Content Optimizer

Summary#

Content Optimizer takes a live URL and a target keyword, scores the page out of 100 against the pages currently ranking for that keyword, and returns a prioritized list of edits. Its internal tool id is content_optimizer and it costs 1 credit.

Purpose#

You have already written the page. It is published, indexed and not ranking. The useful question is not "is this good content?" but "what do the pages beating me have that this one does not?"

This tool answers that in numbers. The decision it supports is what to change first on a page that already exists — and whether the gap is worth closing at all.

Overview#

You supply a page URL and the keyword you want it to rank for, separated by a vertical bar. Metric Vault then:

• reads the live Google results for that keyword to a depth of 30 and keeps the real top ten, with their titles and descriptions,
• pulls related keyword suggestions with their volume and competition,
• measures the keyword's real difficulty,
• and fetches your page itself for its word count and heading structure.

Your page is then scored against that benchmark across ten factors, and the gaps are turned into specific, ordered edits.

The vertical bar matters. With example.com/post | seo tips the tool knows which part is the page and which is the keyword. Given only a URL, it derives a keyword from the domain name, which is rarely what you meant.

Benefits#

• The benchmark is the live top ten for your keyword, not a generic checklist.
• Ten subscores tell you which dimension is weak, so the fix is specific.
• Current vs Recommended states the gap as a number you can close.
• Title and meta description suggestions come with the report.
• 1 credit per run keeps it below the monthly premium threshold.

Use Cases#

Reviving a page that stalled. It ranked on page two and stopped. The subscores usually show which single dimension is dragging.

Checking a draft before publishing. Publish, run the optimizer, fix, and you start from a stronger position than a blind launch.

Settling an editorial argument. "It needs to be longer" becomes a measured word-count gap or it goes away.

Auditing a competitor's page. Any public URL works, so you can score theirs against the same keyword and see why it wins.

Closing out a content project. Score every page produced by AI Article Generator against its brief's keyword.

Requirements#

• A signed-in account.
• A paid plan — 1 credit per run, and any cost above zero is blocked on Free.
• A page that is publicly reachable. Pages behind a login or a firewall cannot be fetched.
• A target keyword. Without one the score has no benchmark.

Permissions#

Any workspace member on a paid plan can run it. Free accounts are refused with This tool needs a paid plan. Free includes the 10 technical SEO tools; upgrade to Pro to unlock the rest. Signed-out visitors see Please sign in to run this tool. The recommendations panel below the result needs Pro or above.

Cost#

1 credit per run. Below the premium threshold of 3, so the monthly quota never blocks it — only the hourly fair-use cap of 100 light-tool calls per hour.

This tool has the shortest shared cache in the product: 12 hours, because the page being scored is usually your own and changes often. A cache hit inside that window still charges the credit. If you have scored the same URL and keyword before, the You already ran this dialog offers Open saved result, which is free. See How credits work and Result caching and freshness.

Navigation Path#

Dashboard → Write & Optimize → Content Optimizer

Inputs#

Step-by-Step Guide#

1. Open Write & Optimize and choose Content Optimizer.
2. Read the three steps on the card: paste URL and target keyword, compare against the real top-10, get fixes ranked by impact.
3. Type your page and keyword into the field, separated by a bar — for example example.com/blog/seo-basics | seo basics.
4. Click Optimize Page · 1 report, or press Enter.
5. Wait for Analyzing… this can take a few seconds for live data. to clear.
6. Read Overall Score, then go straight to Prioritized Fixes.
7. Make the edits, wait for the page to be re-crawled, and re-run to confirm the score moved.

Reading the Results#

The header. Your URL, a Content Optimizer verified badge, the target keyword under Target:, and Scored just now.

The KPI tiles. Up to four, drawn from what the page actually returned.

The rank-prediction banner. A single sentence estimating how the page is likely to place if nothing changes. It is colored by your overall score: green at 80 or above, amber from 60, red below. Treat it as a summary of the score, not as a forecast.

Optimization Subscores — N factors. The most diagnostic block on the page. Each factor is a 0-100 meter, green at 80 or above, amber from 60, red below:

Work on the two lowest meters. Moving one factor from 30 to 80 shifts the overall score more than polishing three that are already green.

Current vs Recommended — close the gap. A four-column table — Metric, Current, Recommended, Gap — covering word count, keyword occurrences, headings, images, internal links, external links and schema blocks. The gap cell shows on target when the two match, a green negative when you are above the recommendation, and an amber +N when you are short. This is the checklist to hand to whoever edits the page.

Top-10 Benchmark. What the ranking pages actually do — the averages your recommendations were derived from. Use it to sanity-check a recommendation that looks extreme.

Prioritized Fixes. The action list, each with an Impact: line and a priority. High priority renders red, medium amber, low green — the color marks urgency, not quality. Work top down; the order already accounts for effort.

Missing Semantic Keywords. Terms the ranking pages use that your page does not. Cover them as concepts inside real sentences. Pasting them in as a list will not move the semantic-breadth score in a way that survives review.

Recommended Headings. The H2 structure suggested for this keyword. Compare with what you have; the difference is usually where the structure subscore went.

Title Tag Suggestions and Meta Description Suggestions. Rewrites for the two elements that decide clicks. For a full set of scored variants with a SERP preview, use Meta Generator.

Readability Issues. Specific passages or patterns causing the readability score, each with a severity.

Schema Recommendations. The structured-data types this page should carry. Validate whatever you add with Schema Tool.

Internal Link Suggestions. Where to link from and to inside your own site.

What to do with it. Take the top three Prioritized Fixes, apply the Current vs Recommended numbers, add the missing semantic keywords in real prose, and re-run after the page has been re-crawled. Anything below 60 usually needs a rewrite rather than an edit.

Examples#

Screenshots#

Tips#

• Always include the bar and the keyword. It is the single biggest difference between a useful score and a meaningless one.
• Score the exact page, not the section index. A category URL will be benchmarked against article pages and always score badly.
• Fix the lowest two subscores first, not the longest fix list.
• Re-run only after the page has been re-crawled. Scoring an hour after an edit usually returns the same numbers.
• Score a competitor's URL against the same keyword to see what a high score looks like in your niche.

Best Practices#

• Brief with Content Brief & Template before writing, score with this tool after publishing. The two use the same benchmark.
• Rewrite for the semantic terms rather than inserting them; the score follows coverage, not repetition.
• Keep a record of the score before and after each edit — Saved Work holds every run.
• Pair with On-Page SEO Checker for the technical side of the same page; this tool grades content, that one grades markup.
• Re-score important pages quarterly, and use Content Decay Detector to find which pages need it.

Common Mistakes#

• Entering a URL with no keyword. The tool then derives a keyword from the domain name and benchmarks you against the wrong SERP.
• Reading the overall score as a ranking prediction. It measures the content gap; links, authority and intent also decide position.
• Chasing word count alone. Length with no semantic breadth moves one meter and nothing else.
• Scoring a page behind a login. The fetch fails and the report has nothing to compare.
• Re-running inside twelve hours and paying a credit for a cached answer.

Limitations#

• The SERP, keyword suggestions and difficulty all come from Google's United States, English-language index. The country picker does not change the market.
• The live results are read to a depth of 30 and the top ten are used as the benchmark.
• The page fetch reads the served HTML, so content that only appears after JavaScript runs may not be counted.
• Related keywords are capped at 30, so the semantic list is a strong sample rather than the entire topic vocabulary.
• The score covers content. It does not weigh backlinks, domain authority, page speed or index status — use Domain Overview and PageSpeed & Vitals for those.
• Only one page is scored per run. There is no bulk mode for this tool.
• Recommendations are model-written on top of measured data; the numbers are real, the prose is guidance.

Troubleshooting#

FAQs#

What does the overall score actually measure? How closely your page matches what the pages currently ranking for your keyword do, across ten content factors. It is a content-gap score, not a ranking forecast.

Why do I need the vertical bar? It tells the tool which part of your input is the page and which is the keyword. Without it, the keyword is guessed from the domain name and the benchmark will be wrong.

Can I score a page that is not mine? Yes, as long as it is publicly reachable. Scoring a competitor against your target keyword is one of the fastest ways to see what a winning page looks like.

Why is this only 1 credit when Gap Finder is 3? Credits track backend cost. This tool makes fewer and cheaper provider calls than a two-domain keyword comparison. See How credits work.

How soon can I re-run after making changes? Wait until the page has been re-crawled and at least twelve hours have passed — inside that window the shared cache may return the same result while still charging a credit.

Does it rewrite the page for me? No. It returns scores, gaps, suggested headings, title and meta options and an ordered fix list. To produce new copy, use AI Article Generator.