A very small travel company offers tours to a Central American country that doesn't frequently enter American vacationers' consideration set. This country actually has fantastic scenery, great hotels, excellent food, and interesting people, plus it's extremely safe. But when people search for vacation destinations, they almost never type this country's name into the search box. SEM (define) is practically useless for this company (although when people do search on the country name, they tend to be quite qualified leads).

For the last two years, the company has gotten word out about itself and its tours with contextual advertising. There have been tons of articles written about this country. It's one of those "undiscovered gems" travel writers love. So when someone reads an article like "Inexpensive and Fabulous Tropical Destinations," the country is invariably mentioned. Luckily, a good number of sites use AdSense (the network the travel company owner uses), so the company's ad appears. People discover the country the same time as they discover the service.

Contextual works. It often works in ways search doesn't. The only issue is making sure both advertisers and publishers are given the controls they need to fit the tactic into their strategy. This is precisely what defines the second generation of contextual advertising networks.

For advertisers, control means the ability to have leads from contextual networks priced differently than those from search. When AdSense debuted, there was an immediate mini backlash: advertisers didn't want to pay the same amount for leads from two different sources, since those leads were of different intrinsic value. Jupiter Research's ongoing executive surveys of search marketers point to the frustration of having search and contextual prices tied to one another. Other networks, such as Yahoo's ContentMatch, made some early friends by offering separate pricing.

Naturally, there's opposing pressure from publishers and advertisers regarding pricing: within price boundaries, the advertiser tries to minimize spend while the publisher tries to maximize it. Assuming neither is trying to cheat the other, both are actually seeking that right-ad/right-person magic moment. This is a good thing; when two sides work cooperatively, the result tends to be better for both.

The challenge is providing enough appropriate controls so each party can operate most effectively. For advertisers, the ability to set their own price is a good thing. But if that price can only be set once for all members of the network, making an intelligent decision becomes difficult. In other words, the only intelligent, rational decision is to set the price down to what you would pay the least-valuable site in the network.

Google has a good solution for this in its Smart Pricing scheme. It dynamically sets prices based on the site. Yahoo may go farther, though, with its new contextual release: the price is set once, for everyone. But it works to ensure all sites in the network are of a high quality. The only problem is this may limit overall network size. So much for bragging rights.

The focus of control has been unfairly balanced toward advertisers. Both sides need the ability to optimize their portion of the program for the entire program to be effective. Fortunately for the future of contextual advertising, this trend is being corrected as the next generation of contextual networks come online.

Control for Publishers

Google really established the keyword-based contextual ad market with its AdSense product. When AdSense came out, the pitch seemed like a dream come true to many publishers: they would take previously low-value inventory, fill it full of relevant ads, and split the revenue. Some sites were cagey enough to ask for CPM (define) prices, but many simply took the money and ran.

Today, it's different. As Yahoo moves more aggressively into this market with its new Yahoo Publisher Network, it will find the market for new publishers is actually quite competitive. Publishers are now in a position to choose from several contextual providers. In fact, they don't have to choose just one. A publisher can agree to take ads from multiple companies, determining for himself which pages should be serviced by which networks.

The next round of competitive development, then, won't necessarily be on the revenue-share front (i.e., who gives away the biggest share). Instead, there'll be an increasing push to give publishers a finer level of control over the ads that appear on their pages.

Yahoo is kicking off this next generation with the single, most interesting (though overlooked) feature of its new network: the ability for a publisher to specify ad categories that may not be reflected in editorial. The publisher of a fly-fishing site, for example, may know his audience also loves high-end stereo equipment. He can request ads for high-end stereo equipment to appear, despite the fact the spiders will never, ever find any such content on his site.

It's entirely possible automated algorithms are better than publishers at figuring out the right ads to display. But the current challenge for the contextual providers is to build their relationships with publishers. The provision of control and the ability to optimize is clearly the next front. Expect to see much more development around this.