Yesterday Facebook announced the launch of its newest don't-call-it-email service, Facebook Messages, and I speculated that this is a move designed to capture more (a lot more) of the valuable online display ad market.
There's two big questions to ask about the new Messenger. One, how does Facebook make money? Two, how does Messenger help them make more? And (ok, there's three really) how does this particular move by Facebook fit into their broader commercial strategy?
Facebook - share of the online display advertising market
Facebook is on track this year to make $1.1bn, primarily from display advertising. This number has been confirmed by Mark Z as "not so far off in either direction that it's causing us any pain", and unlike last year's industry estimates the business feels no need to correct it; it is consistent with a well-known and consistently accurate investment model built by Yahoo in 2006 to support an acquisition approach; and it is consistent with the CPM display ad rates Facebook is believed to command and their share of both the US and UK markets (the two most valuable in the world). For now, it's reasonable to just call that number true as a working hypothesis.
UK online display advertising is worth about £800m/$1.2bn pa (or about 20% of a total £4bn/$6.5bn online ad market); US online display advertising is worth about $8.5bn pa (36% of a $25bn online ad market). Facebook is now the largest single provider of display ad inventory in both markets. In the US it accounts for 23.1% of impressions; in the UK 31%. However, this doesn't equate to a similar percentage of revenues in either market - if it did, Facebook would be looking at 2010 revenues not of $1.1bn but between $2bn and $2.5bn. There's a big disconnect that Facebook has to resolve if it's to extract full value from its current inventory leadership.
Valuing online ad inventory by category
Different types of inventory command very different prices from advertisers, and those prices depend on multiple factors. Publisher-sold inventory tends to command a higher price than "remnant" inventory sold by ad-networks (in the UK the split of publisher to network is about 60/40, but that ratio has been moving in favour of networks for the last ten years). Inventory sold against premium content (newspapers, finance, tech, travel) is relatively expensive. Inventory that is behaviourally targeted, tracking people as they research specific buying decisions, is expensive too. (According to one study it's an uplift of about two and a half times.) Inventory on social networks is far less expensive, partly because the supply is so enormous and has grown so quickly, partly because advertisers fear the quality and propriety of (user-generated) content they'll appear against, partly because much of it is untargeted.
Competing with Yahoo, MSN, AOL
This has two implications for Facebook's commercial ambitions, and those of its main online display rivals at AOL, Yahoo and MSN.
First, because its entire offering is in the social network category and therefore commands about the same effective CPM as untargeted network inventory, Facebook achieves an eCPM between half a dollar and a dollar, vs a whole Internet average of $2.52, an average for portals of $2.60, email of $0.94 and premium/professional content such as newspapers of $7. With 20%-30% of online display ad inventory Facebook is still only capturing 10% of online display ad revenues.
Second, the reality of selling online display advertising is that it involves managing relationships with quite a small number of buying points - a handful of large media-buying agencies that represent the major advertisers. (In the UK and the US, the top ten display advertisers account for about 10% of all spending.) It is far easier, and arguably more effective, for those buying points to leverage their own scale and negotiate with the top three or four portal/networks - that is, Yahoo, MSN, AOL - who can all offer ad inventory against the desired combination of premium content, cheap network space and behavioural targeting than it is for those buying points to shop around every time they want to book a new campaign.
Further, because Yahoo, AOL and MSN operate their own large-scale third-party ad network their reach and share of total online display inventory is much greater than some headline publisher figures suggest - for example in the UK their combined share of online display ad inventory is somewhere in the region of 30% (own calculation from this data and this data). If Facebook is going to compete to sit at the (very small) table when media buyers are deciding where to put their clients' online display money next year, it needs its own network, it needs its own premium inventory, it needs behavioural targeting. And of course it wouldn't hurt to cripple the offering of the existing three incumbents by taking out their biggest and most reliable page farms - by replacing webmail with Messenger.
Problems Messenger solves
Facebook Messenger has the potential to kick the legs out from under MSN, Yahoo and AOL's owned-inventory, which is still heavily email-dependent. If Messenger is the new IM and the new email, the scale of Facebook's major competitors is set to fall. That's important on two levels. Webmail inventory is more valuable than social networking inventory - not quite twice as valuable on an eCPM calculation, which at this scale means purely incremental revenues for Facebook and opening up relationships with advertisers who just don't want social networking inventory. Second, and more importantly from a competitive point of view, reducing the scale of the its major competitors would be a huge win for Facebook. MSN/Yahoo/AOL offer such a compelling proposition to advertisers and ad buyers because they have both scale and all of the ad and content formats a buyer needs. Transfer that scale to Facebook, and the current portals becomes a relatively less attractive proposition and less of an automatic first choice on the roster.
Messenger offers a potential solution to the cookie problem that is faced by existing ad networks. Historically ad networks have relied on cookies to generate user data, and because the largest networks have therefore generated the most data this has given the winners defensible advantages. But people delete cookies; better security systems block them or give the option to block them; and even the largest ad networks aren't on every page. By contrast, many people stay logged into Facebook all the time already, leaving it running in a background tab. Make FB the primary communications platform for the web - replacing both IM and email - and more people will stay logged into it all the time. That's each user automatically identified wherever they are and whatever they do online. (Think privacy concerns will stop them? These are guys who don't even believe privacy is a thing any more.)
Most importantly, Messenger significantly expands Facebook's potential behavioural targeting offering. Behavioural targeting improves yields - eCPMs - significantly, especially vs a network sell. Facebook's existing user data is rich and extensive but add to that piece every word that user would have written as an email or an IM, stored forever, and the potential to target relevant ads to that user - both on Facebook and off - expands significantly. Facebook already knows where you live, it knows how old you are, it knows who your friends are. With Messenger, it will know what you're planning to do tonight also. That half a dollar CPM suddenly looks pretty lowball if it transforms social networking inventory to far more valuable behaviourally targeted advertising.
Part of the wider strategy
The "Like" button was a step towards building out a Facebook ad network by getting Facebook onto every content website in the world. Mary Meeker's awesome presentation earlier today gave several examples where the "Like" button presents new branding and monetisation opportunities. The Friendfeed acquisition last year gave FB a content aggregator, of sorts. That's the two steps left to take for Facebook to take on the big portals and own the online display space. An ad network looks like an easy move for FB to make - it can acquire one or build it out of its existing sales operation. Premium content is less straightforward to generate. All of the big players have made significant investments to build or buy content in the necessary commercial categories. There are a number of ways Facebook might try to enter that space, but as yet very little it has done seems likely to get it there.
Financial implications and next steps
Facebook already carries dramatically more inventory than anyone else, but is less able to monetise it. To cross that gap and become the clear number one player they need to take out the major incumbent portal/networks in online display, they need to apply effective behavioural targeting to their own underpriced inventory, they need to start building out a third-party network to complement their publisher-owned offering and they need to be able to offer premium content to attract the top CPM advertisers. (They don't especially need to generate yet more billions of ad impression pages but it seems that this move will achieve that also, diluting the relative scale and value of the other players.) The combination of the new Like button that is flooding the web with Messenger gets Facebook most of the way there. All they need to add now is premium content, or some other solution that renders the premium content play irrelevant. For Yahoo and AOL, this is probably the end. For MSN and for Google's foray out of search into display the competitive landscape is going to be a lot meaner. And for Facebook - with even more inventory, behavioural targeting across a larger network and the de facto elimination of key competitors, the road to $100bn looks less and less impossible.
(Photo from b_d_solis on Flickr)