Keyword stuffing is a practice in which webmasters load their pages with instances of commonly searched for keywords. The practice is designed to direct traffic to the site by raising the site's ranking in search results, in the hopes of increasing site income. However, keyword stuffing can be an unwise choice; many search engines penalize for keyword stuffing, and they use complex algorithms to detect signs of keyword stuffing, even when it is cleverly done.
The most obvious instance of this practice would be inserting the keyword into a block of text. Often, the author takes the time to make the text look reasonably plausible, in the hopes of avoiding penalization by a search engine. For example, if the webmaster knew that the keyword “wicked goat” was commonly searched for, he or she would try to load it onto a page as many times as possible.
More commonly, keyword stuffing is accomplished by hiding the keyword, so visitors don't see it, but search engines do. Keywords can be hidden in meta tags, and they can also be embedded on a page, in text which is the same color as the background so that it is invisible to the user, or in more crafty ways. In many instances, the keywords are not even related to the content on the site; they are just designed to elevate search rankings so that users will visit, generating advertising revenue. While some amount of keyword stuffing is considered legitimate and even smart Search Engine Optimization (SEO), webmasters who go overboard give the practice a bad name.
Before search engines wised up to the practice, they would return stuffed sites when users searched for common keywords. Users would assume that the site was relevant, since it was in the first page of search results, and they would click on the link. At best, the site might be a bunch of nonsense with interspersed ads; at worst, it might load harmful software on the user's computer. The practice could be extremely frustrating for people trying to browse the web for information, and as a result, search engines refined their algorithms to use a number of markers when determining how relevant a site is.
Most search engines will flag sites with an unusually high number of keywords, terming them spam, rather than legitimate websites, leading to the alternate slang term “spamdexing” to refer to keyword stuffing. In the instance of “wicked goat” above, there are really only so many times that such a phrase could be mentioned in a legitimate web article; if a search engine found 40 instances of the term on a single page, it might assume that the page had been stuffed.
In a variation on keyword stuffing, some webmasters put the same site up at multiple domains. While mirroring a site is a perfectly reasonable way to handle high volumes of web traffic, 20 sites advertising sports equipment in exactly the same text are not necessarily legitimate. Webmasters can also use these sites to create links, which can also raise search engine rankings. As a result of this practice, many search engines look at the source of links to a site to determine whether they are valid, or merely spamming tools.