Geier recalls how Laura Christine, vice president of direct marketing and e-commerce at shoemaker Skechers USA, told attendees at the Internet Retailer 2009 conference in June that she put 94% of the company's entire online budget into retargeting. Christine reduced Skechers' partnerships to seven ad networks from 20. While Geier couldn't recall her exact return on investment, he says the ROI exceeded 200% on more than $2 million in revenue.
Supporting that trend, Didit has released its behavioral retargeting platform, Blizzard, which determines in real time the price for an ad it serves up on any site connected across multiple ad exchanges.
Blizzard connects to Google's DoubleClick, Yahoo's Right Media Exchange, MSN's adECN, and AdNexus through application programming interfaces, giving Didit reach across 92% of the Internet, according to Geier. Among other attributes, the retargeting is based on how many times users come back to a site and what type of items they have in the shopping cart.
The six unidentified companies that began testing Didit's ad retargeting platform several months ago have renewed and increased budgets coming out of the initial trial program, Geier says. The next step will be to retarget display ads in social networks, such as MySpace and Facebook.
Didit CEO Kevin Lee says the way marketers purchase ads is changing. It's only recently that marketers began buying impressions one at a time. Some ad exchanges now allow marketers to change bids in real time, similar to the way people bid for clicks.
The glut of banner inventory and the current economic climate are fueling the change. "You can now specify that you're interested in the phrase 'Caribbean cruise vacation' and pay for the L.A. region in a specific time zone," he says. "It doesn't work well for everyone, but the overall incremental gains clients can achieve are significant."
Buying a list of blanketed travel terms just doesn't work, according to Lee. Marketers can't target ads with precision using a large bucket of generic terms. Knowing the consumer typed in "Caribbean cruise" -- though there are likely fewer people who do, rather than those who type in just "cruise" -- would be much more on target.
When it comes to remnant targeting, often not enough information comes from ad networks or ad exchanges to make a good decision on the relevant advertisements to serve up, Lee says. The consumer ends up seeing too many University of Phoenix ads.
Think about search behavior. It's one of the most important triggers because it's specific. Tapping into someone immediate search behavior can produce better results when it comes to targeted ads, Lee says.
I agree with this theory when it comes to searching for low-cost items. I often make a split-second decision to buy a book, a shirt, a pair of shoes or download a piece of music from Apple's iTunes. But when I consider buying a high-ticket item, such as a car or flight to Italy, I will continue to look at related ads for months. And those ads will look for me, too.