What is content analytics?

Content analytics is web analytics pointed at a specific question: what is each piece of content doing for the site? Not how much traffic the site gets - which posts earn it, which pages hold attention once someone arrives, which ones lead readers toward whatever the site exists to produce, and which are quietly dead weight. The term gets dressed up by enterprise vendors into something involving AI and taxonomies, but for most sites it's a discipline, not a product: a handful of per-page numbers, read regularly, feeding decisions about what to write, fix or delete next.

It matters because content is where most small sites' effort goes, and effort without measurement drifts toward what's fun to write instead of what works. The publishing schedule gets treated as the strategy. Content analytics is the correction: it turns "we published twelve posts this quarter" into "three of them did something, and here's what the other nine taught us".

No single metric ranks your content. A post with huge entrances and ten seconds of engagement is a traffic magnet with a content problem. One with modest traffic and deep engagement plus conversions is quietly your best work. The portfolio view - all five numbers, per page, side by side - is where the reading gets interesting.

Every content question reduces to a handful of per-page metrics, each answering something different:

• Entrances - how many visits started on this page. This is the page's earning power: content mostly gets discovered, not navigated to, so entrances measure its pull.
• Engagement time - how long the page held attention, counted only while it was actually on screen in GA4's definition. The did-they-read number.
• Scroll depth - how far into the piece readers got, with the caveat that scrolling isn't reading. Best for long pieces, where the drop-off point is editorial feedback.
• Assisted conversions - did readers of this page go on to sign up, subscribe, or buy? Content rarely converts directly, so look at what its readers did next rather than demanding a same-session sale.
• Queries and clicks from search - what Google shows this page for and how often it earns the click. This one doesn't come from your analytics at all.

Content's contribution is the hardest to count, because it rarely gets the last click. A reader finds your guide from Google, leaves, comes back twice over three weeks, and signs up from a direct visit - and a last-click reading credits "direct" while the guide that did the persuading counts nothing. This is attribution's general problem landing hardest on content, and it's why content programs chronically look worse in dashboards than in reality.

You don't need enterprise attribution modeling to correct for it, just two habits. First, look at conversion paths that *touched* a piece of content rather than ones that ended there - most tools can show whether converters read a given page in an earlier session, and visitor profiles make that history readable per person once someone signs up. Second, instrument your own distribution: UTM tags on the newsletter and social links to your posts, so "this piece drove signups when we promoted it" is measurable instead of remembered. Neither habit is precise. Both beat crediting your best essay with zero.

Your analytics tool sees what happens after arrival. Search Console sees the step before: which searches your page appeared for, how often, at what position, and whether the searcher clicked. For content decisions that's often the more valuable half - impressions without clicks mean the title isn't earning its ranking, rankings at position 8 mean a refresh might pay disproportionately, and queries you rank for by accident are a list of posts you haven't written yet.

The Performance report filtered to a single page is the routine: its queries, their positions, the CTR on each. Ten minutes per important page, quarterly, and it's free. Most sites with a content strategy have never once done it - which is also why doing it is an edge.

One worked example of what falls out: a post sitting at position 8 for a query with real impressions is the classic striking-distance find - close enough that a content refresh and a sharper title can cross onto page one, where the clicks actually live. We run that exact play on this site's library, and it reliably beats writing something new for the same effort. The report hands you these candidates for free - the only cost is looking.

Measured honestly, content performance is brutally concentrated. Ahrefs' study of over 14 billion pages found roughly 96.6% of them get no traffic from Google at all - a figure with acknowledged bias in its sample, but the shape holds on nearly every individual site too: a small minority of pages earns the large majority of search visits, and the tail earns approximately nothing.

This is the operating fact the whole discipline is built on, not a counsel of despair. It means your site's performance is really the performance of its top pages, so protecting and improving those is worth more than producing average new ones. And it means a long tail of zero-traffic posts is normal, not shameful - the question is only whether each tail post has some other job (supporting internal links, serving existing customers, feeding another channel) or is just archived effort.

Content performance has a time dimension that dashboards hide. A post that ranks well doesn't keep the position by right - competitors publish newer treatments, search intent shifts, and the once-fresh piece slides a position a month until its traffic has halved without any single event to notice. The pattern is called content decay, and on sites more than a year old it's usually the largest invisible traffic loss - a slow drop diffused across dozens of pages.

Detection is simple once you look: in Search Console, compare each top page's clicks over the last three months against the same months last year. A page down 40% year over year while the site holds steady is decaying. Freshness isn't a universal ranking factor - Google applies it mostly where queries deserve it - but for time-sensitive topics, dated examples and stale numbers are exactly what a refresh fixes.

Everything content analytics finds funnels into three actions. Refresh: update the decaying page that used to perform - new data, current examples, a sharper title against today's competing results. Because the page has history and links, a refresh usually beats a new post on effort-for-impact, though results vary too much for anyone honest to quote you an average.

Prune: for pages with no traffic, no conversions and no supporting role, decide deliberately - improve, merge into a stronger sibling, or remove with a redirect. We've merged posts on this site when two of them split the same queries. Consolidation is analytics-driven editing, and it works. Promote: the post with deep engagement but thin traffic doesn't need rewriting, it needs distribution - internal links from your stronger pages, a newsletter feature, the channels its numbers say it deserves. Diagnosing which move each page needs is the actual skill, and the five numbers above are the diagnosis.

The recurring decision content analytics exists to inform: write something new, or work on what exists? The portfolio math above gives the default answer more often than writers like - a site with fifty posts and three performers usually gains more from lifting posts four through ten than from publishing post fifty-one. Existing pages have age, links and rankings to build on, while new pages start from the 96% that earns nothing and climb from there.

New content wins the hour in two situations: when keyword research shows a question your audience asks that no page of yours answers - a gap, not a preference - and when an existing page's problem is that it targets nothing anyone searches, which no amount of polishing fixes. We run this site's library on exactly that logic, and the pattern holds: the months that moved traffic were the ones that filled measured gaps, not the ones that published most.

One page per move per month sounds slow, and outpaces most content teams' actual improvement rate by a wide margin. The routine matters more than the volume.

The whole discipline as a routine:

• Pull the top-pages report by entrances and note anything new in the top ten, plus anything that fell out.
• Check engagement time on the top five - did any of them get worse after a change?
• In Search Console, scan year-over-year clicks for your five most important pages and flag decay.
• Pick one page per move: one to refresh, one to prune or merge, one to promote.
• Write three lines about what you did - next month's review starts by checking whether it worked.

The measurement stack for everything above is GA4 plus Search Console, both free, joined by the linking both tools offer. Dedicated content-analytics platforms exist above that - Parse.ly and Chartbeat serve newsrooms with real-time editorial dashboards and per-author reporting - and make sense when content is the business and a team stares at the numbers all day. Below the enterprise tier, simpler analytics tools cover the same five numbers with less ceremony: Bigdelta shows entrances, per-page engagement and scroll behavior on one screen, next to the conversions the content assisted.

Whatever the tool, the unit of analysis is the page and the habit is the review. Software shortens the pull, but it can't decide which post deserves the refresh.

Content analytics is per-page measurement plus a monthly habit: five numbers per piece, Search Console for the before-the-click half, honesty about concentration and decay, and every finding routed into refresh, prune or promote. Start smaller than feels serious - your ten most important pages, one review, three actions - and let the routine prove itself. The sites that win at content aren't the ones that publish the most. They're the ones that know which of their pages are working, and act on the ones that aren't.