How to run a content audit that finds revenue leaks

Somewhere around the fortieth post, a blog stops being a list of articles and becomes a portfolio: dozens of pages, each quietly earning or not earning, competing for queries, aging at different speeds. And almost nobody manages the portfolio. New posts get all the attention while last year's pages, which hold most of the accumulated rankings, drift without an owner.

A content audit is portfolio management for your published pages: one systematic pass that attaches performance data to every URL and assigns every URL an action. Done right, it routinely finds more traffic and revenue than a quarter of new publishing would, because fixing a page that already ranks is faster than ranking a new one. Done as most guides describe it, it produces a 40-column spreadsheet and no decisions. This is the decision-producing version.

The five leaks you're hunting

Knowing what a leak looks like keeps the audit focused. Five patterns account for most recoverable value on a small business site.

• Decayed winners. Pages that earned real traffic or leads a year ago and have slid, from rankings drifting, information going stale, or competitors publishing better versions. These are the highest-value finds because the demand is proven and the fix is a refresh, not a rebuild.
• Wrong-crowd magnets. Pages pulling healthy traffic that never converts, usually because the query attracts researchers, students, or DIYers rather than buyers. Sometimes fixable with a better next step, sometimes just mislabeled as successes in every report.
• Hidden converters. The mirror image: pages with modest traffic and outsized conversion rates. They're leaks in the sense that you're underinvesting in them: they deserve better internal links, position pushes, and siblings on adjacent topics.
• Cannibals. Two or three of your own pages targeting the same query, splitting the ranking signals between them so none ranks as well as one would. Common on blogs that have published for years without a map.
• Dead weight. Pages with no traffic, no conversions, no links, and no supporting role. A handful is normal. Dozens dilute the site's overall quality signal and your maintenance attention.

Step 1: build the inventory (30 minutes)

Start with a spreadsheet listing every content URL. Export the list from your sitemap or CMS. For each URL, you want a few identity columns you can fill quickly: topic, the query it was meant to target (if anyone recorded it), publish date, and last-updated date. Resist the urge to add twenty columns. Every column must feed a decision, and the decisions come from the data you'll attach next.

Small-site advantage: at 30 to 80 posts, this is an afternoon, not a project. Do it in one sitting so the snapshot is consistent.

Step 2: attach the performance data (45 minutes)

Two sources cover everything a small business needs.

Search Console, for the search side. For each page, pull twelve months of clicks, impressions, and average position, and, this is the part most audits skip, a comparison of the last three months against the same three months a year earlier. The comparison is what makes decay visible: a page can look fine on a twelve-month total while being down 60% year over year. While you're there, note each page's top queries, because "what Google actually shows this page for" frequently differs from what it was written for, and that mismatch explains many wrong-crowd magnets.

GA4, for the money side. Key events, Google's term for the conversions you've marked as mattering, by landing page, over the same period. This single view is the difference between auditing traffic and auditing revenue, and if you can't produce it, stop and set up conversion tracking first; auditing without it means optimizing for the wrong tier of metrics. Add engagement rate per page as a secondary signal for whether visitors who arrive actually stay.

Paste the numbers into the inventory. You now have, likely for the first time, every page's find-rate, click-rate, trend, and contribution to actual business on one screen.

Step 3: assign every page a verdict (60 minutes)

The heart of the audit is forcing a single decision per row. Five verdicts cover every page.

• Keep. Performing, current, converting at a rate you're happy with. No action beyond routine link additions from future posts. On a healthy site this is a big bucket; don't invent work for pages that are earning.
• Refresh. The decayed winners and near-missers: proven demand, slipping delivery. Update facts and dates, deepen the sections queries reveal people want, add your data and experience where the page is generic, and tighten the answer the page opens with. The quality bar for a refresh is Google's own audit question, "does the content provide original information, reporting, research, or analysis?", because a refresh that adds nothing original just resets the date on a mediocre page. The diagnosis-and-repair playbook covers the per-page details.
• Fix the funnel. The wrong-crowd magnets and any page with strong traffic, decent engagement, and no conversions. The content is doing its job; the next step is missing or mismatched. Give the page a relevant call to action, an internal link to the money page that actually fits its reader's moment, or a lead magnet suited to researchers if that's who arrives. When conversion machinery already exists but the crowd is simply never going to buy, relabel the page as a reach asset in your reporting and stop expecting revenue from it.
• Merge. For the cannibals: pick the strongest URL, fold the unique material from the duplicates into it, and redirect the losers to the winner. One consolidated page nearly always outranks the fragments it replaced, and position gains are top-weighted, the top three results capture over half of all clicks, so consolidating two page-two entries into one page-one contender is real money.
• Retire. For dead weight with no rescue case: redirect the URL to the closest relevant living page, or remove it. When in doubt between refresh and retire, ask whether you'd publish this topic today knowing what you now know about your buyers; the answer decides. (Retirement has its own discipline and its own traps, covered by this batch's companion piece on content pruning.)

Verdict discipline: every row gets exactly one, in one sitting, gut-called from the data in front of you. Audits die from deferred judgment, not from wrong judgment, and a mistaken refresh costs an hour while an unfinished audit costs the whole exercise.

A worked round of verdicts

Here's how the judgment feels in practice, five rows from a plausible small-business audit, each decided in under a minute.

*"Bathroom renovation costs 2024", 1,900 clicks last year, down 55% year over year, position drifted 3 → 9, two conversions.* A decayed winner wearing its problem in its title: the year. Verdict: refresh, urgently, update the numbers, re-date honestly, deepen the sections its queries ask for. Proven demand plus fixable staleness is the best trade on the sheet.

*"What is a soakwell?", 2,400 clicks, position 4, strong engagement, zero conversions ever.* A definitional query attracting DIY researchers. Verdict: fix the funnel, but with adjusted expectations: add a "when it's not a DIY job" section linking the drainage service page, then reclassify it as a reach asset. Not every good page is a sales page.

*"Drainage solutions Perth" and "Fixing drainage problems in Perth gardens", both limping at positions 12 and 15 for overlapping queries, splitting ~300 clicks.* Classic cannibals. Verdict: merge into the stronger URL, redirect the other, and expect the consolidated page to beat both, half of position improvements come from ending these civil wars.

*"Our thoughts on the 2019 budget's building incentives", 11 clicks lifetime, no links, no role.* Verdict: retire, redirect to the renovations hub. No ceremony required.

*"Choosing retaining wall materials", 90 clicks, position 8, four conversions.* The quiet star: a hidden converter punching far above its traffic. Verdict: keep, and invest, internal links from the high-traffic pages, a push on its main query, and two sibling posts next quarter. Finding this row is worth the whole audit.

Note what made each call fast: the data was already on the row, and the verdict vocabulary forced a choice. That's the entire design of the process, and why step 3 takes an hour, not a week.

Step 4: sequence by revenue proximity (15 minutes)

You'll finish step 3 with more actions than you'll do this month, so order matters. Rank the non-Keep pages by two questions: how close is this page to money (a decayed comparison page outranks a decayed glossary entry, always), and how good is the effort-to-impact ratio (a two-hour refresh of a page sitting at position 8 on a commercial query is the best trade in SEO). In practice the top of the list is nearly always: refresh the decayed converters, then fix the funnels on high-traffic pages, then merge the worst cannibals, with retirements batched for a single cleanup session.

A reasonable execution pace for a one-person operation is two to three actions a week, which clears a typical first audit in six to ten weeks, comfortably inside a quarter.

Step 5: close the loop

Each executed action gets a one-line log entry with the date, because refreshed pages need before/after judgment at the 8-to-12-week mark, the normal re-ranking lag. In Search Console, watch the refreshed pages' impressions and position against their pre-fix baseline; in GA4, watch conversions per page for the funnel fixes. Expect most actions to work modestly, a few to work dramatically, and a few to do nothing, and let the second audit learn from which was which.

Then put the cadence in the calendar: a full audit yearly, and a 30-minute quarterly mini-version that just re-checks the money pages and scans for new decays in the year-over-year comparison. The first audit is archaeology; the follow-ups are maintenance, and they keep the portfolio from ever again drifting unmanaged for years.

The one-line version

List every page, attach twelve months of Search Console and conversion data, force one verdict per page, keep, refresh, fix the funnel, merge, or retire, and execute in order of revenue proximity. It's an afternoon of analysis and a quarter of two-hour fixes, and it reliably recovers more revenue than the same effort spent writing into the void on top of an unmanaged pile.