Content Performance Metrics: Which Numbers Show a Page Is Working, by the Job It Was Written For
Pick the metric from the job the page does, and a monthly report turns into a list of decisions.
Published by AI Knows Us · Updated 12 October 2026
The content performance metrics worth tracking are the ones that match the job of each page: impressions and clicks for pages meant to be found, engaged sessions and scroll depth for pages meant to be read, and key events such as form submits or purchases for pages meant to convert. A single page should carry one primary metric and two supporting ones. Anything beyond that is decoration.
Start with the job, not the dashboard
Most content reports fail because every page is judged by the same number, usually pageviews. A pricing page and a glossary entry do different work. One is read by a buyer close to a decision and the other brings in people who have only just started to look. Judging them by one yardstick makes the glossary look strong and the pricing page look weak, which is backwards.
Give every page a single sentence of purpose before you open an analytics tool: "bring in searchers who do not know us", "answer a question a prospect asks on a sales call", "move a visitor to book a demo". The metric follows from the sentence.
The five jobs and the numbers that belong to each
| Job of the page | Primary metric | Supporting metrics | Where to read it |
|---|---|---|---|
| Be found in search | Clicks from search | Impressions, average position, click through rate for the main query | Google Search Console performance report |
| Be read | Engagement rate | Average engagement time, scroll depth to the end of the key section | Google Analytics 4 |
| Move the reader on | Click through to the next page | Pages per engaged session, internal link clicks | Analytics events |
| Convert | Key events (form, call, purchase) | Conversion rate for the page, assisted conversions | Analytics plus CRM |
| Earn references | Referring domains to the page | Mentions, citations in other content | Search Console Links report, backlink tool |
What each metric actually tells you
Search metrics
Google documents its Search Console performance report as the source for clicks, impressions, click through rate and average position (Search Console Help, read 12 October 2026). Use it per page and per query. A page with rising impressions and flat clicks has a snippet problem: the title or description is not winning the click. A page with falling impressions has a ranking or demand problem. These are different fixes, and the report separates them for you.
Engagement metrics
Google Analytics 4 counts a session as engaged when it lasts longer than 10 seconds, has a key event, or has two or more page or screen views. Engagement rate is the share of sessions that were engaged, and bounce rate is its opposite (Analytics Help, read 12 October 2026). Compare a page against other pages of the same type on your site. An engagement rate means little in isolation and a great deal next to its neighbours.
Conversion metrics
Count the action a page exists for, and count it per page. Site wide conversion rate hides which content does the work. Record the page where a visitor first arrived as well as the page where they converted, because a long guide often starts a journey that ends on a pricing page.
Reference metrics
Referring domains tell you whether other sites find the page worth citing. Count distinct domains, not total links, since fifty links from one site are one vote.
Metrics that look good and decide nothing
- Raw pageviews. They rise with a viral post that brings nobody useful. Read them alongside engagement and conversions.
- Time on page, alone. A reader who finds the answer in ten seconds is a success. Judge it against what the page should take to read.
- Social shares. Useful as a signal of reach. They rarely connect to revenue.
- Word count. It describes the input. Use it for planning and leave it out of the performance report.
A monthly review that fits on one page
- List your top twenty pages by the metric that matches their job, not by traffic.
- Mark each as rising, flat or falling against the previous three months.
- For each falling page, name the cause from the table above: snippet, ranking, engagement or conversion.
- Choose three pages to change this month. Write down what you changed and the date, so next month's numbers have an explanation.
- Choose one page that is working and link to it from three other pages. Success deserves support.
This sequence is the same one that sits behind a good SEO scorecard. Our page on SEO KPIs shows how to separate brand from non brand traffic, which makes every number above cleaner. If your funnel has several stages, sales funnel content sets out the job for each stage. For the changes themselves, see the repeatable optimisation pass.
Questions about measuring content performance
How many metrics should a content report have?
Three per page type is enough: one primary and two supporting. A report with thirty numbers gets read as noise and produces no action.
How long before a new page shows results?
Check indexing in the first week, impressions after a few weeks and clicks after that. Use the first month to confirm the page is indexed and appearing for its intended query, and judge conversions after you have a few hundred visits to learn from.
Should I measure blog posts and landing pages the same way?
No. Blog posts usually carry search and reading metrics. Landing pages carry conversion metrics. Both can share a reporting template, with the primary metric column changing.
Is engagement rate better than bounce rate?
They are two views of the same data, since bounce rate is the share of sessions that were not engaged. GA4 reports both. Use engagement rate because it names the thing you want.
How do I tell which content leads to sales?
Send key events to your CRM with the first landing page recorded, then group closed deals by that landing page. The result shows which content starts the deals that close.
Measuring how an assistant describes your content
Every metric above counts what happens on your site. A growing share of buyers read the answer to their question in ChatGPT, Gemini or Perplexity and never open a page, and those tools report only what happens on your own site. The page you wrote to answer a question is now also being asked of an assistant, and what it says back is a separate thing from the page. AI Knows Us (aiknowsus.com) was built for that reading: it asks the assistants the questions your content targets and records whether your business or your pages are named, so the content scorecard gains a column for the place where the answer is now given.