Content Audit: A Step-by-Step Process With Template

A content audit is the inventory-and-verdict process: list everything you've published, measure what each piece is doing, and give every URL one of four verdicts — keep, update, merge or remove. Teams dread it as archaeology; in practice it's the highest-leverage week in the content year, because improving assets that already have age and links outperforms creating new ones at zero, and because libraries decay silently — every year, some fraction of your posts slide from earning to rotting, and only an audit notices. Here's the step-by-step, with the template structure built in.

Step 1: Build the inventory

One spreadsheet row per indexable URL. Sources: your CMS export, your sitemap, and a site crawl (the union catches the posts the CMS forgot and the crawler-only strays). Columns for the template: URL · title · publish date · last updated · primary keyword · word count — then the measurement columns you'll fill next. For a first audit on a big library, scope it: blog posts this pass, landing pages next; an audit that tries to boil the whole site usually ships nothing.

Step 2: Attach the numbers

Three data pulls, pasted against URLs: Search Console (twelve months of clicks, impressions, average position per page — the demand-side truth); analytics (sessions, engagement, conversions — the value-side truth); your link data (referring domains per URL, per the backlink audit workflow — the equity at stake, which decides how careful step 4 must be). Add one derived column: traffic trend — this year versus last. Trend is the audit's sharpest instrument: a post at position 8 and climbing needs patience; a post at position 8 after falling from 3 needs a refresh; flat-zero forever needs a harder verdict.

Step 3: The four verdicts

1. Keep — ranking, earning, current. Touch nothing; schedule its next routine review. On healthy libraries this is the biggest bucket, and resisting the urge to fiddle with winners is a discipline.
2. Update — right topic, decaying performance: dated examples, thinner than newer competitors, position sliding. The refresh treatment per the maintenance contract: current facts, added depth where the query grew, internal links to posts published since, meta re-written for CTR where impressions are high and clicks low.
3. Merge — multiple posts splitting one intent, the cannibalization signature (flip-flopping rankings, none winning). Consolidate into the strongest URL, 301 the others in, and inherit their links and history.
4. Remove — no traffic, no links, no strategic role, no salvageable topic. Delete honestly: 410, or 301 only where a genuinely relevant destination exists. Pruning dead weight concentrates crawl attention and sitewide quality signals — thin libraries drag on their own winners.

Step 4: Execute in leverage order

Verdicts are cheap; the queue is the deliverable. Sequence by return: first the high-impression, low-CTR updates (meta and title rewrites — hours of work, measurable in weeks); then page-two posts one refresh from page one (positions 8–15 with links already — the closest money); then merges (careful work — redirect mapping matters when links are at stake); removals last, batched. Feed the queue into the calendar's maintenance slots so the audit becomes work, not a document — the unexecuted audit being the genre's most common output.

Step 5: Make it a cycle

Full audit annually; a light quarterly pass over the top fifty URLs (trend column only — anything sliding gets queued). Each cycle gets faster: the spreadsheet persists, verdicts age predictably, and the library converges toward the shape the strategy intended — fewer, stronger, maintained pieces doing compounding work.

Frequently asked questions

How long does a content audit take?

For a few hundred URLs: a day to build the inventory and attach data, a day of verdicts, and the execution queue spread over the following quarter. The classic failure mode is inverting that ratio — weeks of spreadsheet perfection, zero refreshed posts.

Won't deleting old posts hurt my SEO?

Deleting earning or linked posts hurts; deleting true dead weight helps — sitewide quality is an aggregate, and the crawler's attention is finite. The link-data column exists precisely so nothing with equity gets binned; anything with referring domains gets merged or redirected, never removed cold.

What's the single most valuable thing audits find?

Almost always the page-two cluster: posts with age, links and near-miss rankings that one honest refresh moves to page one. Second: the cannibal pairs quietly capping each other. Both are rankings you already paid for, waiting to be collected — and once collected, the growth lever returns to the usual one, authority earned by links (where we come in).