Doing SEO is not a linear game where "writing good content + technical standards + clean backlinks = getting to the top". I understand the feeling of disappointment when many sites that do things right still lose their rankings after a core update. The truth is that Google's search ecosystem is influenced by layers of "invisible" signals: social consensus, brand reputation (reputation/E-E-A-T), fringe content classifier, and even seemingly small behavioral - brand recognition signals like favicon, layout, readability.
This article will "anatomize" those layers, show why Google seems "biased", and provide a practical playbook so SEO teams can adapt - instead of complain - to recover and grow sustainably. Throughout this entire roadmap, Tan Phat Digital always emphasizes the philosophy: understand the rules of the game - play by the rules - and optimize the speed of adaptation.
1) Google does not hunt for "absolute truth" - Google traces consensus
At the theoretical level, Google tries to model "common knowledge" based on sources considered trustworthy. This is like a signal voting system: when the majority of reputable sources say the same thing, the ranking model will prioritize results that match “consensus”.
In YMYL (Your Money Your Life) fields such as healthcare, finance, public policy…, the level of “consensus” needed is often higher. A contrary article, even well-written, is still difficult to compete with without strong evidence + entity reputation behind it.
Consequence: sites that pursue minority viewpoints or perspectives that go against the flow of mainstream information will have limited visibility, unless they can prove their academic capacity, original research, or have reputation strong enough to "upgrade" their content. content.
Strategic implications
Read the consensus map before publishing: which topics have consensus, which topics are controversial. With controversial topics, use "evidence language": guideline, RCT/meta-analysis (health), legal standards (finance/policy).
Write according to knowledge structure (definitions → evidence → boundaries → counterpoints), reduce the tone of absolute assertions if you do not have enough reputation.
2) Reputation/E-E-A-T: Website and author reputation are more important than you think
Google doesn't just read content; They evaluate who speaks, where, who quotes, how the community responds. Reputation comes from:
Brand signals: mentions in reputable press, academic citations, links from authoritative organizations, clear legal - corporate records.
Author signals: consistent author profile (bio, qualifications, experience, ORCID/Google Scholar with research paper, authentic LinkedIn profile real).
Historical records: domain age, regular content production process, topical authority, level of verification (citations).
Strategic implications
Build brand entities (brand entities) according to schema standards & press records.
Mapping author ↔ topic: author A for medical, B for legal... has a standard bio page, link to external profile.
Promote topical authority: publish in topic clusters, coherent internal links.
Refer to the platform guide at General SEO Handbook: https://tanphatdigital.com/vi/resources/seo-guide
3) Fringe classifier: When content is labeled “fringe”
Imagine Google has a filter to detect “deviant” content standard” compared to mainstream knowledge, related to fake news, conspiracy theories, or community risks. Once enough signs are collected, the site can get a "red score" - reduce system credibility.
The dangerous thing is spillover: a few repeat fringe articles can drag down the overall credibility score, making it difficult for "normal" articles to get to the top.
In niches such as feng shui, spirituality, folk health... the line is very thin between "culture - ritual" and curative claims lacking evidence. Just a few overzealous headlines like "wearing bracelet source verification.
Apply internal risk tags to sensitive topics; Review by a professional editor before publishing.
If you have "accidentally" published risky content, remove/edit, noindex temporarily, do a 60–90 day "clean sprint" with content with strong evidence to remove signal noise.
4) "Small" signals that amplify human behavior Using
Modern SEO is a human-in-the-loop game: users see, feel, click, read, return to the SERP. Little things like blurred/ugly favicon, hard-to-read title, description that doesn't suggest value, cluttered UI... will cause low CTR and pogo-sticking to increase – enough for the algorithm to understand “users don't like it”.
Favicon/logo: clear, consistent brand, strong enough contrast on light/dark background.
Title/Meta: short, specific benefits; Avoid keyword stuffing.
Layout: loading speed, H2/H3 hierarchy, table of contents (ToC), white space, easy-to-read font, dark mode (if available).
Mobile experience: tap targets, sticky TOC, avoid pop-ups that cover content.
If the team wants to check on-page according to new standards, see: Google Standard Onpage SEO Guide 2025: https://tanphatdigital.com/vi/blog/huong-dan-seo-onpage-chuan-google-2025
5) “Unfairness Factor”: Why the feeling of “unfairness” will always exist
Even when you do it right, There are still factors that are out of reach:
Brand effect: with the same thesis, major newspapers automatically have “plus points” in terms of credibility.
Thematic bias: YMYL requires a high threshold of evidence; "Upstream" content is difficult to display without an academic entity backing it.
Signal history: new site, thin history, low authority - subject to longer "rookie tax".
Data advantage: system sees familiar users click on brand X; that pushes up "Censorship" & multi-tool comparison: Don't just look at Google
Some keywords in the "sensitive" field (casino, 18+ content...) may not appear on Google but appear on Bing/DuckDuckGo. This is not just a story about censorship but policy differences and risk tolerance between search engines.
Practical lesson: measure index & traffic from multiple sources. Even though Google holds the majority, tracking Bing/Brave/DDG helps you identify where content lives better, thereby gaining faster ROI for each topic cluster.
7) Real-life playbook: How to “play by the rules” and overcome feelings of injustice
7.1. Clean up technical background & content structure
Sitemap & hreflang: separate sitemap into groups (posts/pages/lang), update
lastmodrealistically; declare hreflang to distinguish vi/en version.Internal link: links by topic cluster, natural anchor; prioritize navigation flow to pillar page.
For example, pillar SEO services page for conversion navigation: https://tanphatdigital.com/vi/services/seo-website
Schema: Article/NewsArticle, Organization, Person, FAQ, HowTo when appropriate.
Core Web Vitals: true speed, optimal LCP/CLS/INP; Reduce third-party bloat.
7.2. Build author & brand "entity"
Author pages: detailed bio, degrees/certificates, academic affiliation if any; Standard portrait photos, same style.
Brand entity: About + Contact + transparency policy; Consistent NAP; media mentions (PR earned), links from professional organizations.
Editorial policy: publicize the content verification process, updates – so users (and algorithms) understand you have professional discipline.
7.3. “fringe” risk management
Topic classification: label content risks; source verification checklist; Mandatory disclaimer for misleading content.
Noindex/Remove promptly for content that exceeds risk thresholds; clean sprint 60–90 days of content with strong evidence to replicate signals.
7.4. Topical authority by cluster
Create a mind-map topic → pillar → cluster; Each cluster has 1 pillar article + 6–15 satellite articles, 2-way cross-linking.
Depth over breadth: dig deep into 1 cluster before jumping to a new cluster to gather authority faster.
7.5. UI/UX is “expensive in silence”
Title/Meta with clear benefits; favicon/logo sharp; ToC stands out; mab (minimum awesome baseline) about design: spacing, line-height, contrast, mobile-first.
Trust blocks: author box, updated date, reference source, link to policy page; Verified rating/review (if any).
7.6. Multi-dimensional measurement - not dependent on 1 index
GSC: impressions/CTR position 1–10, CTR by query & device, index rate, crawl stats.
Logs & Analytics: navigation paths, time to first interaction, exit rate.
Bing Webmaster: to see the difference index/CTR.
Topic-level KPIs: number of articles/clusters reaching top 3, share of voice on SERP features (People Also Ask, Top Stories…).
7.7. Recovery process after core update
Segment diagnostics: index drops by topic, page type, language, device.
Audit “consensus”: what content is out of the mainstream? Is there enough evidence/reputable sources?
E-E-A-T hardening: add authors, sources, make the process transparent.
Content Surgery: merge thin articles, remove cannibalization, expand depth with original data (survey/case).
Internal link refactor: push internal PageRank to pillar.
Clean fringe risks: remove/adjust claims that exceed the threshold.
UI/UX polish: improve CTR (title/meta/faq), improve dwell time (table of contents, summary, TL;DR).
Re-submit: update
lastmod, ping & Submit URL Inspection for formality points.
8) Real case
Situation: A health blog in Vietnam, the author is a doctor, the content has sources but the site is new, low authority; 2 articles have somewhat "difficult" titles such as "folk trick X can replace Y".
Development: Core update scan; traffic -35%.
Handling:Attach disclaimer, edit title from “alternative” → “can support”.
Add evidence structure: systematize RCT/recommendation, add decision tree.
Create pillar “Standard treatment instructions based on guidelines” + 10 cluster articles.
Author hardening: update bio, link to academic profile, publish 1 Q&A article with institute experts.
Internal link smart, push PageRank to pillar; updated
lastmod.
Results (3–8 weeks): Impressions recovered, CTR increased thanks to new title/meta; top 10 returns for 40% of target queries.
Google does not promise absolute fairness; The system prioritizes consensus and reputation to protect users. But that doesn't mean SEO is a game of chance. With a clear playbook, execution discipline, and a pragmatic multi-tool perspective, we can completely recover and grow.
If the team needs a 90-day roadmap with specific KPIs (topic clusters, content calendar, E-E-A-T standards, sitemap/hreflang techniques, CTR/Dwell optimization) – Tan Phat Digital is ready to accompany the team.
Share
