This one is a little rough. A quick summary is that the remnant business has changed. There is as much if not more available today than at any other time, but it's also become harder to grow a company simply buy focusing on remnant.
The two largest ad networks today all owe their business to remnant inventory. Each took the approach, without calling it this then, of aggregating the long tail of the impressions. They focused on two types of sites – smaller sites that either could not or did not have a sales force, and larger sites that had a sales force but still had inventory they could not sell. There was a lot of it, and for companies starting out without brand dollars, focusing on aggregating all of the unwanted impressions seemed a logical way to capture business in an environment where many competed for the few top sites. For the early remnant players, success didn’t necessarily mean aggregating the greatest number of sites, just enough impressions.
If a company could aggregate a lot of unsold impressions, it could reach some economies of scale. And, if it had some level of optimization, it could in turn make enough money off the unsold impressions to continue having access to those impressions. You might think a company would want to show a potentially monetizing ad over nothing or a house ad, but some sites, especially in 1999 and 2000 when dumb money flowed freely, thought associating lower paying ads would detract from their ability to uphold premium prices. As the dollars dried up in 2001, these concerns no longer seemed to matter, and once picky sites all of the sudden started to take calls from the remnant players who, up until then, had occupied, and not unhappily, the role of second tier provider.
The growth of online impressions and retraction of money allowed the remnant players to absorb greater amounts of inventory. They moved up market in the sites they served and clients who purchased. Consequently, sites moved down market, often choosing to focus on traffic and not hire an ad sales staff. In other words, sites started to seek out the ad networks and this increase in demand allowed advertisers to begin testing CPA advertisers, all while networks continued to operate in a primarily blind status, i.e. advertisers could not choose the sites and sites had limited control over which advertisers. Ad networks were, during this time, often servicing much more than simply the remnant.
Ad networks, especially those with their origins in remnant advertisers, have limitations in the number of advertisers they can work with in scale. This has to do with ad networks often using performance data to dictate how and which ads should run on the sites in their network. Google flipped this approach by coming to market with a larger advertiser base, and rather than using performance data to determine what to show where, they used context to self-select a group of advertisers to try. The search giant developed an immense network of sites, but interestingly, there were that many impressions not being serviced to their capacity. This shortfall allowed them to enter the space and grow, but also left space for the existing networks to remain in business.
Google’s entrance did mean that the site mix in the networks shifted, and many networks started dealing with primarily larger sites, the small, very targeted ones going to Google. Many ad networks were back to focusing on remnant as their growth strategy. And, growing a successful remnant ad network today is not as easy as it was four years ago. As mentioned earlier, two factors tend to make remnant inventory a challenge to monetize well. It tends to include mainly untargeted impressions or impressions from sites whose content does not map to a high value advertiser interest, and equally as important, the users have already had a significant amount of exposure to other ads before being shown these ads. A user seeing an ad impression for the first time is much more valuable than one seeing their hundredth ad impression.
With little context and variable quality, just rotating ads will not necessarily earn money. It also explains why lead generators who can take a microscopic focus on a particular section can eek out some money, and, why trying to rotate in a variety of ads without having designed the banners or landing pages can be challenging. If it were that easy, every affiliate network would also be an ad network. And here, too, is where behavioral comes back into play; it can add a necessary layer of intelligence to the untargeted impression to increase the likelihood for click and/or conversion.
The older ad networks still exist because they have reached the economies of scale in pricing/performance where they can leverage higher quality remnant sites (better demographics, few ad impressions per person) to float the lower performing ones. So, while display advertising has definitely started to lust yet again for the brand dollars, the immense functionality of the web (email, MySpace, YouTube, etc.) means that remnant ad impressions will always exist. Providing a remnant solution is a big opportunity, but don’t assume it will be easy.
