Product transparency / Scoring
SEO Blitz Methodology
SEO Blitz Scoring Model v2.0 scores the text signals implemented in its published rule set. The score is a review aid, not a ranking forecast.
What the model evaluates
SEO Blitz reviews a supplied page title and body draft. The selected content mode changes the minimum depth, required lexical cues, some structure expectations, and the weight of each dimension. The result contains eight dimension scores, one rounded overall score, findings, and three prioritized revisions.
The eight dimensions are Clarity, Completeness, Structure, Readability, Specificity, Repetition, Title alignment, and Publication readiness.
How the score is calculated
- Each dimension begins at 100.
- Every triggered rule subtracts its configured deduction from its dimension.
- Each dimension is clamped at zero. Deductions can still produce findings after the dimension reaches zero, but they cannot reduce that dimension further.
- The selected mode applies one weight to each of the eight clamped dimension scores.
- The weighted values are added and the final score is rounded once to the nearest integer, then clamped from 0 to 100.
dimension = max(0, 100 - sum(deductions)) overall = clamp(0, 100, round(sum(dimension * mode weight)))
All six current weight sets total 1.00. Findings are ordered by severity, then deduction, then the active mode weight for the finding's dimension. High severity means 12 or more points, Medium means 6 to 11, and Low means 0 to 5. Severity describes a configured deduction, not measured SEO harm.
Six Content Modes
| Mode | Minimum words | Structure expectation | Required cue groups |
|---|---|---|---|
| Blog article or guide | 300 | At least 3 paragraphs and, above 150 words, at least 2 Markdown-like headings | Reader problem, example, proof or detail, next step |
| Landing page | 200 | At least 2 paragraphs | Offer, deliverable, fit, call to action |
| Product description | 120 | At least 2 paragraphs | Specifications, use case, caveat |
| Local service page | 200 | At least 3 paragraphs | Service area, problem, process, contact, scope |
| Help article | 150 | At least 2 paragraphs and ordered-step line cues above 80 words | Steps, prerequisites, outcome |
| Comparison page | 200 | At least 3 paragraphs and, above 150 words, at least 2 Markdown-like headings | Options, criteria, fairness, recommendation |
Mode cues are case-insensitive word and phrase patterns. A cue match does not prove that the draft covers the subject well. An equivalent unconfigured phrase can also produce a false finding.
Eight Dimensions
| Dimension | Implemented checks | Important boundary |
|---|---|---|
| Clarity | Configured hedging phrases and up to three sentences over 25 words | No passive-voice or jargon detector |
| Completeness | Under 50 words, under the mode minimum, and missing mode cue groups | Cues are lexical, not semantic proof |
| Structure | Markdown-like heading lines, paragraph count, Help step lines, and paragraphs over 250 characters | No rendered HTML hierarchy or visual-layout inspection |
| Readability | Average sentence length, long-sentence count, very-short-sentence majority, and average characters per word | No Flesch, Flesch-Kincaid, syllable, grammar, or comprehension score |
| Specificity | Configured generic claims, digit-plus-unit detail matches, and reader-question term overlap | No factual, citation, example-quality, or semantic verification |
| Repetition | Dominant eligible term and repeated exact three-word sequences | No conceptual similarity or ideal keyword-density calculation |
| Title alignment | Supplied title presence, length, body overlap, repeated title terms, and reader-question overlap | No live title tag or title-to-H1 comparison |
| Publication readiness | Title, mode depth, risky guarantee phrases, limitation cues, and link or next-step cues | No metadata, author, date, live-link, or page inspection |
Overlapping deductions
One draft characteristic can trigger more than one rule because rules can protect different dimensions. Long sentences can affect Clarity and Readability. A short draft can affect Completeness and Publication readiness. A missing title affects Title alignment and Publication readiness. Help mode uses separate completeness and structure expressions for steps. This is intentional v2.0 behavior. It is not independent evidence of ranking impact.
Inputs and display-only metrics
| Field or output | Role in v2.0 |
|---|---|
| Content mode | Score-affecting. Selects weights, minima, cue checks, and structure expectations. |
| Main content | Score-affecting and required for a report. |
| Page title | Score-affecting when present or absent. |
| Primary reader question | Optional and score-affecting for Specificity and Title alignment when its eligibility conditions apply. |
| Intended reader | Planning context. Stored and exported, but not scored. |
| Desired reader action | Planning context. Stored and exported, but not scored. |
| Estimated reading time | Display-only. Calculated as at least 1 minute, using 225 words per minute. It does not affect the score. |
Known boundaries
SEO Blitz analyzes submitted text. It does not crawl a live page, inspect a rendered DOM, validate links, inspect images or alt text, read Search Console or SERP data, verify facts, measure backlinks, or predict rankings. A high score means the draft triggered fewer or lower-weight deductions in this model. It does not prove that the page is accurate, helpful, original, technically sound, or likely to rank.
Detailed documentation
Scoring Model
Review every v2.0 rule, threshold, deduction, mode, and weight.
Limitations
See what the analyzer approximates, does not inspect, and leaves to human review.
Version History
Read the evidence-backed version record and current gaps.
Model Changelog
SEO Blitz Scoring Model v2.0 is the current model. The Version History separates verified repository implementation evidence from unknown public deployment history.
Privacy Model
The current calculator runs in the browser without an account requirement. SEO Blitz stores the current analyzer draft in browser localStorage. In the published Browser Privacy Verification study, submitted test canaries were not observed in captured request URLs, relevant headers, or request bodies during two controlled Chromium runs. Read the Privacy Policy for the broader technical disclosures and scope limits. The study does not establish all hosting, browser, extension, or future-version behavior.
Correction Process
Methodology issues can be reported through Contact. Include the page, issue, expected behavior, actual behavior, and a short non-confidential sample when the issue involves scoring.