Revenue Science Tries Mobile Behavioral Targeting in Japan
The large number of mobile Web users may make Japan fertile ground to test behavioral targeting.
The large number of mobile Web users may make Japan fertile ground to test behavioral targeting.
New York-based behavioral targeting company Revenue Science is using Japan as the launching pad for advertising targeted based on mobile Web user behavior.
For the initiative, Revenue Science has partnered with Japanese Web portal operator mediba and the Digital Advertising Consortium (DAC). The company, which calls the endeavor an important step in moving behavioral targeting beyond the desktop to content on mobile devices, said mediba will create audience segments on the “au one” portal offered by Japanese cellular carrier KDDI.
The behaviorally targeted ads will be distributed when people use their mobile devices to access news and search results pages, and advertisers will be able to deliver relevant and timely ads to members of that audience segment, which will be defined by the device user’s online activity.
Details about just how this will be accomplished are being kept close to the vest by Revenue Science, and CEO Bill Gossman acknowledged some innovative techniques need to come into play, particularly since cookie-based behavioral targeting is not an option with cell phones.
“The notion of a cookie on a mobile phone doesn’t exist,” said Gossman. “Suffice to say, you can address behaviors coming off the mobile phone browser and coming off key strokes in the same way you can do it coming off a PC. It includes clicking and searching and consumption of various types of media… It’s not easy, but it’s possible.”
Revenue Science chose to try mobile behavioral targeting in Japan, rather than the U.S., because mobile communication is more advanced there, with high-speed, 3G systems the norm, and because the Japanese are significantly heavier users of the mobile Web than are Americans.
“If a consumer is frequently downloading a certain genre of music and using their phone to purchase a particular type of movie or video clip and they are searching for various terms, you start to build up a pretty good intuition of who this person is and what kind of consumer passion they have and how to fulfill them,” said Gossman. “It’s based on the handset, but the handset generally maps to a person.”
Neil Strother, a mobile analyst with Jupiter Research, acknowledged he was not familiar with the exact mechanics of the Revenue Science system, but he said BT in the cookie-less mobile world is possible “if you can quantify users by how they behave on their mobile phones.” Such information gathering is usually group-oriented, based on what a number of users have done with their devices.
“For instance, if you know that I go to CNN.com and Weather.com and ESPN.com and Myspace from my phone and [you know] where I live and you have access to some basic demographic data (age, income, education, race) about me, you could parse all this and send me an ad based on my past behavior… an ad for Pepsi or an ad for my local sports team,” said Strother. “It’s a lot of data mining, and — if done right — could be a holy grail in targeted advertising.”
He said this type of behavioral targeting could enable “an ad targeted to all the people who have searched for Pizza Hut within an area and of a certain age group,” or “You could target mobile subscribers who have accessed Best Buy’s site from their phone in the last week.”
Gossman noted Japanese companies are very interested in advancing behavioral targeting for mobile.
“We are working with mediba which is majority owned by KDDI, the second-largest carrier in Japan,” said Gossman. “They have very large penetration in Japan. And mediba’s two other large owners are the second and third largest ad agencies in Japan, so there is a fair amount of interest in solving this problem.”
Revenue Science is betting the venture has enormous potential for advertisers, as it cited the number of mobile phone subscriptions in Japan exceeded 100 million in January, covering 78 percent of the population.