Delivering Personalized Display Ads
One personalization optimization firm which has spent much of the past decade helping e-marketers build recommendation engines to enhance the on-site consumer experience, believes the time is right to expand the of personalization process to off-site advertising creative and messaging.
"Most retailers by now have at least begun (and many have become quite advanced at) personalizing particular product recommendations and brand attributes for consumers who are on their site," says Cheryl Kellond, senior vice president of advertising at ChoiceStream.
"The problem," she adds, " is that that level of personalization is not followed through when it comes to delivering advertising messages to consumers once they've left the site, or to prospective customers. Retailers are well aware that long-term value and share of wallet depend on building strong relationships with customers over time. No way in the world would a savvy tech e-commerce site assume that every visitor just wanted to see recommendations for the same identical products. They know that within the "tech enthusiast" segment are a wide variety of shopping preferences, different product interests, different budgeting styles, and different priorities. Yet somehow there's a huge disconnect between the rich personalized experience retailers and publishers offer customers on-site and the ad messaging those customers see from the brand off-site."
Further, she says, "Advertising is actually one of the smartest places to put these personalized recommendations - because consumers are not just flocking to ecommerce sites any more - the retailers really need to go out and grab every potential in-market consumer and drive them to their site. Personalized display ads help build this relationship with the customer across multiple channels."
To close this gap ChoiceStream is applying the personalization technology they've applied for on-site content recommendations to off-site ad placement. They deploy cookies to learn about consumers based on the shopping-related choices they've made in the past such as purchases, commercial searches, cart abandonments and conversion clicks.
The premise, Kellond explains, is that the hardest piece, the one where most behavioral targeting continues to come up short, is the product recommendation side. "Once marketers move from optimizing their own sites to targeting ads they are still delivering static, somewhat generic ads based on proxy behaviors such as browsing,' she says. "The gap has been figuring out how to use the e-commerce data marketers have already collected to deliver truly personalized ads."
"You've had a situation,' she adds, " where advertisers are getting better and better at finding out whom they want to target but haven't improved their ability to customize the messaging to drive customers back to their site. Retargeting is one of the highest value places to start with these new personalized ads. By itself it is an effective mechanism to drive new sales, but with ads that contain personalized product recommendations, display ads start to deliver an ROI that looks more like search."
Where behaviorally targeted networks leverage page views, clicks and browsing data to generate arget broad-based audience categories and segments, truly personalized ads, Kellond says, must leverage e-commerce-specific transactional and search to predict what types of products and brand elements best fit not only the category interests but product preferences of consumers.
"Behavioral targeting online ads by consumer segment is the easy part," observes Kellond. "It's personalizing them so each consumer gets a unique ad that is meant for them that's the much harder challenge."
The platform enables an e-marketer, for instance, a consumer electronics retailer, to anonymously follow a shopper who searched for plasma TVs, the next time or number of times that shopper is on a site where that electronics store advertisers. The cookie recognizes the customer and a customized ad can be served showing some new products at the site that the recommendation engine thinks the consumer would like, based on a deep analysis of the consumer's shopping specific data and how they relate to buying preferences of people with similar buying patterns and brand affinities.
This, Kellond believes, promises a decisive step beyond both behavioral targeting and re-targeting.
"The more personally ads can be targeted, it gives advertisers a 'shelf life' in ad frequency they don't have with static ads,' she says. "With re-targeting you've had to concentrate on reaching people off-site in a matter of a few days after their last visit. With more varied creative and offer you can target people for far longer, thirty days or sixty days after they've been off-site."
If successful, she believes, it may also prompt advertisers to realize scaling efficiencies that search advertising, the fair-haired child of today's online world, cannot.
"Search, for all the tremendous insight it offers into intention, has a limited reach ultimately," she says. "As advertisers reach a certain threshold of spend, they see they've tapped out already on their ROI. That's because you really can't attain the right scale of reach and frequency in search."
Editor's Note: Session videos of MediaPost's recent OMMA Behavioral conference are now available for viewing at http://www.mediapost.com/events/videos/