It is critical to observe that comScore's data highlighted the impressions by publisher. In regards to this format, it makes sense that Facebook leads the pack. However, the really interesting picture to look at would be the share of ad impressions by ad network/server. While Facebook now has a death-grip on the short tail of the market, the long tail is firmly held by competitors.
As an example, let's look at Google. In comScore's report, Google's sites comprised a very small portion of the impressions served, nearly a tenth of Facebook's impressions. But considering that Google doesn't really serve display ads on their own sites, this isn't particularly surprising. But Google does serve ads on other people's sites. In fact, looking at another piece of Comscore's data from the September Media Metrix report, Google's ad network is the third largest in reach, and reaches about twenty percent more of the online US audience than Facebook (~85 percent and ~65 percent respectively).
In fact, looking at Google's quarterly SEC filings, the ads served on the Google Network Sites account for as many ad dollars as are spent on Google's own sites, including search (only account for 30 percent of revenue, but 51 percent of search revenue and 68 percent of display ad revenue is paid to the publisher). Google even calls out these relationships as a potential threat looking forward:
We rely on our Google Network members for a significant portion of our revenues, and we benefit from our association with them. The loss of these members could adversely affect our business.
We provide advertising, web search, and other services to our Google Network members, which accounted for 30% of our revenues in 2009 and in the nine months ended September 30, 2010. [...] If our Google Network members decide to use a competitor’s or their own web search or advertising services, our revenues would decline. Our agreements with a few of the largest Google Network members account for a significant portion of revenues derived from our AdSense program. [...]
The key takeaway so far is that Facebook is currently a much smaller player than the comScore publisher data indicates when accounting for display ad networks. And this is precisely the area where Facebook has the greatest potential looking forward.
Facebook's growth and strength in the market has largely been due to approaching their business as a platform, rather than as a website. Brands such as Zynga have been created on the Facebook platform, and many other sites have incorporated Facebook's capabilities on their sites through Connect or other features. While Facebook has created one vector of monetization based on these partnerships, namely Facebook Credits, in terms of advertising, they have been operating only as a website. Sure, every pageview on a Facebook game also has ads on the side served by Facebook, but this is hardly integration.
As for the other ad networks, an advertiser serving ads into those networks targets their audience by contextual relevance (either through technology like AdSense or by selecting particular publishers) or through some sort of behavioral targeting through cookies, such as search re-targeting. Facebook could wipe the floor with these other networks by creating an ad serving solution for partner sites that serves ads with the same demographic targeting as can be performed on facebook.com.
As for why they haven't done this yet, a couple of possible reasons. The technology and patents involved might be holding them back, but at this point in Facebook's growth, this could be solved with an acquisition (much like DoubleClick). They wouldn't even really need to acquire for current reach, as their promise (niche demographic targeting) far outstrips any competitor. There's the privacy angle, but Facebook really wouldn't be opening up data to third parties if their servers were serving the ads. It'd be more like facebook.com ads were being extended to 3rd party sites with a revenue share and better ad formats. I think the most likely reason this hasn't happened yet is that Facebook just hasn't gotten around to it. Like most new media companies, their product development teams are far ahead of their ad product teams.
Personally, I hope Facebook recognizes the potential for growth and seizes the opportunity. It would revolutionize the display ad market, increase the CPMs and CPCs publishers make and decrease the total amount of ads web-surfers see each day (and by doing so increase attention paid to ads by eradicating the clutter). It would make display advertising across the web far more relevant to consumers. The change would be good for advertisers, consumers, and publishers. The only ones it would be bad for would be Microsoft, Google, Yahoo!, and the others listed in the graph above. And for them, it would be very, very bad, as there isn't a way to compete. At the end of the day, Facebook is a data company, not a media or technology company. And if their ad network's promise is the ability to target based on Facebook's data, thats a sales proposition no competitor can come close to matching.