Indeed, ads running on Web sites with related content were 61% more likely to be recalled than ads running on sites with unrelated content, according to a recent study conducted in collaboration with CBS Vision using McPheters & Co.'s AdWorks methodology.

Recall of particular ads varied by the type of sites on which they were viewed. Social networks, along with shopping and food sites, generated the highest recall levels of 29% to 39%.

"Targeting by site yields important benefits for advertisers," said Scott McDonald, Condé Nast SVP of research.

Search and portal sites, meanwhile, generated the lowest recall levels, according to the study.

Of note, there were large differences in recall by type of product advertised. "While we have long known that context is important for print advertisers, we welcome proof that the same is true online," said Drew Schutte, SVP and chief revenue officer for Condé Nast Digital.

"These results reinforce the importance of a marketer being associated with category-specific Web sites with established brands," Schutte added.

In the analysis, each of some 400 ads for which recall was measured was associated with the Web sites in which they appeared. Ads were segmented by whether they appeared on Web sites with related content. Recall of ads was measured among Internet users who were directed to surf the Internet at will for 30 minutes.

According to data released earlier in the year by Condé Nast and McPheters & Co., nearly two-thirds -- 63% -- of banner ads were not seen by Web users. Respondents' eyes "passed over" 37% of the Internet ads and "stopped" on slightly less than a third, McPheters found.

In contrast to online ads, TV and magazine ads generated a strong propensity to be seen and recalled, according to the research.

Full-page, four-color magazine ads were determined to have 83% of the value of a 30-second television commercial, while a typical Internet banner ad has 16% of the value.

This news story is important. I recommend that everyone read it.