date with a reason.** "Updated June 2026 to reflect May 2026 core update guidance" is more credible than a changed timestamp. The specific update reason signals editorial care to both human evaluators and the algorithm.

**19. Add outbound links to primary sources.** Each article should link to the primary sources that back its factual claims — Google's own documentation where relevant, studies cited in statistics, tools referenced. Outbound linking to authoritative sources is a Trustworthiness signal.

**20. Get the author's name in relevant external publications.** This is a longer-term E-E-A-T investment: bylines in industry publications, guest contributions to authoritative sites in your niche, mentions in reference articles. External authority is harder to fake than on-page markup, which is why it carries weight.

## The Checklist — Site-Level Pruning & Noindex (7 Items)

These seven items address the site-level quality signal that the HCU classifier evaluates. The `NoindexRule` in RankWizAI identifies candidates for noindex automatically.

**21. Noindex pages with zero impressions in the past 180 days.** If a page has been indexed for more than six months and has never received a single impression, it is not serving any useful indexing purpose. Noindexing it removes it from the aggregate quality calculation.

**22. Noindex paginated archive pages beyond page 2.** `/category/seo/page/3`, `/category/seo/page/4`, etc. are thin pages that rarely rank and dilute crawl budget. Noindex everything past page 2 in all pagination sequences.

**23. Noindex tag pages with fewer than five posts.** Tag pages with one or two posts are thin by definition. Either add more posts to the tag or noindex the page.

**24. Consolidate near-duplicate posts.** Posts that cover the same topic from marginally different angles — "how to find keywords for a blog post" and "blog post keyword research tips" — split ranking authority and both end up ranking poorly. 301-redirect the thinner one to the more comprehensive one.

**25. Remove auto-generated or scraped pages.** Any pages that were generated automatically from templates with little variation, or that republish content from other sources without substantial added value, are core HCU targets. Remove or substantially rewrite them.

**26. Add canonical tags to paginated series.** For multi-part articles or paginated content that should be treated as a unit, add `<link rel="canonical">` tags pointing to the primary version to prevent link equity dilution.

**27. Submit affected pages for re-indexing after fixes.** After applying substantive changes to your priority pages, request re-indexing via GSC's URL Inspection tool. This accelerates re-evaluation rather than waiting for Googlebot's regular crawl schedule.

## Applying Fixes in Bulk Across Hundreds of Posts

The twenty-seven checklist items above describe what to fix on individual pages. The practical problem is applying them at scale when your archive has hundreds of posts.

The workflow that works:

**Phase 1 (Week 1–2):** Run the site-level prune — items 21 through 26. These are structural changes to your index that improve the aggregate quality signal before you start fixing individual pages. Bulk noindex operations are fast to apply and don't require content changes.

**Phase 2 (Week 2–6):** Address content fixes in priority order (highest recoverable-traffic pages first). Use AI-assisted rewriting to address items 1–12 at scale. For each page, the AI draft should be grounded in the specific GSC query data for that page — not a generic rewrite, but a revision that incorporates the actual search queries people are using to find the page. The [guide to fixing old blog posts with AI](/blog/fix-old-blog-posts-with-ai) covers this grounding workflow.

**Phase 3 (Ongoing):** Build the author entity infrastructure once (items 13–20) and apply it across all pages. This is a one-time build that improves the E-E-A-T signal for your entire domain, not just the pages you're actively fixing.

For bulk publishing to WordPress after AI-assisted revisions, the [bulk WordPress content update workflow](/blog/bulk-seo-content-updates-wordpress) covers the review-and-publish pipeline.

## How Recovered Sites Sequenced the Work

The sites in our dataset that showed the strongest recovery signals after HCU-aligned updates followed a consistent sequence:

1. **Structural pruning first.** They used noindex to remove thin, non-ranking pages from the index before touching any content. This improved the site-level quality signal without requiring content work.

2. **E-E-A-T infrastructure second.** Author pages, Person schema, `sameAs` links — built once and deployed across the domain.

3. **Content depth third.** Focused on the twenty to thirty pages with the highest recoverable traffic, not on all affected pages simultaneously.

4. **Measurement throughout.** Per-page before/after baselines set before each change, with 7, 14, and 30-day reads.

## Tracking Across Update Cycles

Because HCU recovery is evaluated at the next core update cycle, you need a measurement framework that persists across multiple months. Set a monthly GSC checkpoint: pull position and impressions for your top 50 pages at the same time each month. Track the trend over the three to four months between the May 2026 update and the expected Q3 2026 core update.

The [RankWizAI ROI tracking feature](/docs/generating-ai-drafts) automates this — every applied fix gets a `RecommendationRoiSnapshot` with 7/14/30/90-day reads. The 90-day read is particularly relevant for HCU recovery because it's long enough to capture changes across a partial update cycle.

## FAQ

**Is HCU a separate update from core updates in 2026?**
No. Google integrated the HCU classifier into the core ranking system in 2024. What you're dealing with in 2026 is the core ranking system applying HCU-style quality scoring as part of every broad core update, including the March and May 2026 updates.

**How many unhelpful pages does it take to trigger the HCU site-level penalty?**
Google hasn't published a specific threshold. Based on patterns in sites that show recovery signals vs. sites that don't, a rough working estimate is that if more than fifteen to twenty percent of your indexed pages receive zero impressions over 90 days, the site-level quality signal is likely depressed.

**Can I use AI to help fix HCU-penalized pages?**
Yes, with the right workflow. AI-assisted content improvements that add first-person experience evidence, specific data, and concrete recommendations can address HCU signals — if the AI is grounded in your actual GSC query data and the output is reviewed before publishing. Generic AI rewrites without grounding in real data will not improve HCU signals and may make them worse.

**Will noindexing bad pages recover my good pages?**
Partially, and over time. Removing thin pages from the index improves the site-level quality signal the HCU classifier evaluates. You're unlikely to see overnight recovery, but the structural pruning is a prerequisite for content-level improvements to have their full effect.

**How do I know if my HCU recovery is working before the next core update?**
Track weekly impressions and average position for your top twenty pages. Gradual improvement in impressions — even if rankings haven't moved dramatically — indicates Google is re-evaluating your content. A useful signal is the "crawl frequency" visible in GSC's URL Inspection tool — if Google is crawling your pages more frequently after you've made quality improvements, that's a sign the improvements are being processed.