ChasNote

That was Battelle’s money quote on a panel last week at EconSM, his way of describing what FM is doing with marketers to bring them into the conversations at leading independent social media sites (Yahoo Publisher Network blog). Something the ad networks and Google AdSense can’t yet do.

David Eun, Google’s VP of content partnerships, at yesterday at PaidContent’s EconSM: “In talking about the polarization between those who focus on the hit-driven head and the user-generated long tail, Eun said YouTube wants to focus on the middle.”

Reading the rest of the quote (I wasn’t there), it appears Eun is talking about content producers. But I’ve been hearing this sentiment from advertisers, too. Marketers are covering segments of their audience that still visit the portals (the “head”), and they are reaching deep down the long tail with direct-response dollars spent with Google AdSense and the ad networks. It’s bringing their brand message to the middle — high quality niche sites that are taking greater and greater share of high quality audiences — and doing it with efficiency that’s still hard.

Rafat Ali at PaidContent rounds up this weekend’s coverage of Google’s acquisition of Doubleclick:

“Competitors like Yahoo, Time Warner and Microsoft, all of whom did bid for DCLK, are talking about anti-competitive nature of the deal. The deal will be subject to a review by either the Department of Justice or Federal Trade Commission.”

This is the first time Microsoft has called an anti-trust foul on someone else, according to the Financial Times!

I can’t imagine the Justice Dept will stop the deal, but Google, very obviously, is hot to move up-market in the online ad world, and Big Media is terrified. Google currently dominates the bottom, CPC direct response advertising, and in some cases (sites that stick to vertical content categories that are well supported with endemic ad dollars, say enterprise technology or auto-buying news) they perform better than the ad networks that are supposedly a notch above them in the food chain. In other cases (sites with high-quality readers, but broad content coverage) the ad-network model makes more sense and more money for publishers. As Google continues to struggle to get into the high-end brand marketing game, they recognize an easy new market to conquer is the ad-network world.

Doubleclick allows them to take their existing publisher network and plug in graphical ad units, served based on criteria beyond key-word context matching. Presto, they just put every ad network out of business.

This still doesn’t get them into the premium brand-ads business. But it’s the platform they’ll need when they do get there.

In the meantime — call it their “brand advertising market research phase” — they get to see what every advertiser pays for site-specific graphical ad programs. This gives them leverage in negotiating with existing publisher partners. “Hey, NY Times, let’s be honest. You sell 26% of your inventory at an average net CPM of $17. Let us have it, and we’ll do 100% at $6, and everybody wins.”

It’s also good market intelligence they can use to steal advertisers from the direct sales efforts of those publisher partners: “Hey advertiser, I’ll get you on this particular site, dayparted, geo-targeted and context matched…. and I’ll do it for a nickel less than the publishers themselves are offering.”
Scary situation for the media titans. When the anti-trust suits don’t pan out, I’m betting the Google-Doubleclick deal will drive an active year for digital media M&A.

Last week’s TNS report (see Clickz) on 2006 ad spending put numbers to an obvious trend: Online advertising is the growth leader (up 17.3%), while print is mostly down and TV (minus Spanish-language) is just barely above flat.

Within online, the report also surfaced a less obvious trend, that large marketers are moving dollars to online display advertising more slowly than smaller-budget advertisers. On average, marketers in 2006 allocated 6.5% of their budgets to online while blue chip brands put only 4% against Internet ads. TNS Senior VP of Research Jon Swallen chalked up the slower adoption by the big spenders to an availability problem: “it’s hard to find inventory to spend more on.”

Huh?! Not enough online inventory? Even at top brands like Yahoo, certain large swaths of the site (Mail, My Yahoo, Groups, etc.) run house ads and cut-rate remnant deals. Ad networks don’t have anywhere near 100% fill rates for participating publishers, and the most efficient ad network of all — Google’s AdSense program — serves PSAs a fair amount of the time.

Market Inefficiency v. Lack of Inventory

Clearly there’s not an inventory problem online, but instead a quality inventory problem. There is limited inventory adjacent to high-quality content, and — maybe this is what Swallen meant — it’s still very hard to find it and sponsor it. There’s a market inefficiency problem that the ad networks and Google aren’t yet solving. The vertical publishers (think CNET’s tech sites or Jupiter Media or Nickelodeon Online) have quality, but not enough of it; the ad networks and Google have scale but can’t yet guarantee quality.

The Story of Google’s Effective Ad Rates

The two sides of Google’s money machine illustrate the point. Google takes a purely rational approach to advertising: Advertisers in a competitive-bidding situation will pay higher and higher cost-per-click rates until the costs approach the immediate revenues generated by that click. The average CPC paid by Google advertisers is around $0.50. When people use Google’s engine to search for something specific, 17% of them will click on one of these advertisers’ ads (see ChasNote). The algorithm that matches keyword advertisers with keywords entered into a search box works great for all involved. Users of Google’s search engine are seeing highly relevant ads (a perfect match 17% of the time!); marketers aren’t paying a nickel more than the value they get from each click; and Google makes about $68 for every thousand search-results pages they serve up.

But Google’s other system, the AdSense publisher network, tries to match those same key-word advertisers with prospective clickers on other sites based on content rather than key-words entered into a search box, and doesn’t work as well. That 17% click-through rate on Google’s own pages, a proxy for how well the ad-server logic delivers ads to people who want to click on those ads, drops to something closer to 1% on non-Google sites. (Consumer mindset, searching versus reading, is another factor.) Direct-response marketers may or may not care about the lower click-through rate: They’re only paying for the completed click.

Brand Marketers Demand Scale, Quality and Safety

Brand marketers, however, are paying for more than click-through performance alone, and they hold media partners to different standards. Even infrequent snafus that put their brand — and the implied endorsement by their brand — alongside homophobic rants by Ann Coulter (ask Verizon or Washington Mutual, see CNN) or on sites that promote illegal behavior (ask Microsoft or Wal-Mart, see Variety) become unacceptable PR disasters. Brand advertisers need a combination of scale, quality and safety (more on this from Battelle, see Searchblog), and the existing artificial intelligence solutions aren’t there yet. The bots get better every day, but to accelerate the migration of brand dollars to the Internet, we still need human insight to facilitate the process.

ClickZ reports on the latest TNS study, which estimates that online display advertising was up 17.3% in 2006. Hispanic TV was up 13.9%. Overall ad spending was up only 4.3%. Blue chip brands spent an average of 4% of their total budgets online, lower than the 6.5% average share industry wide:

“One reason blue-chip advertisers spend less online is it’s hard to find inventory to spend more on, [TNS Senior VP of Research Jon] Swallen believes. ‘Part of it is scale. For Procter & Gamble to put 6.5 percent of its marketing budget into the Internet, they’d change the name from Google to Procter & Gamble,’ he said.“

Hold on — not enough inventory? There is endless inventory online. Google isn’t pushing text-ad PSAs because they’re committed to those charities! I’d argue instead that this is more evidence that brand advertisers are still not confident that the contextual-relevance algorithms at Google and the ad networks can deliver enough quality inventory.

Google’s black-box policy toward how it’s algorithms work presents a major challenge to legitimate businesses that are contemplating name changes, site redesigns or other adjustments to their URL structure. From Techdirt:

“Decisions such as name changes or shifts from .net to .com domains are being influenced by how they’ll effect search results, with many companies frustrated with the feeling of being held ransom by their Google rank or other factors. This frustration grows when search companies aren’t particularly helpful to businesses making legitimate changes, preferring to let their algorithms, which often ignore certain kinds of behavior because they signal fraud or other funny business, handle everything. While it’s certainly well within the rights of a company like Google to tinker with its algorithm, it and other search engines’ vast amount of influence over sites’ traffic and business could make many people feel they have some responsibility to make things a little easier for the legitimate companies that aren’t looking to scam the system.

Another snafu moment for brands whose ads ended up sites they didn’t want to be on. (Thanks for sending, Lester!) From CNN:

“At least three major companies want their ads pulled from Ann Coulter’s Web site, following customer complaints about the right-wing commentator referring to Democratic presidential candidate John Edwards as a ‘faggot.’ Verizon, Sallie Mae and Georgia-based NetBank each said they didn’t know their ads were on AnnCoulter.com until they received the complaints….. Verizon, Sallie Mae and NetBank said the ads were put on a variety of sites by a third party company. In many cases, advertisers do not know which sites feature their ads.”

…. An analyst friend who covers Google reminded me that not all Google pages are search-results pages carrying ad links — only 70-75% of Google pages carry ads — which explains the difference between the $68 eCPM figure and the $92 figure. Multiply $0.54 average CPC with 17% average click-through rate with 74% pages that carry ads with 1000 to convert to revenue-per-thousand-pagviews and you get $68.

For the full calculation, and its implications on Google getting serious about selling programs to brand advertisers, check out my post from last week.

A well-informed dinner companion last night told me Google’s average effective CPM on Google-owned search results pages is $68. Another guest at the table said Google’s average CPC for paid search (not AdSense) is $0.54, and another said click-through rates on paid ads on search-results pages average 17%. Multiply those together an you get a $92 eCPM.

Compare that the average CPM for television spots in the US: $11.

It’s no wonder Google is struggling with how to enter the brand-advertising game. Sure, there’s $80 billion in US TV spending. But the CPMs are 10-15% of what Google gets (in some parts of their business) today. Plus selling to brand advertisers requires expensive humans instead of robots and auction platforms.

UPDATE: Battelle has posted Google’s K-1 posted at Searchblog, and calls out a remarkable stat: “Astonishingly, cost of sales is 8 percent. Very low and very efficient.” Another obstacle that makes it hard for Google to get excited about hiring a bunch of human salespeople to call on brand advertisers. They’ll either succeed at automating the buying and selling of brand programs — I’m a skeptic — or they’ll go down trying.

The K-1 also let’s me extend (or perhaps check) my voodoo math. If Google Network (AdSense, etc) is $4.16BB of Google’s overall $10.6BB, that says $6.45BB is from advertising links on Google’s own pages. $6.45BB divided by $68 per thousand pages, times 1000 is about 95 billion pageviews over 12 months, or about 7.9 billion pageviews per month. Comscore Media Metrix estimated pageviews for “Google Sites” to be 11.8 billion for the month of December 2006. Assuming monthly traffic grew over the course of the year, that math about jives.

ANOTHER UPDATE: An analyst friend who covers Google reminded me that not all Google pages are search-results pages carrying ad links — only 70-75% of Google pages carry ads — which explains the difference between the $68 eCPM figure and the $92 figure. Multiply $0.54 average CPC with 17% average click-through rate with 74% pages that carry ads with 1000 to convert to revenue-per-thousand-pagviews and you get $68.

From Mediapost: Starting in Q2, Google will begin to share with AdSense advertisers the list of sites on which a campaign ran. This is big news, given Google’s historical aversion to transparency on both sides of the AdSense program — it hasn’t before told advertisers where they will (or did) run, and it hasn’t told publishers their actual share of the ad revenue Google generated from the pages of their sites. Could it be that Google’s competitors are starting to apply pressure??

“A report in The New York Times suggested that the move came in response from growing market share gained by contextual advertising firm Quigo, but a spokesman for Google denied the connection.

Stylman said it’s unlikely that Google would change its strategy based on a smaller competitor’s system, but that Quigo’s approach was gaining significant share in the market.

“I don’t think that Quigo is a significant enough player for them to change their entire system based on them,” he said. “Having said that, Google’s intelligent enough to keep an eye on market trends, and Quigo is definitely gaining some market traction.”

Quigo last week announced a multi-year partnership with Forbes to serve contextually targeted ads on its sites including Forbes.com, ForbesAuto.com and ForbesTraveler.com.”