Most content audits I've seen stop at a spreadsheet: every URL, its traffic, its word count, maybe a "keep/kill/update" column nobody revisits. That's not an audit — it's an inventory. An audit should end with a decision about what to do next, not a longer to-do list.
Here's the version I actually run.
Start with intent, not URLs
Before opening a crawler, write down what decision this audit needs to support. Are you trying to find quick-win rankings? Consolidate overlapping pages before a re-platform? Justify headcount for a content team? The questions you ask the data change completely depending on the answer, and skipping this step is why most audits produce a pile of observations nobody acts on.
Segment before you score
Pull the full URL list, then split it three ways before scoring anything:
- Pages with traffic, no conversions — these are usually a mismatch between search intent and page intent, not a content-quality problem.
- Pages with rankings, no traffic — often stuck on page two, or ranking for a query with declining volume.
- Pages with neither — the real kill/consolidate candidates, but only after checking they aren't quietly supporting internal linking to pages that do perform.
Scoring everything on one blended scale flattens these into noise. They need different fixes.
Match findings to a real competitor gap, not a generic checklist
Generic audit checklists ask if a page "covers the topic well." That's not answerable in a vacuum. Pull the top three ranking pages for the same query and compare structure, depth, and what they answer that yours doesn't. The gap is usually narrower and more specific than "needs more content" — often it's a missing comparison table, an unanswered objection, or a subtopic competitors treat as a full section and you treat as one paragraph.
End with a prioritized short list, not a full report
The deliverable isn't the spreadsheet — it's five to ten pages with a specific next action each, ranked by effort versus likely impact. If the audit doesn't produce something a content team could start executing on Monday morning, it wasn't finished.