Methodology / Boundaries
SEO Blitz Scoring Limitations
SEO Blitz Scoring Model v2.0 is a text-rule review tool. This page separates implemented analysis, lexical approximations, unsupported signals, and work that still requires people or external tools.
What SEO Blitz analyzes
- The text supplied in the main content field.
- The separate title supplied by the user.
- The selected content mode and its configured minima, cue groups, structure expectations, and weights.
- Word, sentence, paragraph, title, heading-pattern, repetition, and fixed lexical metrics defined in the Scoring Model.
- Exact normalized term overlap between the optional primary reader question, title, and body where SPC-03 and TTL-06 apply.
What SEO Blitz approximates
Several rules use lexical patterns as review cues. These checks are reproducible, but they do not understand meaning.
- Mode completeness uses configured words and phrases. A match can be irrelevant, negated, or superficial. A valid synonym can be missed.
- Headings are detected from Markdown-like raw lines. SEO Blitz does not inspect a rendered HTML heading hierarchy.
- Link readiness uses words such as
guide,next step, orread more. It does not confirm that a link exists. - Specificity uses a fixed generic-claim list and digit-plus-unit matches. It does not decide whether a statement is precise or supported.
- Brief alignment uses exact normalized term overlap. It does not understand synonyms, entities, morphology, or search intent.
- Repetition uses exact terms and exact three-word sequences. It does not detect conceptual repetition.
What SEO Blitz does not analyze
| Area | Unsupported capability |
|---|---|
| Live pages | No live URL crawling, fetching, rendered DOM inspection, HTTP status checks, redirect checks, or index-status checks. |
| Titles and metadata | No extraction of an HTML title tag, no title-to-H1 comparison, no meta-description checks, and no canonical or robots validation. |
| Links | No internal-link crawling, external-link verification, destination validation, broken-link verification, citation validation, or source-quality analysis. |
| Media | No image analysis, image relevance review, alt-text inspection, video analysis, or other media analysis. |
| Authority | No backlink analysis, referring-domain analysis, link-authority score, domain-authority score, author check, credential check, or publication-date check. |
| Search data | No SERP measurement, competitor comparison, ranking data, Search Console data, keyword-volume data, keyword-difficulty data, or click-through data. |
| Technical SEO | No Core Web Vitals, page-speed, layout-stability, mobile, schema, sitemap, or JavaScript-performance audit. |
| Language quality | No Flesch score, Flesch-Kincaid score, passive-voice detector, jargon detector, grammar checker, spelling checker, boilerplate analysis, or unanswered-question analysis. |
| Meaning and truth | No factual verification, semantic truth evaluation, semantic example analysis, search-intent determination, originality check, plagiarism check, or AI-generation detection. |
| Outcomes | No ranking prediction, ranking guarantee, indexing prediction, traffic prediction, conversion prediction, outcome attribution, revenue prediction, or AdSense approval prediction. |
Parser and language limits
- Sentence splitting uses periods, exclamation marks, and question marks. Abbreviations, decimals, code, and unusual punctuation can produce imperfect boundaries.
- Paragraph splitting uses blank lines or a newline followed by an uppercase character. The result can differ from visual paragraphs.
- Dense paragraphs are measured in characters, not words or rendered height.
- Title length uses JavaScript string length. Some emoji and combined Unicode characters can differ from perceived character count.
- Tokenization, stopwords, cues, and patterns are English-centered. Non-English and mixed-language drafts can receive misleading findings.
Scoring interaction limits
- One writing pattern can deduct points in more than one dimension.
- Dimension scores clamp at zero. Later deductions can still appear as findings without lowering the overall score further.
- Hard thresholds can make a small input change add or remove a finding.
- The rounded overall can hide small differences between unrounded results.
- Estimated reading time is display-only and has no scoring effect.
- Mode word minima are configured review boundaries, not evidence that a length is sufficient or necessary.
What requires human judgment
- Accuracy, evidence quality, source selection, and factual context.
- Audience fit, search intent, originality, usefulness, expertise, and first-hand experience.
- Brand voice, ethical risk, legal risk, safety, and final publication decisions.
- Whether a concise page is complete or a long page is padded.
- Whether a finding is a useful revision prompt or a false positive.
What requires external SEO tools
Use suitable external tools for live crawling, rendered DOM inspection, technical SEO, structured data, Core Web Vitals, broken links, backlinks, keyword research, SERP measurement, Search Console data, analytics, and index coverage. SEO Blitz does not substitute for those systems.
Correct interpretation
A higher SEO Blitz score means the submitted draft received fewer or lower-weight deductions under v2.0. It does not mean the page will rank, index, attract traffic, convert, or satisfy a search engine. The score is a review aid, not a ranking forecast or publication approval.
Return to the Methodology overview or inspect the complete rule reference.