Methodology / Rule reference
SEO Blitz Scoring Model v2.0
The complete public rule map for the shipped analyzer. It documents all 30 score-affecting rule sites, exact v2.0 weights, thresholds, deductions, and mode behavior.
Calculation
Every dimension begins at 100. Triggered findings subtract points only from their listed dimension. A dimension is clamped at zero before weighting. The eight weighted dimension values are added, rounded once to the nearest integer, and clamped from 0 to 100.
dimension = max(0, 100 - sum(rule deductions)) overall = clamp(0, 100, round(sum(dimension * selected mode weight)))
All current mode weights total 1.00, so the weight denominator is 1.00. A rule that deducts 8 points from a dimension weighted at 0.15 changes the unrounded overall by 1.2 points, unless that dimension is already at zero.
Findings are sorted by severity, deduction, and then the active mode weight for their dimension. The first three are priority revisions. High severity is 12 points or more. Medium is 6 to 11 points. Low is 0 to 5 points.
Exact content-mode weights
| Mode | Clarity | Completeness | Structure | Readability | Specificity | Repetition | Title alignment | Publication readiness | Total |
|---|---|---|---|---|---|---|---|---|---|
| Blog article or guide | 0.12 | 0.18 | 0.15 | 0.12 | 0.14 | 0.07 | 0.10 | 0.12 | 1.00 |
| Landing page | 0.13 | 0.18 | 0.12 | 0.10 | 0.15 | 0.07 | 0.13 | 0.12 | 1.00 |
| Product description | 0.13 | 0.17 | 0.10 | 0.12 | 0.18 | 0.06 | 0.10 | 0.14 | 1.00 |
| Local service page | 0.12 | 0.19 | 0.13 | 0.11 | 0.14 | 0.06 | 0.10 | 0.15 | 1.00 |
| Help article | 0.16 | 0.16 | 0.17 | 0.13 | 0.12 | 0.06 | 0.08 | 0.12 | 1.00 |
| Comparison page | 0.12 | 0.19 | 0.15 | 0.11 | 0.15 | 0.06 | 0.10 | 0.12 | 1.00 |
Mode minima and expectations
| Mode | Minimum | Paragraph minimum | Extra structure | Required lexical cue groups |
|---|---|---|---|---|
| Blog | 300 words | 3 | At least 2 Markdown-like headings above 150 words | Reader problem, example, proof or detail, next step |
| Landing | 200 words | 2 | None | Offer, deliverable, fit, call to action |
| Product | 120 words | 2 | None | Specifications, use case, caveat |
| Local | 200 words | 3 | None | Service area, problem, process, contact, scope |
| Help | 150 words | 2 | Ordered-step expression above 80 words | Steps, prerequisites, outcome |
| Comparison | 200 words | 3 | At least 2 Markdown-like headings above 150 words | Options, criteria, fairness, recommendation |
The mode minimum is a configured review threshold, not a universal content-length requirement. CMP-01 applies below 50 words. CMP-02 applies from 50 words to one word below the mode minimum. RDY-02 also applies below the mode minimum.
Help mode has two step expressions
The Completeness cue accepts a numbered raw line, step plus a number, or words such as first, second, third, then, next, and finally. STR-03 is stricter. It expects a numbered raw line, step plus a number, or first,, second,, then,, next,, or finally,. A Help draft can pass one check and fail the other.
Mode-required cue alternatives
CMP-03 performs case-insensitive lexical matching. These alternatives are the public reproduction contract. Variations not listed here can fail even when a human considers them equivalent.
| Mode | Cue group | Accepted alternatives |
|---|---|---|
| Blog | Reader problem | reader problem, reader question, reader needs, reader wants, intended reader, target reader, or audience |
| Blog | Example | example, for instance, scenario, case, or e.g. |
| Blog | Proof or detail | because, so that, measurement, data, study, evidence, proof, result, percent, inch, feet, cm, kg, or trial |
| Blog | Next step | link, read more, see also, related, next step, guide, checklist, workflow, or learn more |
| Landing | Offer | your or you paired with archive, team, page, tool, audit, service, product, guide, or report; we paired with review, audit, deliver, build, help, or offer; the audit; or the deliverable |
| Landing | Deliverable | deliverable, you get, you receive, include, includes, what you, or what we |
| Landing | Fit | good fit, poor fit, right for, not for, best for, or work best |
| Landing | Call to action | get started, contact, request, book, schedule, try, sign up, cta, call to action, or next step |
| Product | Specifications | inch, inches, cm, mm, kg, lb, watt, led, cord, base, arm, finish, size, weight, material, battery, replaceable, dimension, or dimensions |
| Product | Use case | built for, designed for, work best, made for, renters, students, remote workers, or ideal for |
| Product | Caveat | not designed, before you buy, before you order, before you choose, configured desk, outlet, space, or depth checks, won't, cannot, is not, limitation, or limitations |
| Local | Service area | A north, south, east, or west phrase followed by a word; downtown, service area, nearby, neighborhood, city of, or in followed by a word |
| Local | Problem | leak, leaking, damage, repair, broken, clog, stain, fail, emergency, issue, problem, or symptom |
| Local | Process | we paired with help, inspect, explain, repair, replace, shut, or turn off; first, then, during the visit, or what we do |
| Local | Contact | call, contact, schedule, book, appointment, reach us, before calling, or shutoff valve |
| Local | Scope | may require, follow-up, followup, parts, access work, configured what we will or what we won't, cannot, does not, or scope |
| Help | Steps | A raw line starting with a number plus . or ); step plus a number; first, second, third, then, next, or finally |
| Help | Prerequisites | before you start, configured you need or you will need, required, prerequisite, configured tool you or tools you, or make sure |
| Help | Outcome | Configured phrases beginning with when you, once fixed, once done, once finished, result, expected, or should now |
| Comparison | Options | vs, versus, or followed by a word, alternative, option plus one digit, two options, two choices, or two products |
| Comparison | Criteria | criteria, compare by, compared on, configured by price, by size, by space, by need, or by use, factor, tradeoff, trade-off, whereas, or while |
| Comparison | Fairness | whereas, while, on the other hand, downside, upside, better for, worse for, or configured pro and con or pros and cons |
| Comparison | Recommendation | choose, pick, recommend, best for, right for, if you, or go with |
All 30 score-affecting rule sites
Rule IDs are stable public references for v2.0. A looping site can emit more than one finding. Severity below follows the deduction produced by that finding.
| Rule ID | Dimension | Trigger or threshold | Deduction | Severity | Modes | Notes |
|---|---|---|---|---|---|---|
CLR-01 | Clarity | At least 2 total matches from the fixed hedging phrase list | min(2 + count, 10) | Low at 2 to 3 matches, Medium at 4 or more | All | Maximum deduction 10 |
CLR-02 | Clarity | Each of the 3 longest sentences over 25 whitespace-separated words | 5 at 26 to 30 words, 8 over 30 | Low or Medium | All | Up to 3 findings |
CMP-01 | Completeness | Word count below 50 | 18 | High | All | Mutually exclusive with CMP-02 |
CMP-02 | Completeness | Word count at least 50 but below the selected mode minimum | 10 | Medium | All | Mode minima are 300, 200, 120, 200, 150, and 200 |
CMP-03 | Completeness | One selected mode-required lexical cue group does not match | 8 per missing group | Medium | All, with mode-specific groups | Can emit 3 to 5 findings by mode |
STR-01 | Structure | Over 150 words and fewer than 2 Markdown-like heading lines | 6 | Medium | Blog, Comparison | Does not inspect rendered HTML headings |
STR-02 | Structure | Below the mode paragraph minimum and over 80 words | 5 | Low | All | Minimum is 2 or 3 by mode |
STR-03 | Structure | Over 80 words and no strict ordered-step raw-line cue | 7 | Medium | Help | Uses the stricter Help step expression |
STR-04 | Structure | Each of the first 3 paragraphs over 250 characters | 4 each | Low | All | Character count, not word count |
RDB-01 | Readability | Rounded average sentence length over 25 words | 8 | Medium | All | Mutually exclusive with RDB-02 |
RDB-02 | Readability | Rounded average sentence length from 22 through 25 words | 4 | Low | All | Mutually exclusive with RDB-01 |
RDB-03 | Readability | At least 1 sentence over 25 whitespace-separated words | min(long sentence count * 2, 8) | Low at 1 to 2, Medium at 3 or more | All | Maximum deduction 8 |
RDB-04 | Readability | More than half of sentences have 1 to 5 words, with more than 5 sentences total | 3 | Low | All | Uses whitespace-separated sentence tokens |
RDB-05 | Readability | Over 50 words and rounded average visible non-whitespace characters per word over 5.0 | 3 | Low | All | Not a syllable or grade-level calculation |
SPC-01 | Specificity | At least 3 whole-word matches from the fixed generic-claim list | min(2 + count, 10) | Low at 3, Medium at 4 or more | All | Maximum deduction 10 |
SPC-02 | Specificity | Over 120 words and fewer than 2 digit-plus-unit matches | 5 | Low | All | Recognized units include percent, dimensions, weights, time, money, users, customers, and pages |
SPC-03 | Specificity | Reader question has at least 2 eligible terms and none occurs in the body | 5 | Low | All | Exact normalized lexical overlap only |
REP-01 | Repetition | Top eligible keyword occurs at least 6 times and is strictly over 8% of all word tokens | 6 | Medium | All | Not an ideal keyword-density target |
REP-02 | Repetition | Any exact normalized 3-word sequence appears at least 3 times | 5 | Low | All | Reports the highest-count sequence |
TTL-01 | Title alignment | Trimmed supplied title is empty | 12 | High | All | Also triggers RDY-01 |
TTL-02 | Title alignment | Supplied title is shorter than 30 JavaScript UTF-16 code units | 5 | Low | All | Only when a title exists |
TTL-03 | Title alignment | Supplied title is longer than 65 JavaScript UTF-16 code units | 4 | Low | All | The display badge changes above 60, but scoring changes only above 65 |
TTL-04 | Title alignment | At least 1 eligible title term exists and none occurs in the body | 6 | Medium | All | Exact normalized token membership |
TTL-05 | Title alignment | One eligible title term occurs at least 3 times in the title | 4 | Low | All | Checks the supplied title only |
TTL-06 | Title alignment | Reader question has at least 2 unique eligible terms and none overlaps title terms | 5 | Low | All | Exact normalized lexical overlap only |
RDY-01 | Publication readiness | Trimmed supplied title is empty | 10 | Medium | All | Also triggers TTL-01 |
RDY-02 | Publication readiness | Word count below the selected mode minimum | 10 | Medium | All | Can overlap CMP-01 or CMP-02 |
RDY-03 | Publication readiness | Visible text matches a configured guarantee for rankings, traffic, or results | 12 | High | All | Fixed expressions, not general factual analysis |
RDY-04 | Publication readiness | Over 150 words and no configured limitation or caveat cue | 5 | Low | All | Presence check, not quality review |
RDY-05 | Publication readiness | Over 200 words and no configured link or next-step cue | 4 | Low | All | Does not parse or validate a link |
Parsing and threshold details
Heading detection
STR-01 detects raw lines that look like Markdown headings, using one to six leading # characters or one to three leading * characters. It does not inspect a rendered H1 through H6 hierarchy. HTML pasted into the body is converted to visible text for most analysis, so rendered HTML heading elements do not satisfy this rule as heading nodes.
Lexical link cues
RDY-05 and the Blog next-step cue look for words such as link, read more, see also, related, next step, guide, checklist, workflow, or learn more. The analyzer does not perform broken-link detection, destination validation, HTTP status checking, internal-link crawling, or external-link verification.
Digit-plus-unit specificity
SPC-02 recognizes a digit, an optional decimal, and one of these units: %, percent, inch, inches, ft, feet, cm, mm, kg, lb, lbs, hours, minutes, seconds, days, weeks, months, years, dollars, $, users, customers, or pages. Singular and plural time, user, customer, and page forms are accepted. Examples that match include 25%, 3 hours, 12 cm, $40, and 200 users. A bare number, a spelled-out number, or a unit outside the list does not match. This is lexical pattern matching, not semantic understanding.
Fixed phrase lists
CLR-01 checks: in order to, due to the fact, a number of, it is important to, it should be noted, for the purpose of, in the event that, basically, actually, very, really, quite, various, numerous, and several.
SPC-01 checks: best, top, leading, premier, world-class, quality, comprehensive, cutting-edge, state-of-the-art, robust, scalable, seamless, innovative, trusted, ultimate, premium, and expert.
RDY-03 checks configured forms of guaranteed first-page, page 1, number 1, or top ranking; guaranteed ranking; promised ranking; first-page ranking; top of Google; number 1 ranking; guaranteed traffic; and guaranteed results. RDY-04 checks configured cues for a limitation, caveat, drawback, risk, warning, before-you statement, check-whether statement, or a statement that something does not or cannot apply.
Title thresholds
The title length badge changes when the supplied title exceeds 60 code units. TTL-03 deducts scoring points only when it exceeds 65 code units. These are two separate current UI behaviors. The difference can be confusing and is a follow-up product issue, but it does not change v2.0 scoring documentation.
Non-scoring fields and outputs
Intended reader and desired reader action are stored as planning context and included in exports, but they do not affect v2.0 scores. The primary reader question can affect SPC-03 and TTL-06. Estimated reading time is display-only. It is calculated with 225 words per minute and does not affect any dimension or the overall score.
Interpretation boundaries
The score reports how the submitted text matches these rules. It does not represent Google's algorithm, confirm optimization, predict rankings, predict indexing, predict traffic, or prove that a page will convert. A passing rule is not a search-engine endorsement.
For a concise explanation, read Methodology. For unsupported signals and human-review requirements, read Limitations. For evidence-backed model dates, read Version History.