Scope: what this audit covers and what it does not
The question this audit answers is whether a content programme is working towards business goals, with standards, distribution and upkeep that make each piece earn its cost, and what should change first. It scores the programme that produces content rather than the pages it has produced: goals and measures, audience research, pillars and the calendar, briefs, standards, expert input and AI review, people-first quality, distribution and repurposing, topic organisation for search, measurement, ownership and upkeep.
The page-by-page decisions to keep, refresh, consolidate or retire belong to the content inventoryA record of every published content page with its owner, review date, audit decision and changes, kept so the next review knows what exists and what happened. and audit that the content governance guide runs, recorded in the content inventory template; this scorecard asks only whether that inventory exists and is kept. Single pages are scored for search by the on-page SEO audit, and the brand's social channels by the social media presence audit. Search content production week to week is run by the SEO content operations playbook.
It does not re-teach the work. The content strategy framework and the content marketing guides linked in each area own their methods. Every scorecard sits in the audit library.
Before you score: access and evidence
Ask for the strategy or plan, the calendar, the style guide, access to the analytics reports used for content and the content inventory if one exists. Without the analytics reports you cannot see whether any strand does its job, so the measurement rows stay provisional until you have them. Fix an evidence window long enough for pieces to have been found, read and acted on, which for search content means looking well past publication, and keep the same window for every rescore.
Collect:
- the written content strategy or plan, with goals and the measures attached to each strand;
- the calendar for the evidence window and the briefs behind a sample of pieces;
- the style guide, the review steps and the record of who approves what;
- distribution records for a sample of pieces: where each went beyond publishing;
- performance reports for content grouped by strand and job;
- the content inventory with owners and review dates, if one exists.
Choose the sample across strands and authors rather than the pieces the team is proudest of. A programme is judged on what it usually produces.
Criteria and scoring
Rate each row missing (absent or broken), partial (present but incomplete, inconsistent or out of date) or solid (matches the "what good looks like" column). The scale is your judgement against that column, not a number. Rows marked gate decide the band on their own when they score missing. A row that does not apply, such as AI review for a team that does not use AI in drafting, is marked not applicable rather than missing.
Strategy
| Criterion | What good looks like | Score guidance |
|---|---|---|
| Goals and measures per strand (gate) | Each content strand has a job tied to a business goal and a measure that shows whether it is doing that job | Missing: content published to keep a schedule, with no job or measure / Partial: goals stated for the programme, not per strand, or measured by views alone / Solid: a job and a measure for every strand |
| Audience research | Topics come from customer questions, sales and support conversations and research, not from guesses | Missing: topics from the team's own ideas / Partial: search data only, or research done once / Solid: questions from several sources, kept current |
| Pillars and calendar | A few pillars organise the calendar, which plans by job and capacity rather than by date alone | Missing: a list of dates and titles / Partial: pillars named, calendar still sized to a fixed output / Solid: pillars drive the calendar, sized to real capacity |
Goals are scored against the content goals and KPIs guide, which sets a goal per funnel stage and replaces activity measures with ones that change decisions. Research is scored against the audience research guide, and the plan against the content pillars and calendar guide, which explains how each Content PillarOne of a small set of recurring themes that a brand's content is planned around. becomes topics and how the Editorial CalendarA shared schedule of what content will be published, on which channel, when and by whom. is sized to capacity with an evergreen base.
Production standards
| Criterion | What good looks like | Score guidance |
|---|---|---|
| Briefs for every piece | Every piece starts from a brief with audience, angle, evidence and the action it should prompt | Missing: a title and a word count / Partial: briefs for some pieces, or without the action / Solid: a brief for every piece, used at review |
| Editorial style guide | A style guide covers voice, terms and formatting, and writers use it | Missing: none / Partial: written but out of date or ignored / Solid: kept current from editors' corrections and used |
| Expert input | Pieces that need expertise draw on interviews with people who have it | Missing: expert topics written from other articles / Partial: experts asked for drafts they never send / Solid: prepared interviews, quotes checked with the expert |
| AI drafts reviewed | AI-assisted drafts go through human review for accuracy, originality and voice before publishing | Missing: AI drafts published unchecked / Partial: edited for voice, claims not checked / Solid: every claim sourced, voice restored, a review gate nothing passes without |
Briefs are scored against the content briefs tutorial, the Editorial Style GuideThe house rules a writer checks while drafting, so spelling, names, numbers, formatting, links and sources are handled the same way by everyone. against the style guide tutorial, expert input against the guide to interviewing subject-matter experts and AI review against the guide to AI-assisted content with review, which reads Google's guidance on helpful content as it is published rather than through summaries.
Quality
| Criterion | What good looks like | Score guidance |
|---|---|---|
| People-first content | Content shows real experience and expertise and is written for readers rather than for rankings | Missing: pieces written around keywords, no named author or evidence / Partial: experience shown on some pieces / Solid: authorship, evidence and first-hand experience throughout |
| Written for the web | Pages lead with the answer, are easy to scan and work on a phone | Missing: the point arrives late, walls of text / Partial: headings and lists, answer still buried / Solid: point first, scannable, readable on a phone |
People-first quality is scored against the E-E-A-T guide, which explains, from Google Search Central's helpful content documentation as it states it, why E-E-A-TExperience, Expertise, Authoritativeness, Trust. matters and how it is shown on a page and across a site. Writing is scored against the writing for the web guide.
Distribution
| Criterion | What good looks like | Score guidance |
|---|---|---|
| Planned distribution | Each piece has planned channels beyond publishing, chosen for where its audience is | Missing: published and left / Partial: shared on the brand's own social accounts by habit / Solid: channels chosen per piece and planned in the brief |
| Repurposing | Strong pieces are reworked into other formats on purpose rather than copied across | Missing: never reused, or excerpts pasted everywhere / Partial: reused without adapting to the format / Solid: chosen source pieces, adapted and tracked |
Distribution is scored against the distribution channels guide, which treats search and email as the base, and repurposing against the repurposing guide.
Measurement and search
| Criterion | What good looks like | Score guidance |
|---|---|---|
| Performance read by job (gate) | Content is judged against its job, grouped by strand, with assisted value considered alongside direct results | Missing: page views for the whole blog, or no reports / Partial: reports by page, not grouped by strand or job / Solid: grouped by strand, read against each job, assisted value considered |
| Topic organisation | Topics are organised into clusters with one page owning each query and links between them | Missing: pages competing for the same queries / Partial: clusters planned, links or ownership incomplete / Solid: a hub and supporting pages, one owner per query, linked |
Performance is scored against the guide to measuring content performance, which separates being found, engagement and contribution to an outcome and ends in a monthly review. Topic organisation is scored against the topic clusters guide and the topic clusters framework.
Upkeep
| Criterion | What good looks like | Score guidance |
|---|---|---|
| Owners and inventory | Every page has an owner and sits in an inventory with a review date | Missing: no inventory / Partial: an inventory without owners or review dates / Solid: owner and review date for every page, reviewed on schedule |
| Updates and corrections | Out-of-date pieces are updated, consolidated or retired, and corrections are made openly | Missing: old pages left as they were / Partial: updates made, dates and corrections unexplained / Solid: updates noted, corrections open, retired pages handled |
Ownership is scored against the content governance guide, which also runs the page-by-page inventory, and updates against the guide to updating and correcting content. Evergreen ContentContent that stays relevant and useful long after it is published, rather than being tied to a date or a news moment. still needs an owner: it is the part of a library most likely to keep ranking after its facts have moved.
How to score and prioritise
Score the gates first. Content with no job, or no way to read whether it does that job, cannot be judged, and every other score loses its point. If goals score missing, the research, standards and distribution rows describe how well the team makes something whose purpose nobody has set, so mark them provisional. If performance reading scores missing, nobody can say which strand to grow and which to stop.
Then rank the remaining weaknesses in this order:
- Quality and accuracy risks: unreviewed AI drafts, missing expertise and out-of-date pages that still rank. These put the brand's name on claims it cannot stand behind.
- Distribution, because a good piece nobody sees has the same result as a weak one.
- Production efficiency, style and calendar refinements next.
- Repurposing last, unless the calendar runs short of ideas.
Within each group, rank by the reach the weakness touches. An out-of-date guide that brings in most of the programme's search visits outranks the same fault on a page nobody finds.
Reading the result: what to do with each band
Red means a gate scored missing. Hold new topics and new formats, write a job and a measure for each strand, set up reporting grouped by strand, and decide nothing about growth until the new measures have been read. Rescore once they have run for a full reporting cycle.
Amber means the gates hold at least partially but several other rows are partial or missing. This is the usual outcome. Work a short ordered list, starting with accuracy and upkeep, and bring each change to the monthly content review so its effect is seen.
Green means the gates are solid and most rows are too. Move effort into the strands that do their job best, into distribution tests and into the inventory's refresh cycle, and rescore each year or after a change of team or strategy.
From findings to a prioritised fix list
Write each partial or missing row as a finding in plain words: what you saw, the evidence and why it matters. A dated page with a claim that is no longer true, or a strand with no measure at all, is worth more than a general remark about quality. Then turn the findings into one fix list with a line per fix: the finding, the fix, the owner, the guide that explains it and the rescore date. Order the list by the priorities above, not by how easy each fix is.
Group the lines by who does the work, because goals and measures sit with the content lead and whoever owns the business goals, reporting with the analyst, standards and reviews with the editor, distribution with the channel owners and upkeep with each page's owner. Hand the page-level decisions the audit uncovers to the content inventory, so this list stays about the programme. Present the band, the gate findings and the first few fixes to whoever funds the programme, and keep the full scorecard as the working copy for the rescore.
Illustrative example
A home energy advice company audits a busy content programme
Eskavor Energy Advice, an invented home energy advice company, publishes about 12 articles a month across 4 strands and cannot say what any of them achieves. It runs the audit over the last 6 months. Goals and measures score missing: the plan sets a publishing target and no strand has a measure. Performance reading scores missing, because the only report is total page views for the blog. Updates score missing: 7 guides on home energy grants, among the most visited pages, describe terms that changed about a year ago. Owners and inventory score missing, and planned distribution scores missing, because nothing is shared beyond the blog. Briefs score partial and writing for the web solid. The band is red on the goals gate. The fix list puts a job and a measure for each of the 4 strands and a report grouped by strand first, updating or withdrawing the 7 grant guides with open update notes second, an inventory with an owner and review date for every page third and a distribution line in every brief fourth, and holds new topics. After 3 months, both gates score solid, upkeep scores partial and the band moves to amber.
Common mistakes
- Scoring pages instead of the programme. A strong article does not make a strong programme. Score how content is made, and leave page decisions to the inventory.
- Accepting views as a measure for every strand. A strand meant to win enquiries is not doing its job because it is read. Score goals on whether each strand has its own measure.
- Judging the sample the team chooses. The proudest pieces are not typical. Sample across strands and authors.
- Treating publishing frequency as a strength. A full calendar with no job per piece scores missing on the gate, however busy it looks.
- Ignoring pages that still rank. Old pages with moved facts carry the most risk because people still find them. Score upkeep before style.
- Fixing everything at once. New strands, new formats and new reports together make the next review impossible to read. Work the list in order and rescore on the same evidence window.
