Online Growth Spurt Is Not Business as Usual

Computerworld - E-business is booming, and online marketing is back. U.S. online retail sales are expected to reach $65 billion this year and will continue to grow by a compound annual growth rate of 17% through 2008 to top $117 billion, according to a report issued by Jupiter Research. Meanwhile, after taking a beating for much of 2001 and 2002, online advertising made an amazing comeback in 2003, fueled in part by the popularity of search, which accounted for almost 31% of all online activity in the first half of 2003, according to a joint report from the Internet Advertising Bureau (IAB) and PricewaterhouseCoopers. Piper Jaffray & Co. is predicting 20% compound growth for the next five years in online spending and calling for 21% growth in 2004, rising from $6.7 billion to $8.1 billion.
But other factors beyond the popularity of search engines are contributing to online marketing's resurgence, including the continuing fragmentation of audiences and media, a growing dissatisfaction with the accuracy of traditional bastions of marketing research, and the pressure on agencies and holding companies for more accountability for their marketing dollars.
This rebirth isn't accompanied by the click-based metrics of online marketing's heyday, before the dot-com bubble burst. Instead, a long list of new targeting and tracking capabilities is attracting marketers dissatisfied with traditional methods of measuring audience behavior. For example, advertisers now have the ability to optimize their campaigns by building sophisticated user-based response models that come very close to mimicking what database and CRM experts have been doing off-line.
Campaign improvements are no longer achieved by simply rotating out poor-performing media placements. Instead, marketers are able to build sophisticated, predictive models measuring variables such as ad placement, ZIP code and PRIZM data, exposure count, bandwidth, and viewers coming from work or home. These new tools can automatically improve targeting models on the fly. This approach to optimizing campaigns is like night and day for heavy data users accustomed to optimizing on only two variables.
Consider, for example, that performance can improve by 20% to 30% just by understanding the subtle differences between targeting by time of day rather than by day of week. Combining deeper analysis of multiple variables with more effective graphic ad formats provides online advertisers with a truly efficient, metric-driven, ROI-based capability to combat the diffusion of their audiences.

Or take the case of pop-up ads. They proved very effective in 2003, but even though eMarketer reports that pop-ups represented less than 4% of the overall ad footprint in 2003, they seem to