Content Audit Methodology: Scoring Traffic, Rankings, and Conversions to Decide What to Do
1. Why You Need a Three-Dimension Audit: Traffic Alone Misleads You
The most common approach to a content audit is to export all 150 pages, rank them by GA4 traffic, keep the top 20%, and delete the bottom 20%. On B2B websites, this method often kills good pages by mistake.
Here’s the counterintuitive part: B2B buying cycles run 3 to 9 months, so a long-form article that gets only 3 visits a month might be read over and over by a future customer. By 90-day traffic alone, it belongs in the trash. By conversions, it might be the only page that brought in an inquiry this quarter.
So a content audit has to look at all three dimensions together: traffic answers “is anyone reading this,” rankings answer “does Google think you deserve to rank,” and conversions answer “did the people who read it take action.” Only by crossing all three can you decide whether a page should be updated, merged, redirected, or taken down.
| Type of Misjudgment | What not to do | Consequence | The right way |
|---|---|---|---|
| Traffic only | Delete pages with under 10 monthly UVs | Kills decision-reference pages for future customers | Layer in 90-day conversions and primary keyword rankings |
| Rankings only | Keep all pages in the top 20 | Keeps thin pages with high rankings and zero conversions | Layer in inquiry CVR and CTA placement |
| Conversions only | Keep all pages with leads | Keeps thin pages that convert by accident on brand terms | Layer in primary keyword rankings and organic traffic trends |
This article covers what to do after you’ve taken inventory. For how to inventory 150 pages and classify them, see Site-wide content inventory. For judging whether a single article should be updated, rewritten, or merged, see Refreshing Old Articles. The three-dimension audit is the batch execution layer for both.
2. How to Pull Data for the Three Dimensions: Tools, Windows, Thresholds
The data windows for all three dimensions must match, or the comparison is meaningless. Use a consistent 90-day window and align GA4 and Search Console reporting definitions. Don’t use 30 days for one and 90 days for another.
| Dimension | Tools | Window | Core fields | Healthy Threshold |
|---|---|---|---|---|
| Traffic | GA4 | Last 90 days | Average monthly UVs, rolling trend | Monthly UVs ≥ 50 and a stable trend |
| Rankings | Google Search Console | Last 90 days | Average position for primary keywords, impressions | Average position ≤ 20 for target keywords |
| Conversion | GA4 events + CRM | Last 90 days | Inquiry count, CVR | CVR ≥ site average |
| Cannibalization | GSC query grouping | Last 90 days | Pages ranking for the same term | ≤ 1 page per site in the top 10 |
Three measurement traps that are easy to fall into. In GA4, exclude internal traffic and bots from UV first. In GSC, “average position” swings wildly below 100 impressions, so check impressions before position. Conversion definitions must align with How to calculate lead attribution correctly and never count WhatsApp or phone clicks as form inquiries.
Thresholds are not fixed numbers. For a B2B industrial site, 50 monthly UV is healthy; a SaaS site may need 500. Use your own distribution over the past 12 months to calculate percentiles. Only below the 25th percentile is an outlier. Do not apply published industry numbers directly.
3. Action decision matrix: six page types, six actions
Put the three dimensions into one table. Every page lands in a cell, and the action follows. This is the core decision framework of the audit, and the dividing line between “changing content on a hunch” and “acting on data.”
| Traffic | Rankings | Conversion | Action |
|---|---|---|---|
| High | Top 20 | ≥ average | Keep, minor update once a year |
| High | Top 20 | Below half the average | Leave traffic alone, change CTA and form fields |
| Declining | Top 20 | Normal | Refresh first: update data + add internal links |
| High | Dropped out of top 30 | Normal | Rewrite title and H1, check whether a competitor has overtaken you |
| Low | Dropped out of top 50 | No conversions | Merge into a sibling page, 301 to the new URL |
| Low | Dropped out of top 50 | Has conversions | Keep and watch for a quarter, add internal link equity |
The decision tree in one sentence: check conversions first. A page with inquiries is never deleted, no matter how low its traffic. Then check rankings: pages with the main keyword in the top 30 get refreshed first. Finally check traffic: only merge or retire a page when all three dimensions are weak.
Merging is not simple copy-paste. When two articles cover the same process parameters, the old URL must 301 to the one you keep. Do not leave a 404. For when to merge and when to split, see keyword cannibalization. For how to distribute internal link equity after a merge, see internal linking system.
4. Two scenarios: different calls for a B2B long-form article and a lead-gen page
Scenario 1: a long-form article on “stainless steel flange PN16 pressure ratings,” 12 monthly UV, 2 inquiries in the past 90 days, main keyword at position 18. Verdict: conversions yes, ranking in the top 20, traffic low. Action: keep it and add internal links, one from the product page and one from the process page, then check next quarter whether UV can climb to 30.
Scenario 2: a lead-gen page on “2024 industry white paper analysis,” 2,800 monthly UV, main keyword at position 8, zero inquiries in 90 days. Verdict: traffic and ranking are good, conversions have collapsed. Action: leave the ranking alone and change the bottom CTA from “Download the white paper” to “Book a 30-minute technical call,” then follow Inquiry Form Field Design and Conversion Rate and cut required fields from 9 to 6.
A counterintuitive extension: an About page that only ranks for brand terms, gets 300 monthly UV and brings in 8 inquiries should not be cut for “not earning organic traffic.” In a B2B buying cycle, customers always return to the About page after reading every product page. It is the trust closer, see Trust pages and AI credibility.
5. Quarterly audit cadence and regression
Keep each audit under 150 pages. If you have more, split them into batches by template and run product pages, blog posts, and whitepaper download pages as separate rounds, so decision fatigue doesn’t make you rush the last 50 pages.
| Steps | Action | Output | Timeframe |
|---|---|---|---|
| 1 Pull data | Export 90 days of data from GA4, GSC, and your CRM | One row per page, three dimensions | Week 1 |
| 2. Score | Assign an action label to each page using the matrix | Six page-type lists | Week 1 |
| 3. Execute | Update, merge, 301, revise CTA | Log of changes pushed live | Weeks 2 to 3 |
| 4. Submit | Request reindexing in GSC | Indexing queue cleared to zero | Week 4 |
| 5. Review | Revisit rankings and inquiries next quarter | Before-and-after action comparison table | After 90 days |
An audit isn’t just a quarterly task. On-site mining on-site search terms, questions that keep coming up in inquiries, and Competitor content gaps analytics are all everyday sources of signal. The quarterly audit is the window where you deal with these signals in one pass, not the only action you take all year.
6. How to archive and revisit action records
For the four types of changes each quarterly audit produces, updates, merges, 301s, and CTA revisions, archive them row by row in a shared spreadsheet. At minimum, include the original URL, the action taken, who executed it, the go-live date, the GSC indexing status two weeks later, and the change in inquiries after 90 days. This table isn’t for keeping a record after the fact, it’s the baseline you compare against next quarter. Without it, you can’t tell whether last time’s title change actually lifted CTR or just caught a natural swing in industry traffic.
One field that’s easy to miss when archiving is “basis for the decision.” Why did you fold this page into its sibling page at the time, why did you keep that zero-traffic page? Paste a screenshot of that row of three-dimension data into the table. Six months later, many decisions that seemed obvious at the time have no recoverable rationale once they’re separated from the original data, and the next time you hit a similar page you’re back to guessing.
FAQ
How often should you run a content audit?
Below what monthly UV can a page be deleted outright?
After merging two articles, what do you do with the old URL?
Why does the average position in GSC fluctuate so much?
Should a B2B website delete all its zero-traffic pages?
How do you decide whether a page needs a refresh or a rewrite?
Final thoughts
The value of a content audit isn’t in deleting pages, it’s in reallocating effort. Of 150 pages, maybe only 30 actually need changes; the other 120 should either be kept or left alone. Turn the quarterly audit into a three-dimensional table, and decisions stop relying on gut feel.
Three steps to put it into practice: this week, export the last 90 days of data from GA4, GSC, and your CRM and merge it into one row per page; next week, tag every page with a disposition based on the matrix, starting with the “declining + top 20” refresh tasks; on the first day of next quarter, review the rankings and inquiries for the pages you acted on to confirm the decisions were right.
One line worth remembering: traffic decides whether a page is worth looking at, ranking decides whether it can be found, and conversion decides whether it’s worth keeping. Cross all three dimensions, and you’ll know whether a page stays or goes.