Page View

Every time you open a web page, your computer sends a request to the server where the page is stored. That server delivers the page back and logs the event. This single event is exactly what’s called a page view. If you open the homepage of a news site and then three articles, that’s four page views from one person. So a page view doesn’t count people, but loaded pages. That’s the most important distinction to keep in mind.

Why publishers stare at their view counts

Most news sites make their money through advertising. Ad space is paid for based on visual contacts, and every page view delivers such contacts. More views therefore directly mean more revenue. Typical prices are a few euros per thousand delivered ad impressions. A site therefore needs millions of views per month to be able to live off it.

That’s why view counts also show up in business reports and stock market announcements. If the number grows, a media company is considered healthy. If it falls, the share price often drops too. For investors, the page view is one of the few easily comparable metrics on the web.

But the number has a downside: it says nothing about quality. Someone who closes a page in annoyance after two seconds generates the same view as someone who reads for ten minutes. That’s exactly what makes sensational headlines so tempting. Experts therefore also look at dwell time, meaning how long visitors actually stay on the page.

How counting works and where errors creep in

Technically, counting usually happens via a tiny piece of code embedded in the page. It reports to the measurement system that the page has been loaded. The Google Analytics service is well known for this. Alternatively, one can analyze the server’s log files, meaning the list of all requests that were answered.

Both methods count differently, and neither is perfect. Ad blockers in the browser can disable the tracking code, causing the view to be missing from the statistics. Server logs, on the other hand, count too much: automated programs such as search engine crawlers also fetch pages, even though no human is reading along. Such programs crawl the web on their own and make up a significant share of traffic on some sites.

Related metrics narrow down the picture. The visit, in English session, groups together all views by one person within a single session. The unique visitor counts each person only once per time period, no matter how many pages they open. A website with a million views therefore has significantly fewer visitors than views. Anyone comparing figures from two providers must always check which metric is meant.

From blog counters to AI chatbots

View counts show up everywhere content is published. Video platforms display them openly under every post. Anyone running their own blog or online shop finds the number in their site’s admin dashboard. Wikipedia also publishes how often each article was accessed.

Currently, this metric is under pressure, and that has to do with AI. Chatbots and AI answers in search engines summarize content directly. The user gets their answer without ever clicking through to the source. Many publishers therefore report significantly declining view counts from search. This is one of the reasons why media companies are suing AI providers or signing licensing agreements with them.

For you as a reader, this mainly means one thing: if a site advertises a view count, that’s a reach figure, not a quality judgment. And when a report mentions a drop in page views, it’s almost always about money.