The Science of (Web) Advertising

In the hit television series Mad Men (Mad being short for New York’s Madison Avenue, the historic epicenter of the advertising industry), the 1950s ad business is presented as a schmoozy affair. Buying and selling decisions were often made over cocktails at a favored watering hole with a preferred sales rep.

The Internet promised to impose scientific rigor on this clubby business. The effectiveness of advertising in traditional media was difficult – often impossible – to measure. But the Web could track consumer response, at least in theory.

Frank Addante, founder and chief executive of the Rubicon Project, a young and aggressive Los Angeles startup, says, to date, the Web’s potential as an advertising medium is far greater than what it’s achieved. Web advertising accounted for $27 billion in 2007, according to Forrester Research. A staggering figure, Addante points out, given that it’s only 7 percent of the total dollars spent on radio, TV and print ads.  Yet, 33% of consumer time is spent online.

Surprisingly, despite the massive online advertising spend and the incredible amount of inventory, 80 percent of ad space on the Web still goes unsold.  Websites turn to ad networks to monetize unsold ad space. Ad networks act as middlemen between advertisers and websites.  As the number of ad networks proliferates exponentially – today, there are over 300 ad networks; seven years ago, only 15 – it’s difficult for websites to identify and manage relationships with the networks that will bring them the most revenue. 

Ad networks come in all shapes and sizes: international networks; local networks; price-based networks such as CPM, CPC and CPA; vertical networks targeting sports, travel, women and the gay audiences; format focused networks for text, display and now video; newly formed ad exchanges; and big media companies are now launching ad networks.  Ad networks are now in fashion, yet the lack of ad networks’ access to quality inventory combined with the deficiency of an advertising technology solution for publishers has rendered the majority of web ads largely ineffective, with little value placed on them. Consequently, Web ad space still goes unsold or for a fraction of the price of comparable ads that run in traditional media.

The task is especially difficult since “many Web publishers don’t have any idea which ad networks will best monetize which ads,” Addante says. “Although the data is available, it has not been made actionable. On the other side, ad networks have had very little visibility into the quality of the traffic on these sites. The dissatisfaction among both the ad networks and web publishers has resulted in a lot of market frustration and churn.”

Enter the Rubicon Project, an advertising technology company that optimizes ad networks to help websites make mad cash.  The company has attracted a massive network of websites (sellers) and ad networks (buyers), and applies sophisticated proprietary technology to match each ad to the best money-making opportunities. That technology, developed by an in-house team of advertising technology experts, exploits demographic, contextual and geographic factors to deliver ads that will perform best with a given publisher’s audience. And, it makes those matches in real-time.

Forrester Research forecasts the total number of dollars spent on Web advertising will nearly double in three years to $50 billion by 2011. Raj Kapoor, Mayfield Fund managing director, expects that the Web will take a larger slice of the total advertising pie due to the medium’s ability to track results. Mayfield recently led a $15 million Series B round of financing in the Rubicon Project, bringing the total capital raised to $21 million.

Demand is already proving high. In just five months, between October 2007 and March 2008, more than 4,000 Websites applied to join the Rubicon Project’s private Beta program, including some venerable names like America Online and eHarmony, as well as newer sites like Reunion.com, Slide.com, ZoomInfo.com and Friendster.com.

During that time, the Rubicon Project optimized more than 7 billion online ads across 60 of the top ad networks, in many cases delivering between 33 to 300 percent increase in advertising revenue for websites. The beauty of the Rubicon Project, Addante says, is it gets “smarter” with every ad it serves. Specifically speaking, that’s 7 billion pieces of market data, already collected, that the technology can decipher to predict future trends.  The company is changing the way advertising is bought and sold online to great effect. Today, it can place 5 billion ads per day – all without running an exorbitant bar tab.