Traffic

Traffic is the English word for movement or flow. On the internet, it refers to the stream of visitors and data arriving at an offering. For a news site, this usually means counting how many people visit it and how many pages they click on. For a network operator, it instead means counting the amount of data transported, i.e. the number of bits transferred. Both meanings fall under the same word, and context determines which one is meant. The comparison with a road works well: you can count cars or weigh the cargo being transported.

Why visitor numbers determine money

Very many websites make their money through advertising. Ad slots are paid for per thousand impressions. Anyone with twice as many visitors can therefore roughly sell twice as much advertising. That’s why traffic is a hard business metric for media companies, shops, and platforms, not a matter of vanity.

Conversely, traffic also costs money. Every visitor triggers computing work on servers and generates data transfer. Cloud providers bill exactly for that. A sudden surge, for example after a viral video, can drive up a bill sharply or even crash the site.

For AI companies, this topic has taken on new urgency. When chatbots answer questions directly, fewer people click through to the original sources. Publishers have reported significantly declining visitor numbers from search engines since 2024. The dispute over licensing agreements between publishers and AI companies is essentially about exactly this lost traffic.

How visitors are counted and data streams are measured

For websites, three metrics are common. Page views count every call of a page. Sessions bundle together everything a visitor does in one go. Unique visitors attempt to count individual people, no matter how often they come back. The three figures often differ by a multiple, which is why comparisons between companies without specifying the metric are worthless.

Measurement is usually done with a small piece of code in the page’s source code that reports every call to an analytics service. Alternatively, server log files are evaluated. Both methods have gaps: ad blockers suppress the tracking code, and a user with a phone and a laptop is easily counted twice.

A large portion of traffic also doesn’t come from humans. Search engines send out programs that automatically crawl pages, so-called crawlers. In addition, there are now collection programs from AI companies that scan texts to train their models. Reputable statistics filter out this automated share, because otherwise it inflates the numbers.

Traffic in quarterly figures and in your own browser

In the business reports of large platforms, traffic almost always appears, often as the number of monthly active users. If it rises, the stock price usually reacts positively. In reports about data centers, on the other hand, it’s about data volumes, given in terabits per second. In both cases, the same word stands for very different things.

In everyday life, you encounter traffic with your phone plan, where a monthly data allowance is agreed upon. Once it’s used up, the connection is throttled. The term search engine optimization also belongs here: it describes all attempts to rank higher on Google and thereby gain more visitors.

A common misconception is that more traffic automatically means more success. What matters is what the visitors do. Ten thousand people who click away after two seconds bring less than a thousand who buy something or sign up for a subscription. Experts therefore always look at the quality of the traffic, not just its size.