Every draft SEO Bison writes comes with a score. A number with no explanation is worth nothing, so this page explains exactly how it is built, including the parts that are less flattering. It describes the scoring that runs in the app today.
The short version
- A draft is scored on eight dimensions, each out of 100, and combined into an Overall score using fixed weights.
- The scoring is rules, not opinion. It reads the draft’s text and metadata and applies fixed checks, so the same draft always scores the same.
- A draft must reach 75 on its weakest dimension, not on average. Below that, it gets automatic rewrite passes.
- The score is a quality check on the draft. It does not predict rankings.
The eight dimensions, weights and checks
Overall is a weighted average. SEO carries the most weight at 30%. Points below are out of 100 within each dimension.
| Dimension | Weight | What it checks |
|---|---|---|
| SEO | 30% | Where the main (primary) keyword appears: the title (20 points), meta description (10), first 100 words (15), an H2 heading (15), the closing paragraph (10); keyword density between 0.5% and 2.5% (15); length up to 1,200 words (10); other target keywords used (5). |
| AEO (answer engines) | 15% | Whether the draft is easy to lift an answer from: an FAQ (25), key takeaways (15), a list (15), a table (15), at least two headings phrased as questions (15), short sentences (15). |
| GEO (generative search) | 15% | Whether a generated answer can quote it: your store or a product named in the text (20), a list (10), a table (10), real citations (15), a table of contents (15), key takeaways (10), short sentences (10), four or more H2 sections (10). |
| Content quality | 15% | Length up to 1,200 words (25), reading level around grades 6 to 9 (25), three or more H2 sections (15), a list (15), at least six paragraphs (10), short sentences (10). |
| E-E-A-T | 10% | Trust signals: a “Last updated” line (25), an author or publisher (15), your store or a product named (20), a list backed by figures or a table (15), product facts checked against your catalog (10), length up to 1,400 words (15). |
| Technical SEO | 5% | Article structured data (20), an FAQ (20), mobile-friendly (20), a page-speed baseline (20), and an H2 plus paragraphs (20). |
| UX | 5% | Reading level (30), short sentences (20), a list or table (20), at least one image (15), responsive layout (15). |
| Metadata and linking | 5% | A title of 40 to 65 characters (25), a meta description of 120 to 165 characters (25), 3 to 8 links to your own products (30), no “click here” or “read more” anchors (20). |
Where the score is high by design
Here is the candid part. Several dimensions mostly check for structure: an FAQ, key takeaways, a table, lists, a table of contents, headings phrased as questions. SEO Bison drafts are built to include all of those, so AEO, GEO, content quality, UX and metadata tend to score high by design. A high number there tells you the template was followed, not that the writing is brilliant.
Technical SEO is the clearest case. Sixty of its hundred points are for things that are assumed true rather than measured per post: Article structured data (added when you publish), a mobile-friendly Shopify theme and a page-speed baseline. A well-formed draft therefore lands at or near 100 almost every time. We would rather say so than have you read it as a differentiator.
SEO is the dimension that actually varies. It depends on whether the main keyword lands in the title, meta description, opening, a heading and the close, and on how often it appears. That is why it is the heaviest weight, and why the first thing to look at on a low score is the main keyword’s placement.
Two checks depend on what happened in your store: E-E-A-T’s 10 points for product facts checked against your catalog are only awarded if that check actually ran on the draft, and GEO’s evidence points count real citations that survived SEO Bison’s link check, in line with the rule that drafts only cite sources found in live research.
The quality floor: 75 on the weakest dimension
A draft passes only when the lowest of the eight scores reaches 75. A post that is 100 on seven dimensions and 40 on one fails. When that happens SEO Bison runs up to two targeted rewrite passes (within a per-post cost limit), then one full rewrite if the bar is still not met. A targeted pass is kept only if it raises the weakest dimension without lowering the overall score by more than two points, and the full rewrite is kept if it improves the weakest dimension or the overall score. Otherwise the earlier draft stands.
One detail worth knowing: the floor check runs while the draft is being written, before the fact check and the final citation count are known. The score you see afterwards includes both, so GEO or E-E-A-T on the finished post can read lower than the gate saw. The headline Overall for a fresh draft is also capped at 95, so a first draft never shows a flat 100.
Improve never makes it worse
Improve with SEO Bison rewrites a draft toward its weak dimensions. If the rewrite scores worse, it retries once with explicit “do not regress” feedback. If the result still scores lower overall than your current draft, SEO Bison keeps the original. It works on drafts, not on articles already published in Shopify, and each plan includes an allowance of improvements (see pricing).
What the score cannot tell you
It cannot tell you where a post will rank, how much traffic it will earn, or whether the topic is right for your store. It does not use search volume, and SEO Bison does not track rankings, clicks or impressions. To pick topics with a realistic shot, start with keyword research, and to check the rest of your store, run the Shopify SEO audit. For practical guides, see the blog.