Web publishers, for whom online advertising is generally the bread and butter of their livelihood, have, naturally, been watching for changes in the online advertising market with anticipation. A look at economic projections has not been particularly reassuring, especially considering the generally-accepted idea that advertising is the line item to take a hit on a company’s budget when times are tough.

I wrote a month ago about a eMarketer study that showed search advertising growing at a faster clip than display advertising, which suggested that, in lean times, advertisers are more interested in measurable performance. A survey of 175 CMOs and marketing executives conducted by Epsilon shows that, indeed, among those who control the pursestrings, spending is generally being pared back. But, like the eMarketer study’s projections, those surveyed are increasingly moving their shrinking ad budgets online.

Here are the nuggets:

• 65% say their ad budgets are shrinking due to the poor economy
• 63% say their spend on interactive/digital advertising has risen (14% say it has decreased; 23% report it staying the same)
• 59% say spend on traditional media has dropped (13% increased; 29% stayed the same)
• when it comes to new media, 42% (the highest percentage) said that they plan on increasing spend in social media; blogs were #2 at 35%

True to the old adage, when the going gets tough, the tough get going, marketers are doing their best to squeeze as much performance possible out of every dollar spent. In the world of marketing and advertising, fortunately, new media and advertising technology is giving advertisers some opportunities.

YieldBuild is happy to announce that we’ll be hosting Lunch 2.0 at our SoMa digs on Thursday, October 30th. It’ll be the day before Halloween, so candy-loving sweet tooths will be especially happy, although we’ll have a spread of more traditional fare that won’t cause as many cavities.

If you’re interested in attending, please join the Lunch 2.0 Group on Facebook, and follow the instructions on how to RSVP a spot for yourself when the organizers blast out a message about it.

Alan Patrick at Broadstuff has an interesting take on Tim O’Reilly’s sobering keynote speech at Web 2.0 NY last week. Tim O’Reilly warned that applications that focused on capricious novelty, like throwing sheep using SuperPoke on Facebook, or iBeer on the iPhone, are taking away the brainpower that could be solving mankind’s problems, like global warming and financial market meltdowns. A serious splash of cold water, ironically from the man widely credited with coining the term “Web 2.0″.

Alan’s inference was that Web 2.0 as it stands now can not support themselves through advertising alone: the pie just isn’t big enough for serious companies in the Web 2.0 space to carve out huge valuations for themselves when the behemoths, like Google, Msft, Yahoo and AOL, have garnered most of the ad spend.

He goes on to argue that poor quality content is a natural consequence of an everything-is-free market, which doesn’t reward high quality. This, of course, goes against Chris Anderson’s thesis, explored in depth in Wired in February (and one that he’s developing into a book, Free), that an advertising-based model is the future.

I disagree with Alan (and Tim) on a few points:

• some of the sillier apps, like iBeer, are the fruit of developers testing the app markets. One dubious success can give an entrepreneur the cash, connections and confidence to pursue greater aims. Remember that one of Jimmy Wales’s first contributions to the world of online media was Bomis.
• Google et. al. do eat up a big portion of the pie, but they are supporting ad income for a wide swathe of  Web 2.0 sites (Digg is supported by Microsoft ad revenue; HubPages, in large part, by Google, to provide just a couple of examples).
• By almost all measures, the online ad pie continues to grow, despite overall ad spend shrinkage as economies tighten worldwide. This is because the wider move from offline to online in users’ habits outweighs ad budget contractions.

It’s not hard to feel Tim O’Reilly’s frustration. There are thousands of brilliant programmers whose genius isn’t readily apparent when we use one of their disposable, forgettable apps, and we do face considerable challenges that need some of the best minds to work on them. But I would argue that silly novelties are just one part of a greater phenomenon, in which people grow familiar with and appreciate technology enough to reward those who create it, and that greater trend is something worth feeling optimistic about.

MarketingSherpa has released the results of a study looking at the clickthrough performance of the most popular ad unit sizes:

In some ways, the general picture this data paints is not surprising:

• the 300×250 medium rectangle is almost always placed contextually, juxtaposed with content. This is always a good strategy to get visitors to notice.
• the 728×90 leaderboard is typically placed at the very top of the page, so it is one of the first things that visitors notice.
• the 160×600 wide skyscraper is usually positioned in the navigation column, often below navigational/informational links or widgets. I’d suspect the performance of this ad unit has the widest variance (although I don’t have the data to say one way or another)

Although the 728×90 leaderboard signficantly outperforms the 468×60 banner, clickthrough performance is not solely a function of size. In fact, placement is probably a better indicator. Take a look at the YieldBuild Lab Heat Map to see the most common placements of ad units across the top 100,000 (by traffic) websites.

Ad optimization has been enjoying growing attention in the media, as publishers seek to maximize the online ad revenue potential of their traffic. The continued growth of the contextual advertising mainstays, Google AdSense and Yahoo Publisher Network (YPN), alongside a proliferation of generalized and niche display ad networks, has given publishers plenty of options to mull over.If your goal is to maximize the total revenue of your site, here is a list of work you have to do:

• Decide which ad size, color, and format ad to display
• Choose the most lucrative ad network for each spot on each page, and deciding when is the best time to run it.
• Pick the right location for the ads, which involves:- The right number of ads (avoiding a cluttered page)- The right mix of ads (ensuring different ads work well together)- Determining the point at which visitors will ignore the ads, and preventing it from happening.- Creating visibility across ad networks and monitoring performance

There are four variables that affect your ads’ performance (something we call The Four Pillars of Online Ad Optimization):

• Ad Size: altering an ads dimensions can affect its performance, although bigger isn’t always necessarily better
• Ad Format: background colors, border styles and colors, fonts, and other visual characteristic optimization can yield an over 10% improvement in performance alone
• Ad Placement: where ad units show up, even at the pixel level, can affect performance, although some general rules (like keeping ads above the fold, and near navigational elements) will help most publishers
• Ad Network: different ad networks will pay out differently, although each might have requirements for placement, caps and other restrictions that must be complied with

Because this is an exceedingly complex problem space, but there’s so much to be gained by solving it for your site, many publishers use YieldBuild, which does all this work automatically for them. Others prefer a more manual, hands-on approach, and instead try their hand at homegrown optimization, using either A/B or multivariate testing. There are rules of thumb, however, that will generally give you solid performance, although with plenty of room for additional gains by using YieldBuild or a rigorously-applied optimization program. Here are some rules of thumb—tailored for AdSense, but that have pointers that will work with any ad network—that will help you get started:

• AdSense Formats & Formatting - selection of colors, border styles and ad unit sizes
• AdSense Placement - selection of optimum layout and positions for your ad units
• AdSense Optimization - selection of a testing protocol and optimization methodology 

 And, if you’re using a popular forum or blogging platform, take a look at these specific recommendations:

• AdSense Optimization for Forums (Discussion Boards)
• Optimize AdSense for your Blog
• AdSense for Blogger (Blogspot)
• AdSense for WordPress

Behavioral targeting has been one of the most promising avenues explored to optimize online advertising efficiency. Some behavioral targeting ad networks, such as Revenue Science and Tacoda, have built a business model around identifying web visitors’ browsing patterns (and implied interests and buying intents) and delivering relevant ads.

Because privacy concerns continue to percolate through the industry (both NebuAd and Phorm have battled privacy advocates, mostly because both firms have collaborated with ISPs to gather user data), there has been plenty of discussion about how to achieve the promise of BT without it spiraling into a PR nightmare for the ad networks involved. One company seems to have a solution, by decoupling user behavior data from ad serving.

Startup BlueKai, founded by some ex-Revenue Science folks (they are based in Seattle, as is RS), has developed an behavioral data exchange, in which aggregate user data can be bought and sold by advertisers and Web publishers. It is similar to list aggregators in the offline world, although, because the data is far richer and more reliant on behavioral data than demographic, user lists can be sliced and diced in many more ways. Buyers effectively buy a cookie associated with a profile targeted by the buyer, who might have searched for a certain class of product or service of interest to the BlueKai client.

Data on users who have visited video sites, or looked for information on hybrid cars, can be collated and sold, for instance, to video ad networks or online car retailers, for instance. In fact, Autobytel and Tremor are two of BlueKai’s first partners.

How does BlueKai avoid trespassing on privacy concerns?

• user data can not be resold
• user data expires within one month of purchase
• users can manage their individual profiles, granting full opt-out privileges, by accessing a registry hosted by BlueKai
• users who do participate are rewarded by a donation to a charity

Tremor’s and Autobytel’s experience using BlueKai will likely be watched carefully. Regulators’, and their constituents’, sensitivity around advertiser overreach won’t necessarily make BlueKai’s sell particularly easy.

The Google-Yahoo deal announced three months ago is running into some trouble, from advertising trade associations that contend that the deal violates U.S. antitrust laws. They got the attention of the U.S. Department of Justice, who hired veteran antitrust attorney Sanford Litvack to review the evidence and see if the DoJ has case with which to block the deal. As was the case when the DoJ went after Microsoft a decade ago, murmurs have started bubbling through the blogosphere that Google is simply too successful for its own (or, rather, our collective) good.

What exactly is the case against the deal? It’s useful to remind ourselves why antitrust legislation exists. It’s not to punish successful companies. It’s in place to prevent companies with considerable market power from harming both consumers and entrepreneurs. A monopoly can use its dominant market position to charge more to consumers because there aren’t enough competitors to bring down pricing to market-determined levels. And a monopoly can exert its power to shut entrepreneurial competitors—which usually act as the wellspring of innovation—out of the market. The end result of unrestricted monopoly is a reduction in market efficiency, and, more long-term, the technological decline of an industry where the dominant player sees more benefit in muscling out competitors than investing in innovation.

Whether this scenario is the case with respect to Google remains to be seen. But as it’s portrayed by the litigants, Google already dominates the search advertising market, and working with Yahoo brings no. 2 into its fold. Google is not known for transparency with respect to how much of a cut it takes from advertisers, the gap between what advertisers pay via AdWords and what Google pays out to publishers via AdSense. Competitive forces would force Google to pare down its share to a minimum, in order to entice advertisers to spend with them. But without viable competitors, Google can take as large a chunk as it wants, and advertisers will simply have to pay to play.

How much does Google control?

• A study by Attributor in March shows Google serving ads on 35% of ad-supported Internet traffic (double that if you include DoubleClick).
• Our own survey of the top 100,000 sites by traffic shows about 60% of ads on the front page are served through Google.
• Search advertising is projected to be a $10.4 billion market this year, and continues to enjoy double-digit growth rates

The months ahead will see a decision by the DoJ on whether to pursue action against the Google-Yahoo deal. Then we’ll see whether the US government sees Google’s future growth as fueled by its own (or acquired) innovation or deft strategy, or through limiting inroads to the sizeable search ad market to its competitors.

Yoni Baciu, one of YieldBuild’s engineers, has overlaid imprints of 20,000 ads randomly selected from the Web’s top 100,000 Websites (by traffic) on one blank slate. The result is YieldBuild Lab’s Heat Map which renders the most popularly-selected positions on a Web page with a brighter red glow.

What does the Heat Map tell us?

First, header position ads, and sidebar ads (particularly on the right side), are popular, and ad density also decreases as you go down the page. The darkest portions of the Heat Map are to lower-left.

This is not to say that this a protocol for placing ads if you want to maximize performance. Generally, ads should be within a users’ line of sight, close to navigation elements, and embedded in read content. But site-specific recommendations are usually the result of careful testing (or implementation of YieldBuild, which finds the best-performing regions for each page algorithmically).

Interesting, Google’s own recommendations for AdSense publishers suggest heavier weighting towards the top of content (although not necessarily the header) and also favoring the left more than the right.

If Google’s recommendations for AdSense publishers prove to be generally true for publishers using other networks as well, then that darker-red zone to the middle-left probably presents some opportunity to increase revenue, although without YieldBuild or testing, there’s no way for a publisher to know.

Master New Media’s Robin Good spoke with YieldBuild CEO Paul Edmondson recently, and posted the video of his interview with Paul Edmondson on his site. They cover what YieldBuild is, how it benefits publishers in terms of its ease of use and performance, why ad optimization like this matters, and how YieldBuild compares to Pubmatic and the Rubicon Project.

For a full transcript, see Robin’s post.

Both of the conventions are over, and the US electorate has only two more months of campaign misery to endure. Television commercial spots attacking Obama and McCain will saturate the airwaves from now until early November. Those who’ve gotten sick of it all could turn to the Web, but even that is no longer necessarily a safe haven from partisan mudslinging. The election which has embraced online social networking has also been a boon to contextual advertising, particularly Google.

Here are some interesting figures on the campaigns’ spending and performance on online advertising:

• The majority of Obama’s online ad spend—59%, or $3 million—from January through July of this year, was spent on Google ads. Yahoo followed with $618K, and local advertiser Centro a bit more than half a million dollars. [ClickZ]
• ComScore data shows that for the first half of this year, the Obama campaign paid for 12 times as many display ads as the McCain campaign. [MarketWatch]
• Nielsen Online data points to heavier spending by the McCain campaign on search advertising compared to the Obama campaign. June saw 7 million search ad impressions from McCain, and 1.15 million from Obama. [National Journal] The gap widened in July, when McCain paid for 15.1 million search ad impressions, while Obama barely budged at 1.2 million. [WSJ]
• ‘Typo-squatted’ domains (like barckobama.com or johnmcain.com) are squirreling away online spend. A Symantec study showed that 47 of the 160 misspelled variants of Obama’s site were held by typo-squatters. [AFP].

Naturally, advertising is only one part of the candidates’ presences online, but growing spend demonstrates both its effectiveness and the growing traffic directed at the two candidates.