Scope: what this audit covers and what it does not
The scorecard is applied page by page. It covers what sits on the page and in its head: whether the content fits the query it targets, the title and meta descriptionAn HTML summary of a page that search engines may use as the snippet under its title link., the heading outline, how thoroughly the topic is covered, the signals that show who wrote it and why it can be trusted, the links into and out of the page, image handling and structured data. Image handling includes whether each informative image carries descriptive alt textThe text alternative to an image, read by screen readers and used by search engines.. Score one page per row of a spreadsheet, then compare.
It does not teach writing for search. Reading intent from the results page belongs to the search intent guide, and outlining a page belongs to the headings and content structure guide. New pages are covered at launch by the SEO page publishing checklist, which is ticked once rather than scored. Anything that applies across the whole site, such as indexation, canonicals or speed, belongs to the technical scorecard elsewhere in the audit library. Score those here only if a single page behaves differently from its template.
Before you score: access and evidence
Pick the pages first. A sensible shortlist mixes pages that earn revenue or leads, pages that have lost clicks over recent months, and pages that rank just off the first results page. For each one, write down the main query it is meant to answer, then open the live SERPThe page of results a search engine shows. for that query in a clean browser and note what kind of results win: guides, product pages, comparisons, tools or videos.
From Search Console, export the queries and clicks each page receives for a fixed window, so you can see whether it ranks for what you think it targets. Check whether a second page on the site competes for the same query; if one does, the keyword cannibalisation guide explains how to decide which page should own it. Record the author, the last meaningful update date and any reviewer shown on the page.
Criteria and scoring
Rate each row missing, partial or solid for every page. Intent match is the gate.
| Criterion | What good looks like | Score guidance |
|---|---|---|
| Intent match (gate) | Page type and angle match what wins the results page for its query | Missing: wrong page type / Partial: right type, wrong angle or depth / Solid: clear fit |
| Single owner per query | No other page on the site competes for the same main query | Missing: several pages compete / Partial: some overlap / Solid: one clear owner |
| Title tag | Specific, distinct, front-loads the topic and fits the display width | Missing: generic or duplicated / Partial: accurate but weak / Solid: specific and compelling |
| Meta description | Summarises the page's answer and gives a reason to click | Missing: absent or duplicated / Partial: generic / Solid: specific to the page |
| Headings and outline | One main heading; subheadings follow the questions a reader has, in order | Missing: no structure / Partial: skips or vague labels / Solid: clear outline |
| Opening answer | The first lines confirm the reader is in the right place and answer the core question | Missing: long preamble / Partial: answer buried / Solid: early answer |
| Content depth | Covers the subtopics a reader needs, without padding | Missing: thin / Partial: gaps against competing pages / Solid: complete and focused |
| Freshness | Facts, screenshots and references reflect the current state | Missing: clearly outdated / Partial: partly stale / Solid: current |
| Experience and expertise signals | Named author with relevant background; first-hand detail where it matters | Missing: anonymous and generic / Partial: author but little substance / Solid: credible and specific |
| Trust signals | Sources cited, claims supported, contact and ownership clear | Missing: unsupported claims / Partial: some sources / Solid: well supported |
| Internal links in | Relevant pages link to it with descriptive anchors | Missing: orphan or near orphan / Partial: few or vague anchors / Solid: well linked |
| Internal links out | Links onward to the next useful page in the topic | Missing: dead end / Partial: generic links / Solid: deliberate next steps |
| Images | Images add information, carry descriptive alt text and are compressed | Missing: decorative stock, no alt text / Partial: some gaps / Solid: useful and optimised |
| Structured data | Markup matches the page type and visible content | Missing: none where useful / Partial: errors or mismatch / Solid: valid and accurate |
Treat the experience and trust rows as judgements about the page, not about the brand. The SEO Starter Guide frames helpful, reliable content in plain terms, and that is the standard to score against. For the structured data row, score Schema MarkupStructured data vocabulary. only where the page type has a relevant type; a missing markup block on a plain article is not a failing.
How to score and prioritise
Score intent first. If a page fails it, stop scoring that page's other rows in detail and mark it for rewrite or consolidation, because tuning headings and links on the wrong page is wasted effort. For pages that pass, score the remaining rows quickly and consistently; two people scoring the same page should land on the same level, so agree examples of partial and solid before starting.
Prioritise across pages, not within them. Multiply in your head the page's importance by how far it falls short: a priority product guide with several partial rows outranks a minor blog post with several missing ones. Pages that rank just off the first page and score amber are usually the fastest wins in SEOImproving visibility in unpaid search results., because modest improvements can be enough to move them. Pages that compete with each other are resolved before either is improved, otherwise the work splits the signal further.
Reading the result: what to do with each band
Red means the page fails intent. Decide between a rewrite to match the winning page type, a merge into the page that should own the query, or retirement with a redirect. Do not patch it.
Amber means intent fits but several rows are partial or missing. Give the page a structured tune-up: rework the opening and outline, fill the depth gaps, refresh the facts, add author and source signals, and fix internal links in both directions. Rescore after the change has been crawled.
Green means the page meets the standard. Leave it alone apart from a freshness check on a fixed rhythm, and use it as the internal example of what partial and solid look like for the next scoring round.
From audit to action plan
Collect every page's band in one sheet and sort by importance, then band. Each red or amber page becomes a brief for a writer with the target query, the winning page type, the rows to fix and the pages that should link to it. Group pages by topic so a cluster is reworked together and its internal links are fixed in one pass rather than piecemeal.
An accounting software blog shortlist
A small accounting software company scores 24 priority blog pages. Five fail intent: four are opinion pieces targeting queries where the results page shows step-by-step how-to guides, and one competes with a product page for "invoice template". Eleven score amber, mostly on depth, freshness and missing author information. Eight are green. The team merges the competing post into the product page, rewrites the four opinion pieces as how-tos, and gives the eleven amber pages a tune-up in two topic batches. After three months, nine of the reworked pages gain clicks; two of the rewrites still lag and are rescored with fresh results-page notes.
Common mistakes
- Polishing pages with the wrong intent. A better title on the wrong kind of page does not change what searchers want. Fail intent, then rewrite.
- Scoring every page on the site. Shallow passes over hundreds of pages produce noise. Score a shortlist well.
- Counting words as depth. Depth is coverage of what the reader needs. A longer page full of filler scores partial at best.
- Treating author boxes as trust. A name without relevant substance or sources does not move the trust row. Score the evidence on the page.
- Fixing competing pages separately. Decide which page owns the query first, then improve that one.
- Rescoring before the change is crawled. A rewrite that has not been picked up yet looks like a failure. Wait until the new version is indexed.
