How to succeed in advertising, redux

Six months ago, I penned an article called "How to succeed in advertising" in which I argued that the Internet was going to have a major impact on how advertisements are bought and sold in the future. I argued that the Internet would lead to a shift toward performance-based advertising, as each advertiser could measure the actual performance of every ad.

On the Internet, advertisers can measure not only click-through ratios (the number of people who click on a given ad, divided by the number of impressions), but also sell-through numbers (actual customers, divided by the number of impressions). The supposition was that, if advertisers could measure exactly the response to their ads, why would they settle for underperformance?

This article generated a completely bipolar response. The "content" providers lashed out with a fiery tongue. The standard response, elegant in its simplicity, was: "You couldn't be more wrong." Those in this camp were quick to point out that brand or image advertising could not be measured using such overly analytical direct-marketing techniques. Furthermore, they insisted, the public needs both Seinfeld and the Super Bowl, and it's brand advertising that allows them to exist.

At the other pole, the people who pay for ads bombarded me with praise. "We knew this day would come," they chortled. "Why should we pay for something that doesn't work?" Ad buyers are much more supportive of increased scrutiny than ad sellers.

Looking toward the future, my analytical mind favors the argument of the buyers, and not simply because they control the dollars (or because they were more supportive of my viewpoint), but rather because the continued evolution of technology puts tremendous pressure on traditional impression-based advertising.

Internet ad buyers are learning at a rapid pace. Six months ago, they began cross-checking their ad expenditures by analyzing simple click-through ratios. During renegotiations with ad-based sites, they would point out that dollars per click-through were what mattered, not dollars per impression. Some sites might have many impressions, they argued, but few viewers who respond to ads.

On the other hand, some sites with specific context have much higher click-through rates. Ad buyers call these sites "performers," and any smart ad buyer can recite their performers and nonperformers by heart.

Of course, being the adaptive creatures that they are, many content providers tried to game the system, artificially inflating their impressions using technology that "spidered" the sites at night. As the game shifted to click-throughs, the mythical consumers began to click on the ads.

However, ad buyers continue to evolve as well, and now are looking beyond the click-though and analyzing the lifetime behavior of customers that click over from a particular site. These e-merchants now can tell you not only which advertisers produce click-throughs, but also whether those customers actually follow through and buy, how much they buy, and what the margins are on their purchases. This is surely information ad sales representatives never wanted to see revealed.

Many Silicon Valley companies are jumping in to help arm ad buyers with detailed analytical software. Companies such as AdKnowledge, Personify, and Epiphany are developing market analysis packages that help marketers justify their ad expenditures across the Internet and other media. This technology will result in much more rational behavior on the part of ad buyers. Privacy advocates and ad sales reps will be equally displeased with the depth of information in these customer-by-customer reports.