When an end user clicks a link on a website they are viewing using their browser it will usually send a request to the server where the destination page is hosted. This request will include an HTTP header field usually referred to as the referrer field which tells the new server the last page the user was on. This page will naturally be the page on which the link was located.
This information is extremely useful to marketers as it will allow them to better understand the sources of their traffic and the profiles of websites that link to them. This will allow landing pages to be better optimised and future link building strategies to be adapted based on this information. It will also give some indication if the previous page was a search engine as to the search phrases used in the discovery of the site. The advent of secure searching on Google will reduce the availability of this referrer information over time.
It is not completely reliable information however, as it is possible to fake the referrer field or to block the provision of this information. Some users are keen to block this information as well as their true IP from web servers to avoid their browsing habits and user profile being logged (sometimes indefinitely) by hundreds of websites across the World Wide Web.