# External content SEO / AEO / GEO research log **Purpose:** Dated, source-linked notes from web research to inform Ordio blog workflow updates. Do not duplicate full internal guides—link to [AEO_GEO_BEST_PRACTICES_2026.md](AEO_GEO_BEST_PRACTICES_2026.md) and [CONTENT_CREATION_WORKFLOW_2026.md](CONTENT_CREATION_WORKFLOW_2026.md). --- ## 2026-04-04 — Lexikon „DATEV“ (Ökosystem, Lohn, Schnittstellen) **Scope:** HR/Payroll-Einordnung für Arbeitgeber — **kein** Arbeitnehmer-Online-Login-Guide, keine Preisliste, keine Produkt-Roadmap-Gerüchte. Abgrenzung: DATEV = Anbieter/Software-Ökosystem; Lohnabrechnung = Prozess/Dokument. **Sources (high level):** - [DATEV – Unternehmensportal](https://www.datev.de/web/de) — Cloud-Anwendungen, MyDATEV, Lohn- und Personal-Lösungen (Überblicksseiten). - [DATEV – Lohn- und Gehaltsabrechnung](https://www.datev.de/web/de/unternehmen/loesungen/lohn-und-personal/lohn-und-gehaltsabrechnung) — Rollen Selbstabrechnung vs. Steuerberater, digitale Lohnunterlagen. - [Wikipedia – DATEV](https://de.wikipedia.org/wiki/DATEV) — DATEV eG, Genossenschaft, Geschäftsfelder (neutraler Rahmen; keine wörtliche Übernahme). - SISTRIX SERP/PAA: hoher Anteil Arbeitnehmer-/Preis-/App-Vergleichs-Noise → `paa-questions-manual.json` mit 15 HR-Fokus-Fragen. **Decisions for Ordio content:** 1. **Angle:** Arbeitgeber:innen — Datenfluss Zeiterfassung → Lohn, Schnittstellen, Rolle Steuerberater, GoBD-**Kontext** mit Link zum Lexikon „Revisionssicherheit“. 2. **Vergleichstabelle:** DATEV vs. Lohnabrechnung vs. Steuerberatung (Beruf/Mandat). 3. **Ordio:** kontextuell `/payroll` + `/arbeitszeiterfassung`; keine Markenverwechslung („Ordio ist DATEV“). 4. **Featured Image:** keine DATEV-Logo-Grafik (Marken/TM-sicher). **Post folder:** `docs/content/blog/posts/lexikon/datev/` (`SERP_ANALYSIS.md`, `CONTENT_OUTLINE.md`). --- ## 2026-04-04 — Lexikon „Gefahrenzulage“ (Tarif/BV/AV, Abgrenzung Zuschläge) **Scope:** SERP/PAA und HR-Glossar-Konkurrenz (u. a. Staffomatic); **kein** Anspruch auf eine bundeseinheitliche Höhe; Steuer/SV nur high-level (§ 3 Nr. 4 EStG vs. typisches Arbeitsentgelt). Arbeitssicherheit (Gefährdungsbeurteilung) bewusst von Lohnanspruch getrennt. **Post folder:** `docs/content/blog/posts/lexikon/gefahrenzulage/` (Details: `KEYWORD_DECISION.md`, `SERP_ANALYSIS.md`). --- ## 2026-03-24 — Lexikon „Grundfreibetrag“ (EStG / BMF / Lohnsteuer) **Queries / Tools:** Serper (gl=de, hl=de): `Grundfreibetrag`, `Grundfreibetrag 2026`, `Grundfreibetrag Lohnsteuer`, `Grundfreibetrag EStG`, `Tabellengrundfreibetrag Lohnsteuer`, `Grundfreibetrag Verheiratet Splitting` (v. a. Video/Serp-Noise); Fetch `gesetze-im-internet.de/estg/__32a.html`, `bundesfinanzministerium.de/.../das-aendert-sich-2026.html`; Fetch Papershift + Nettolohn Lexikon-URLs. **Primary legal source:** - [§ 32a EStG – Einkommensteuertarif](https://www.gesetze-im-internet.de/estg/__32a.html) — ab VZ 2026 erste Stufe „bis 12 348 Euro (Grundfreibetrag): 0“. **Official / high-trust German:** - [BMF: Die wichtigsten steuerlichen Änderungen 2026](https://www.bundesfinanzministerium.de/Content/DE/Standardartikel/Themen/Steuern/das-aendert-sich-2026.html) — Grundfreibetrag 12.348 € (+252 €), Kinderfreibetrag 9.756 €, Beispiele Lohnsteuer. **Competitors (Lexikon):** papershift.com/lexikon/grundfreibetrag; nettolohn.de/lexikon/grundfreibetrag.html; plus SERP-Top steuertipps, Wikipedia, VLH, IHK, lexware, smartsteuer, gesetze-im-internet. **Decisions for Ordio content:** 1. **Angle:** HR/Payroll – Grundfreibetrag im **Tarif** und **Lohnsteuerkontext** erklären; nicht als eigene Zeile auf dem Lohnzettel suggerieren. 2. **Zahlen:** 12.348 € nur mit **2026** + Verweis § 32a/BMF/Projekt-Konstanten. 3. **SERP noise:** `universal-search-box.com` ohne Text – für Tiefe ignorieren; in `SERP_ANALYSIS.md` vermerkt. **Post folder:** `docs/content/blog/posts/lexikon/grundfreibetrag/` --- ## 2026-03-24 — Lexikon „Grundrentenzuschlag“ (SGB VI / DRV / HR) **Queries / Tools:** Serper `Grundrentenzuschlag`, `Grundrente Grundrentenzuschlag Unterschied` (gl=de, hl=de); Fetch `gesetze-im-internet.de/sgb_6/__76g.html`. **Primary legal source:** - [§ 76g SGB VI – Zuschlag an Entgeltpunkten für langjährige Versicherung](https://www.gesetze-im-internet.de/sgb_6/__76g.html) — Definition Grundrentenzeiten, Grundrentenbewertungszeiten, Höchstwerte, Faktor 0,875, max. 420 Monate. **Official / high-trust German:** - [DRV FAQ Grundrente / Grundrentenzuschlag](https://www.deutsche-rentenversicherung.de/SharedDocs/FAQ/grundrente/grundrente_faq_liste.html) - [BMAS Fragen und Antworten Grundrente](https://www.bmas.de/DE/Soziales/Rente-und-Altersvorsorge/Leistungen-Gesetzliche-Rentenversicherung/Grundrente/Fragen-und-Antworten-Grundrente/fragen-und-antworten-grundrente-art.html) - [DRV Presse 2025 – Überprüfung Grundrentenzuschlag](https://www.deutsche-rentenversicherung.de/DRV/DE/Ueber-uns-und-Presse/Presse/Meldungen/2025/250115_ueberpruefung-grundrentenzuschlag.html) — Einkommensanrechnung / Aktualität **Competitors (Lexikon / Ratgeber):** papershift.com/lexikon/grundrentenzuschlag; dgb.de/service/ratgeber/grundrente/; rentenblicker.de/wiki/g/grundrentenzuschlag/; wsi.de; bpb.de (Kontext Reform). **Decisions for Ordio content:** 1. **Angle:** HR/Payroll — Arbeitgeber **meldet und zahlt Beiträge**, **berechnet** den Grundrentenzuschlag **nicht**; kein Posten auf **Lohnabrechnung**. 2. **Zahlen (Freibeträge, Durchschnittsrente):** nur mit **Jahr** und Verweis auf **DRV/BMAS**, keine festen Euro als „immer gültig“ ohne Quelle. 3. **Pipeline:** SISTRIX PAA „no result“ → `paa-questions-manual.json`; SISTRIX-Competitor-Noise (`universal-search-box`, leere ruhr24-Analyse) aus `competitor-analysis.json` entfernt vor Depth-Analyse. **Post folder:** `docs/content/blog/posts/lexikon/grundrentenzuschlag/` --- ## 2026-03-23 — Answer-first structure, AEO, skyscraper alignment **Sources (external):** - [Answer Engine Optimization (AEO) – AEO Mastery](https://aeomastery.org/blog/answer-engine-optimization/) — AEO definition; schema + site structure; modular extractable answers. - [What is AEO? – AI Search Rankings](https://www.aisearchrankings.com/guides/what-is-aeo.php) — AI answers as “position zero”; citation-oriented optimization. - [AEO explained – LLM VLab](https://www.llmvlab.com/guides/answer-engine-optimization) — conversational search, being quoted as source. - [AEO best practices – ClickRank](https://www.clickrank.ai/aeo-best-practices/) — bullets, structure, factual tone. - [Featured snippets + AEO – BOHO SEO](https://bohoseo.com/blog/how-to-optimize-for-featured-snippets-and-write-for-aeo/) — inverted pyramid; concise answer under headings that mirror queries. **Takeaways aligned with our stack:** 1. **Inverted pyramid / answer-first:** Place a **short direct answer** immediately under **question-shaped H2s** (we document 120–150 char capsules in [CONTENT_DEPTH_GUIDELINES.md](CONTENT_DEPTH_GUIDELINES.md)); avoid answer-leads under declarative H2s. 2. **Extractability:** Lists, tables, and clear H2/H3 hierarchy help both SERP snippets and LLM citation—matches our `validate-section-depth`, table rules, FAQ schema. 3. **Schema + discovery:** FAQPage / Article JSON-LD and `llms.txt` are already part of the product; research reinforces keeping FAQs aligned with body (validators) and maintaining LLM file hygiene per [docs/seo-strategy-2026/LLM_FILES_MAINTENANCE.md](../seo-strategy-2026/LLM_FILES_MAINTENANCE.md). 4. **Skyscraper:** External guides emphasize **comprehensive topical coverage** and **modular sections**—consistent with [SKYSCRAPER_CONTENT_QUALITY_ENFORCEMENT.md](SKYSCRAPER_CONTENT_QUALITY_ENFORCEMENT.md) and `compare-content-to-competitors.php` gap lists. **Caveats:** Third-party statistics in marketing articles (e.g. CTR drops) were not adopted as facts; use for directional priority only. --- ## 2026-03-23 — Content ops as code, git editorial, validation pipelines **Sources (external):** - [Choosing a CMS for Technical Teams in 2026: Git-Based vs. Headless vs. Collaborative Markdown – HackMD](https://homepage.hackmd.io/blog/2026/02/04/tech-teams-cms-2026) — Git as coordination layer; trade-offs vs headless CMS for dev-heavy teams. - [ContentOps as Code: Implementing a CI/CD – TrySteakhouse](https://blog.trysteakhouse.com/blog/contentops-as-code-ci-cd-pipeline-geo) — Treat content like software; automate validation in delivery. - [Content as Code: Git-Based Workflows – Headless CMS Guide](https://headlesscms.guide/guides/content-as-code-git-based-workflows) — Versioning, review, rollback patterns for file-based content. - [MDX vs headless CMS: When Git Files Beat Admin Panel – Potapov.me](https://potapov.me/en/make/mdx-cms-alternative) — When Markdown/MDX in repo beats CMS for small/medium technical content sets. - [Building an Editorial Workflow with GitHub and Slack – Hyprnote](https://hyprnote.com/blog/building-an-editorial-workflow-github-slack/) — PR-based editorial flow complements file-first content. **Takeaways aligned with our stack:** 1. **Single source in Git:** Prefer **one canonical working file** per artifact (`content-draft.html` for blog; `content.md` for templates) and **iterate via commits**, not duplicate filenames—documented in [BLOG_CONTENT_EDIT_WORKFLOW.md](BLOG_CONTENT_EDIT_WORKFLOW.md). 2. **Validation as a pipeline:** `make blog-post-validate` / `**blog-post-validate-strict`** mirror “content + CI checks”; full validator matrix on every PR is optional ([CI_BLOG_CONTENT_GUARDRAILS.md](CI_BLOG_CONTENT_GUARDRAILS.md)). 3. **Scripted apply step:** `update-post-content.php` and `apply-and-validate-post.sh` keep JSON escaping safe—analogous to “build” steps in content-as-code posts. 4. **Future format (Phase C):** External consensus often favors Markdown/MDX for author ergonomics; we deferred MD migration until justified ([CONTENT_TOOLING_ROADMAP.md](CONTENT_TOOLING_ROADMAP.md)). 5. **Delivery vs staged artifacts:** Same “content as code” idea applies to FAQs—**one canonical payload** (`faqs` in post JSON) and **separate pipeline outputs** (`faq-research.json` → `faq-answers-optimized.json`) that feed scripted apply steps, not three hand-maintained parallel sources. See [FAQ_SOURCE_OF_TRUTH.md](FAQ_SOURCE_OF_TRUTH.md). 6. **Validators target shipped state:** FAQ quality checks read published post JSON, matching the principle that CI/gates validate what you ship, not intermediate generator files alone. --- ## 2026-03-23 — Answer-first without templated repetition (GEO / editorial) **Sources (external):** - [How to create answer-first content that AI models actually cite – Search Engine Land](https://searchengineland.com/guide/how-to-create-answer-first-content) — Direct answer up front; structure for extraction without sacrificing readability. - [Answer-first content: GEO writing pattern – Neo Core](https://theneocore.com/answer-first-content-geo-writing-pattern-for-ai-overviews/) — Lead with the answer, then context; inverted-pyramid style for generative surfaces. - [Skimmable, answer-first content blocks – Hashmeta](https://hashmeta.com/blog/how-to-build-skimmable-answer-first-content-blocks-that-boost-engagement-and-rankings/) — Short paragraphs, lists, clear H2/H3 hierarchy. - [Answer-first writing: structure AI wants to cite – aruntastic](https://aruntastic.com/blog/answer-first-writing-structure-ai-citations/) — Position and self-contained answers under relevant headings. **Takeaways aligned with our stack:** 1. **Keep answer-first** under question-shaped H2s (see [CONTENT_DEPTH_GUIDELINES.md](CONTENT_DEPTH_GUIDELINES.md)) but **avoid identical lead labels** in every section—variation and natural first sentences reduce “templated AI” perception while preserving extractability. 2. **Bold and lists** remain useful for scanability; use them **selectively**, not as a copy-paste prefix on every block. 3. **Operational guardrails:** Repo inventory via `audit-blog-repetitive-phrasing.py` / `make audit-blog-repetition`; per-post `validate-aeo-capsule-diversity.php` complements `validate-aeo-capsules.php` (length). --- ## 2026-03-23 — Skyscraper 2.0, topical authority, refresh cadence **Sources (external):** - [Skyscraper Technique Updated (2026) – ContentMation](https://contentmation.com/seo/skyscraper-technique-2026) — Modern skyscraper: differentiation, original data, distribution; technique still viable when bar is higher. - [SEO Skyscraper Technique: Complete Guide – Search Atlas](https://searchatlas.com/blog/seo-skyscraper-technique/) — Find strong pages, build better asset, outreach; notes algorithm and saturation context. - [Topical Authority and Content Clustering – SEO HQ](https://seohq.github.io/topical-authority-content-clustering) — Pillar/cluster model, internal linking, structured depth as authority signal. - [Content Marketing and SEO: How to Create Content That Ranks (2026) – Atlas Marketing](https://atlasmarketing.ai/surviving-algorithm-changes-a-guide-for-business-seo/) — E-E-A-T and resilience to algorithm shifts; quality and usefulness over tricks. **Takeaways aligned with our stack:** 1. **Differentiation over length:** Skyscraper wins on **unique value** (format, examples, freshness, first-party insight)—not padding. Matches [SKYSCRAPER_CONTENT_QUALITY_ENFORCEMENT.md](SKYSCRAPER_CONTENT_QUALITY_ENFORCEMENT.md) and `compare-content-to-competitors.php` gap fixes (substance, no filler). 2. **Topical authority:** Depth should sit in a **cluster context**—pillar, siblings, internal links. Operational docs: [docs/reference/architecture/content-clusters/CLUSTER_MAPPING.md](../../reference/architecture/content-clusters/CLUSTER_MAPPING.md); rules: `.cursor/rules/blog-cluster-mapping.mdc` (and `content-clusters.mdc`). 3. **PAA coverage ≠ competitive depth:** Completeness can show PAA addressed while **word ratio vs `competitive-depth-analysis.md`** still fails—use [CONTENT_PIPELINE_AUDIT_2026-03-23.md](CONTENT_PIPELINE_AUDIT_2026-03-23.md) as the canonical warning; `**compare-content-to-competitors.php**` is the ratio and gap list of record before publish. 4. **Refresh cadence:** Competitive lexikon/ratgeber benefits from **SERP re-validation + compare pass** when SERPs shift, not only annual figure updates—see `.cursor/rules/blog-evergreen-annual.mdc` for year-dependent fields. **Caveats:** Vendor “X% more traffic” claims were not adopted as facts; use for directional priority only. --- ## 2026-03-23 — Editorial speed without quality trade-offs (pipelines + checkpoints) **Sources (external):** - [Creating helpful, reliable, people-first content – Google Search Central](https://developers.google.com/search/docs/fundamentals/creating-helpful-content) — People-first content; avoid search-engine-first churn; quality and usefulness over production volume alone. - [What high-maturity content operations actually look like – Contently](https://contently.com/2025/12/28/what-high-maturity-content-operations-actually-look-like-top-10-platforms-for-2026/) — Mature ops embed expertise and review in the workflow (vendor context: treat platform rankings as directional, not endorsements). - [ContentOps as Code: Implementing a CI/CD – TrySteakhouse](https://blog.trysteakhouse.com/blog/contentops-as-code-ci-cd-pipeline-geo) — Automate repeatable checks; keep human judgment for strategy and final quality (already cross-linked in earlier log entry). **Takeaways aligned with our stack:** 1. **Layered validation:** Fast local iteration can use **narrow** checks (e.g. FAQ-only) when the change set is narrow; **pre-publish** still uses the full strict chain—mirrors “automate the boring parts, keep review for judgment” without skipping what ships. See [CI_BLOG_CONTENT_GUARDRAILS.md](CI_BLOG_CONTENT_GUARDRAILS.md) and [BLOG_WORKFLOW_EFFICIENCY.md](BLOG_WORKFLOW_EFFICIENCY.md). 2. **Single command surfaces:** Documented Make targets and `apply-and-validate-post.sh` reduce copy-paste errors and tool fragmentation (same idea as content-as-code pipelines in git). 3. **People-first / helpful content:** Google’s framework reinforces that efficiency gains should come from **clear process and structure**, not from dropping depth, factuality, or unique value—consistent with skyscraper and anti-fluff rules. **Caveats:** Third-party “maturity model” and platform list articles are used for workflow principles only, not product recommendations. --- ## 2026-03-23 — Research-backed outlines, quality over volume **Sources (external):** - [Creating helpful, reliable, people-first content – Google Search Central](https://developers.google.com/search/docs/fundamentals/creating-helpful-content) — Satisfying, original content; avoid churning pages that do not add value; focus on audience usefulness over production volume alone. - [How to create a content brief (with examples) – Ahrefs Blog](https://ahrefs.com/blog/content-brief/) — Briefs align writers (and tools) with search intent, structure, and subtopics before drafting—maps to our outline + `competitive-depth-analysis.md` + section briefs. **Takeaways aligned with our stack:** 1. **Outline = brief:** Treat `CONTENT_OUTLINE.md` + pipeline outputs (`competitive-depth-analysis.md`, PAA, competitor H2s) as the single planning surface before HTML or `content.md`—see [CONTENT_CREATION_DATA_CHECKLIST.md](../CONTENT_CREATION_DATA_CHECKLIST.md) for which data files to expect after each orchestrator. 2. **Depth ≠ word stuffing:** External guidance matches our rule: differentiation and usefulness beat “longer for its own sake”—use `compare-content-to-competitors.php` and unique-value checks, not raw word-count chasing. 3. **Refresh when reality shifts:** When SERP or competitor set moves materially, re-run the relevant collection step and update the outline; do not rely on stale `data/*.json` for a competitive pass. **Caveats:** Marketing “brief templates” vary by vendor; we adopt the *practice* (intent, structure, gaps), not a specific vendor format. --- ## 2026-03-23 — Version control vs backup, test restores **Sources (external):** - [Getting Started — About Version Control](https://git-scm.com/book/en/v2/Getting-Started-About-Version-Control) — Version control records snapshots of files over time; distributed workflows assume **multiple copies** (clones) as part of normal collaboration. - [Git Basics — Undoing Things](https://git-scm.com/book/en/v2/Git-Basics-Undoing-Things) — Recovery often uses history and refs; **practice** undo/restore flows so they work under pressure. **Takeaways aligned with our stack:** 1. **Git/GitHub = primary for committed blog JSON**; local `blog-snapshots/` (gitignored) addresses **pre-commit** and **bulk-script** windows—see [guides/BACKUP_GUIDE.md](guides/BACKUP_GUIDE.md). 2. **Test restores:** Periodically run `restore-from-snapshot.py --dry-run` (and rely on normal `git checkout` / revert for committed state) so recovery paths are not “first used in an emergency.” 3. **Org-level risk** (repo deletion, account issues) is a **policy/mirror** topic, not solved by committing duplicate post trees into the same repo. **Caveats:** General VC documentation does not replace your GitHub org backup settings; no vendor DR statistics cited here. --- ## 2026-03-23 — Agent-assisted editorial pipelines, review, validate what ships **Sources (external):** - [Best practices for coding with agents – Cursor](https://cursor.com/blog/agent-best-practices) — Plan before large edits; keep context focused; verify outputs (lint, tests, behavior)—maps to our Plan mode + `make blog-post-validate-strict` before publish. - [Subagents – Cursor Docs](https://cursor.com/docs/context/subagents) — Isolated context for exploration vs main thread—aligns with using built-in Explore/Shell for wide search or noisy commands while keeping the primary content brief in chat. - [Content workflow efficiency (internal)](BLOG_WORKFLOW_EFFICIENCY.md) — Layered validation (narrow vs strict) and single-command surfaces; we mirror “automate checks, keep judgment for humans” without skipping what ships. **Takeaways aligned with our stack:** 1. **Skills as routers:** Project skills under `.cursor/skills/` should point to `CONTENT_SYSTEM_INDEX.md` and rules—not replace validators—see `docs/ai/cursor-playbook.md` (**Content workflows**). 2. **Human checkpoints:** Outline gate, SERP review, and strict pre-publish targets remain **human-in-the-loop** quality bars; agents speed drafting and refactors, not final sign-off. 3. **Validate the shipped artifact:** FAQ-only shortcuts (`make blog-post-validate-faq`) do not replace full strict validation when body or meta changed on the same branch. **Caveats:** Vendor blog posts are directional; no adoption of uncited productivity statistics. --- **Last Updated:** 2026-03-23