In years past, Web builders came up with a list of keywords for a Web page and placed that list in the keyword meta tag. The keyword meta tag was designed to be a reference for indexing page content, and the early search engines relied on it heavily.

However, some Web builders began to abuse the system by stuffing the keyword meta tag with terms that had little or no relevance to the page content. The result was Web pages that got traffic by virtue of high rankings for common search terms, but those pages often had little or no content related to the keyword the visitor was searching for.

This pollution of the keyword meta tags eventually rendered them unreliable tools for indexing Web pages or ranking their relevance. As a result, all the major search engines now ignore the keyword meta tag completely or give its contents very low priority in their page-ranking algorithms.

The following list puts the possible keyword locations in approximate order from highest to lowest priority.

• Domain name
• Page title
• Headings (enclosed within h1, h2, h3 tags)
• Body text—The first 2 to 3 KB usually counts more than the rest of the text (if the search engine scans more than that).
• Meta tags—Keywords appearing in the description meta tag still seem to count, as do keywords in some of the Dublin Core meta tags. Some search engines don't ignore the keyword meta tag completely but, rather, discount it heavily—especially if the keyword doesn't appear elsewhere on the page.
• Links—Even keywords buried in the URL, name, or id attributes of a link count in page ranking
• Alt text—Keywords in alt text attributes count toward page ranking. This is yet another reason to take the time to create meaningful alt text for all images.

Keyword repletion is another important factor in page ranking. A page that contains multiple instances of the keyword will generally rank higher than a page on which the keyword appears only once. However, the interaction of keyword placement and repetition is one of the areas where it's almost impossible to second-guess the ranking algorithms.