Stop trying to rank for keywords 🔗 **Source:** [Stop trying to rank for keywords](https://www.jonoalderson.com/conjecture/stop-trying-to-rank-for-keywords/) > [!note]+ Document Notes > tags: [SEO, search, content-strategy] > author: Jono Alderson > source: jonoalderson.com > reference: [https://www.jonoalderson.com/conjecture/stop-trying-to-rank-for-keywords/] > date created: 2026-02-23 12:00 > > # 🏷️ Title: Stop trying to rank for keywords > <!-- deep-read:sections start --> ## 🧭 Table of Contents - [📝 Summary](#-summary) - [📖 Main Content](#-main-content) - [🗒️ Notes](#️-notes) - [📚 References](#-references) - [🗣️ Paraphrase](#️-paraphrase) - [📰 Newsletter](#-newsletter) - [✅ Action Items](#-action-items) - [🧪 Hypotheses](#-hypotheses) - [📊 Methodology](#-methodology) - [❓ Questions](#-questions) - [🔮 Future Research](#-future-research) - [💡 Implications](#-implications) - [🧠 Learning Styles Assessment](#-learning-styles-assessment) - [📊 Accessibility Assessment](#-accessibility-assessment) - [🌍 DEIB Assessment](#-deib-assessment) - [✅ Logical Validity](#-logical-validity) - [🛡️ DARVO Analysis](#-darvo-analysis) - [🎯 Conclusion](#-conclusion) ### 🔗 References - Include links cited in or relevant to the article. ## 📝 Summary - The article argues that starting SEO with “we want to rank for X” is backwards. - What matters is becoming a distinctive, trusted entity that is the obvious answer for a specific audience in a specific context. - Stakeholders: marketers, SEOs, product teams, founders, content teams. - Risks: continuing to churn interchangeable “keyword pages” that get collapsed into AI/search summaries, and failing to differentiate. - Bottom line: treat keywords as diagnostic input; invest in unique capabilities, evidence, and legibility so search visibility emerges as a byproduct. ## 📖 Main Content - Search engines and AI don’t reward sites because they target a keyword; they infer entities (people, brands, products) and decide which is most credible and useful. “Pages are inputs. Rankings are outputs.” - The typical SEO playbook—find a keyword, create a landing page, write 1,000 generic words, add headings and links—produces content that is easy for systems to collapse into a single canonical answer. You become undifferentiated training data for someone else’s synthesis. - The core strategic shift is from “how do we rank for X?” to “what are we uniquely positioned to produce, for which audience, in which context, that genuinely reduces their friction?” Jobs To Be Done is used as a framing: people “hire” solutions to make progress, not to consume keywords. - Differentiation is central. The article lists structurally different artefacts competitors can’t easily copy: original datasets, tools, deep real-world case studies with numbers and trade‑offs, opinionated frameworks, and detailed public product documentation. - Being the best product is not enough; systems rely on signals such as reviews, citations, discourse depth, and author credibility. SEO is about being *understood* as the best via these signals, at the entity level. - Authorship and expertise must be legible: not “Marketing Team” bylines on generic pieces, but identifiable experts with a visible footprint across talks, interviews, and citations. This is framed as building a “real, legible entity on the web.” - The article proposes a sequence: clarify capabilities; understand audience progress; design unique utility; gather evidence; then amplify so the work is referenced and embedded in the wider ecosystem. Only *after* this should teams look at search behaviour and keywords, using them as a language map, not a content wishlist. - For objections (early‑stage brands, commodity e‑commerce), the answer is not “give up on differentiation,” but document real trade‑offs, data, compatibility questions, sourcing, usage, and community knowledge. “Specificity beats generic authority.” - AI is described as removing “slack in the system.” When systems synthesise and collapse similar documents, slightly‑better‑than‑average content decays, while distinctive, experience‑backed artefacts compound. - The closing move: stop trying to rank for keywords; ask instead “What can we build or publish that would make us the obvious answer for a specific audience, in a specific context?” Visibility then follows from being useful, legible, referenced, and remembered. ### 🗒️ Notes - “You do not choose what you rank for. You are inferred.” — rankings are downstream of how systems model your entity. - Standard “rank for X” workflow: pick high‑volume keyword → matching landing page → 1,000 words of generic advice → headings + internal links → wait. Works less and less as search and AI collapse redundancy. - “Most keyword landing pages are not resources. They are bait. And increasingly, bait is ignored.” - Keyword‑first thinking: - Centres channel, not audience. - Centres metric, not outcome. - Treats demand as something to *capture* instead of to understand or shape. - Better starting questions: - What are we uniquely positioned to produce? - Which audiences are we trying to help make progress? Where do they struggle? - What can we publish/build/expose that genuinely reduces friction? - Differentiation examples: - Original datasets from your operations. - Tools that solve recurring problems. - Detailed project breakdowns with numbers and trade‑offs. - Opinionated frameworks that reflect how you actually work. - Public documentation of edge cases. - If you can’t answer “what do we uniquely know or see?”, your SEO problem is really a positioning and capability problem. “You cannot out‑optimise commoditisation.” - Systems infer “best” from: - Reviews, ratings. - Mentions, citations, and discourse depth. - Author credibility and consistency across contexts. - “Being the best and being understood as the best are different problems. SEO sits squarely in the second.” - Anonymous or generic corporate bylines are weak in a synthesis‑driven world; expertise must be attributable and connected. - Clear semantic structure and relationships (people ↔ orgs ↔ content) help machines model you more accurately, echoing prior arguments about semantic HTML. - Recommended sequence: 1. Capabilities — non‑trivial knowledge, builds, or insights. 2. Audience progress — who’s trying to do what, where, and how. 3. Unique utility — what materially helps them in a way others don’t. 4. Evidence — how you *demonstrate* experience. 5. Amplification — how the work is seen, referenced, embedded. - Then use keywords as: - A map of language. - A way to see how audiences describe problems. - A lens on geographic/cultural nuance, fragmentation, and confusion. - Early‑stage brands: advantage is proximity to the problem; document trade‑offs, failures, and learning. - Commodity e‑commerce: differentiate via depth (compatibility guides, transparent testing, comparisons, real installation videos, community Q&A). - “In an AI‑summarised world, indistinguishable suppliers are abstracted away.” - Distinctive, experience‑backed artefacts compound; generic landing pages decay. - “If your first question is ‘what do we want to rank for?’, you are optimising the mirror, not the muscle.” - Final principle: “Keywords still matter… But they are a diagnostic tool, not a strategy. Stop trying to rank for keywords. Start trying to become the answer.” ## 📚 References - Jono Alderson, “Stop trying to rank for keywords” — https://www.jonoalderson.com/conjecture/stop-trying-to-rank-for-keywords/ - Jobs To Be Done overview (for context) — https://jobstobedone.org/ - Google Search Quality Rater Guidelines (for E‑E‑A‑T context) — https://static.googleusercontent.com/media/guidelines.raterhub.com/en//searchqualityevaluatorguidelines.pdf ## 🗣️ Paraphrase - The author criticises keyword‑first SEO strategies, arguing that rankings are the outcome of how search and AI systems infer the quality and distinctiveness of entities, not the reward for ticking optimisation checkboxes. - He proposes a focus on unique capabilities, audience progress, and hard‑to‑replicate artefacts that clearly demonstrate expertise, so that search visibility becomes a natural byproduct rather than the central goal. ## 📰 Newsletter In this week’s WordPress‑focused SEO corner, Jono Alderson challenges the classic “we want to rank for X” brief. For WordPress builders, this means moving beyond spinning up yet another “ultimate guide” post and instead using WordPress as a platform for distinctive artefacts: custom data dashboards, interactive tools, deeply‑documented case studies, and transparent product docs. Pair this with solid semantic markup, author profiles, and clear entity relationships, and your site becomes the obvious answer—not just another AI‑summarised snippet. If you’re planning your next content sprint, start with “what can we uniquely show?” instead of “which keyword should we chase?” ## ✅ Action Items #### DevRel - Document and publish opinionated implementation guides showing real trade‑offs, numbers, and constraints. - Run workshops on entity‑centric SEO and “becoming the answer” for your developer audience. - Create example apps/tools that surface original operational data and embed them in docs. - Produce content campaigns around real user progress stories, not keyword themes. - Facilitate community discussions that generate citations, mentions, and discourse depth. #### OSPO - Encourage open sourcing internal tools/datasets that can safely be shared to create unique public artefacts. - Establish policies for attribution and authorship so contributors are legible entities. - Align vulnerability, SBOM, and provenance work with public documentation that search systems can model. - Track how upstream contributions and citations increase organisational entity prominence. #### Community - Help maintainers and contributors document real‑world case studies and implementation notes. - Run “document your edge cases” sprints producing highly specific, reference‑worthy content. - Encourage Q&A formats and community knowledge bases that surface lived experience. - Create transparent funding or sponsorship reports as differentiated, citable artefacts. #### WordPress - **Core:** Publish deep, number‑backed explanations of major decisions (performance, UX, security) as citable resources. - **Core Performance:** Maintain public, repeatable benchmarks and dashboards with methodology clearly explained. - **Core AI:** Document AI experiments, evaluation sets, and guardrails with explicit trade‑offs. - **Core Program Team:** Make cross‑team decision logs and retros public where appropriate as differentiation signals. - **Mobile:** Share detailed implementation chronicles (offline UX, telemetry decisions) as posts or dev notes. - **Playground:** Publish reproducible labs and debugging walkthroughs for complex scenarios. - **Community:** Turn meetup and WordCamp talks into well‑attributed, well‑structured written artefacts. - **Training:** Focus curricula on real workflows and “jobs to be done” rather than keyword‑themed lessons. - **Marketing:** Shift from “SEO content calendars” to campaigns built around unique utilities and tools. - **Support:** Turn recurring complex tickets into detailed guides authored by named experts. - **Design:** Document design systems, tokens, and rationale publicly with examples and metrics. - **Documentation:** Emphasise authored, experience‑based guides over anonymous generic summaries. - **Photos:** Publish clear, transparent policies and curation rationales as reference‑worthy docs. - **TV:** Ensure videos are accompanied by rich, structured transcripts and attributed show notes. - **Polyglots:** Capture translation edge cases and cultural nuances as searchable guidance. - **Accessibility:** Publish thorough, example‑rich A11y audits and implementation checklists. - **Meta:** Make .org roadmap, incident reports, and architecture docs first‑class, linkable resources. - **CLI:** Provide realistic task‑driven examples and case studies of CLI usage in production. - **Hosting:** Share benchmarking methodologies, reliability SLOs, and incident postmortems. - **Tide:** Transparently document scoring models, false‑positive handling, and iteration history. - **Openverse:** Publish ingestion and moderation process docs with examples and metrics. - **Themes:** Offer detailed theme case studies (performance, UX, maintainability) from real sites. - **Plugins:** Document plugin security practices, update policies, and long‑term support histories. #### The WPOCC - Design funding programs that reward creation of distinctive, open, experience‑backed artefacts. - Support maintainers in publishing governance and release‑process docs with explicit trade‑offs. - Treat funded outputs as long‑lived, citable references, not just announcements. ### FAIR - Encourage local, verifiable metadata for authors and artefacts to strengthen entity modelling. - Make governance artefacts (charters, policies, grant ledgers) public, structured, and linkable. - Provide reusable templates so other projects can more easily create their own legible artefacts. ## 🧪 Hypotheses - H1: Sites that invest in unique, experience‑backed artefacts (datasets, tools, deep case studies) will see stronger long‑term organic visibility than sites focused on keyword‑targeted summaries. - H2: Clear, attributable author expertise (visible footprint across the web) will correlate with higher rankings and inclusion in AI‑generated answers. - H3: Commodity e‑commerce sites that provide detailed compatibility, testing, and usage documentation will be less likely to be abstracted away by AI summarisation than those with templated pages. - H4: Early‑stage brands that systematically document real‑time learning and trade‑offs will gain entity recognition faster than those publishing generic thought leadership. ### 📊 Methodology - Longitudinal study comparing cohorts of sites: one group produces unique artefacts; another publishes generic keyword‑targeted articles. Track rankings, featured snippets, and presence in AI overviews. - Correlational analysis of author entities: measure footprint (talks, citations, social graph) vs. performance of their content in search and AI answers. - Controlled comparison within an e‑commerce vertical: redesign a subset of product lines with deep compatibility/testing content; compare visibility and conversion to control group. - For startups, track time‑stamped “build in public” documentation vs. generic blog output and subsequent growth in branded search, mentions, and high‑authority citations. ### ❓ Questions - How can smaller teams practically scope “structurally different” artefacts without overextending? - What’s the minimum viable proof of expertise that meaningfully shifts entity‑level inference? - How will AI answer synthesis evolve in weighting original data/tools vs. narrative content? - Where is the line between “being legible” and over‑exposing internal operations or privacy‑sensitive information? ### 🔮 Future Research - Quantitative models linking specific types of artefacts (tools vs. datasets vs. deep dives) to different search and AI outcomes. - Frameworks for measuring “sameness” and “distinctiveness” in large content corpora. - Best‑practice patterns for author identity and entity modelling across platforms. - Empirical studies on how users perceive and trust entity‑centred vs. keyword‑centred content. ### 💡 Implications - Practice: SEO roadmaps should begin with capabilities, audience progress, and unique utilities, with keyword research as a later supporting activity. - Policy: Organisations may need guidelines around transparent documentation, data sharing, and authorship. - Security: Public documentation of operations must be balanced against security posture; redaction patterns are needed. - Education: SEO and content curricula should emphasise systems thinking and entity modelling over mechanical keyword tactics. - Funding: Grants and budgets should prioritise durable, reference‑worthy artefacts rather than volume‑based content production. ### 🧠 Learning Styles Assessment - Estimated reading level: roughly Grade 10–11 (Flesch‑Kincaid style estimate based on sentence length and vocabulary). #### 👀 Visual learners - Present: clear structure with headings and bullet lists. - Improve: add diagrams illustrating entity models, “generic vs. differentiated” flows, and before/after examples. #### 🔊 Aural learners - Present: none explicitly. - Improve: short audio summary or podcast‑style discussion of the main argument. #### 🗣️ Verbal learners - Present: strong conceptual framing, clear contrasts, and repeated key phrases. - Improve: brief glossary of terms like “entities,” “synthesis,” and “Jobs To Be Done.” #### 🧩 Kinesthetic learners - Present: implied actions (ask these questions, build these artefacts). - Improve: step‑by‑step exercises (e.g., audit a content portfolio for sameness, design one structural artefact). #### 🧠 Logical learners - Present: clear causal chains (inputs → modelling → outputs), explicit sequences and lists. - Improve: simple diagrams or tables mapping tactics to outcomes and risks. #### 🧑‍🤝‍🧑 Social learners - Present: framing that encourages collaboration between marketing, product, and subject‑matter experts. - Improve: discussion prompts and team workshop outlines to redesign SEO strategy. #### 🧘‍♀️ Solitary learners - Present: plenty of conceptual material to reflect on individually. - Improve: reflection questions at the end (“what can I uniquely produce?”). #### 🌿 Naturalistic learners - Present: minimal direct environmental framing. - Improve: analogies to ecosystems (e.g., monoculture vs. biodiversity in content) to emphasise differentiation. ### 📊 Accessibility Assessment #### Screen reader semantics - Present: semantic headings and list structure (based on article excerpt style). - Improve: ensure landmark roles, language attributes, and descriptive link text are used consistently. #### Keyboard navigation - Present: likely standard article behaviour. - Improve: confirm visible focus indicators, skip links, and logical tab order. #### Contrast and readability - Present: typical blog layout; likely adequate contrast. - Improve: verify WCAG 2.2 AA contrast ratios and adjust typography (line height, spacing) if needed. #### Text resizing and reflow - Present: responsive design is probable. - Improve: test at 200% zoom and with prefers‑reduced‑motion settings to prevent layout issues. #### Media and captions - Present: article is text‑first. - Improve: if any media are added (charts, videos), include alt text, captions, and downloadable transcripts. #### Forms and interactions - Present: minimal or none on the article page. - Improve: where forms exist (newsletter, comments), ensure labelled controls, clear errors, and adequate target sizes. #### Standards reference - Map observed or potential gaps to WCAG 2.2 AA criteria (1.4.x contrast, 2.1.x keyboard, 1.3.x structure, etc.). ### 🌍 DEIB Assessment #### Cultural breadth - Present: content is largely Western digital‑marketing centric. - Improve: more global examples and references to varied markets and languages. #### Inclusive language - Present: generally neutral, professional language. - Improve: define specialised jargon and avoid insider/consulting idioms where possible. #### Neurodiversity support - Present: clear structure, headings, and chunked content. - Improve: even more explicit signposting, consistent patterns, and optional summaries or checklists. #### Representation - Present: conceptual; no people depicted. - Improve: if adding imagery or examples, reflect varied identities, regions, and roles. #### Psychological safety - Present: direct but not hostile; challenges ideas, not individuals. - Improve: explicit invitations to adapt ideas to different scales and contexts, avoiding “one right way” tone. ### ✅ Logical Validity #### Strong reasoning - Uses causal logic: entity signals → system inference → rankings; consistent with known search behaviour. - Employs reductio of keyword‑first tactics by showing their convergence to sameness and subsequent collapse by AI. - Provides concrete alternatives (datasets, tools, case studies) that logically follow from the goal of distinctiveness. #### Fallacies - Mild risk of hasty generalisation: implies most keyword landing pages are “bait” and ignored; real‑world exceptions exist. Could be softened by acknowledging niches where simple optimisation still works. - Potential survivorship bias: focus on systems that reward differentiation may underplay cases where authority alone sustains visibility. #### Tone of voice - Confident, provocative, and slightly contrarian (“Stop trying to rank for keywords”). - Uses sharp contrasts (“bait vs. resources,” “mirror vs. muscle”) and short declarative sentences for emphasis. ### 🛡️ DARVO Analysis - Deny: No denial or minimisation of harm; instead, the author acknowledges systemic shifts that make old tactics ineffective. - Attack: No personal attacks; criticism is aimed at practices and mindsets, not individuals. - Reverse victim/offender: No attempt to recast those challenged as persecutors; the tone is advisory, not defensive. ## 🎯 Conclusion Search and AI systems reward entities that are clearly understood, differentiated, and genuinely useful—not those that merely target keywords. Reorient SEO around capabilities, audience progress, unique utilities, visible expertise, and amplification. Use keywords to understand language and gaps, then build artefacts that make your organisation the obvious answer; let rankings follow from that reality. ## 🧱 Refactor Appendix (append after 🎯 Conclusion) ### 📚 Literature Split Pad - Purpose: Stage Konik CEE units that become per-source Literature notes. - Split level: **H3**. - Include verbatim quotes under 25 words and a [view highlight] link. ### [H1] 5–10-word present-tense claim > [!quote] [Stop trying to rank for keywords](https://www.jonoalderson.com/conjecture/stop-trying-to-rank-for-keywords/) by [[Jono Alderson]] via [[jonoalderson.com]]. Published 2026-02-21. > “Keywords still matter… But they are a diagnostic tool, not a strategy.” Two lines: context and use. Cite data if relevant. [view highlight](https://www.jonoalderson.com/conjecture/stop-trying-to-rank-for-keywords/#H1) #xref #articleseed ID:rw_[MISSING] ### [H2] 5–10-word present-tense claim > [!quote] [Stop trying to rank for keywords](https://www.jonoalderson.com/conjecture/stop-trying-to-rank-for-keywords/) > “Search visibility is a byproduct of being useful, legible, referenced, and remembered.” Two lines: context and use. [view highlight](https://www.jonoalderson.com/conjecture/stop-trying-to-rank-for-keywords/#H2) #xref ID:rw_[MISSING] --- ### 🔖 Atomic Split Pad (Zettel Candidates) - Purpose: Promote reusable claims into Atomic notes. - Split level: **H4**. - Each item: one idea, present tense, specific, testable. #### [Z1] Entity modelling drives search visibility more than keyword targeting - Scope/condition in one line. - Pointer to page/time or dataset. - Links: [[SEO]] [[Entity-Search]] #### [Z2] Unique artefacts compound while generic content decays - Scope/condition in one line. - Pointer to page/time or dataset. - Links: [[Content-Strategy]] [[Differentiation]] --- ### ✏️ Fleeting Queue (Extract by Selection) - Purpose: Capture quick actions or sparks for Fleeting notes. - Extraction: **Select one line** → “Refactor: Extract selection to new note (content only).” - Keep each line short and self-contained. - [F1] Audit existing content for sameness vs. structural differentiation — context/https://www.jonoalderson.com/conjecture/stop-trying-to-rank-for-keywords/ #fleeting #inbox - [F2] Design one original dataset or tool that competitors cannot easily copy — context/https://www.jonoalderson.com/conjecture/stop-trying-to-rank-for-keywords/ #fleeting #inbox - [F3] Map authorship and expertise signals to entity visibility goals — context/https://www.jonoalderson.com/conjecture/stop-trying-to-rank-for-keywords/ #fleeting #inbox --- <!-- deep-read:sections end -->