Darren Rowse, of ProBlogger fame, guest-blogged at the Inside AdSense blog today with Six ways to experiment with AdSense and grow your earnings, and, speaking from years of intense experience, had a lot of useful advice for budding AdSense-monetizing publishers to make the most of their AdSense placements.

His key advice: test (experiment), track, repeat. As his post makes clear, there is no magic formula that works most of the time. Each site, page and placement can vary a bit, which makes certain rules of thumb moot. This is exactly why YieldBuild works as well as it does to optimize publishers’ AdSense: it tests, tracks and repeats. (Thousands of times!)

Let’s run down Darren’s list of 6 ways to experiment:

1. Ad Position:
   Darren says... Check out the Google heatmap. Ads near/alongside, or at the end of content, can do well, but experiment.
   YieldBuild adds…
   If you’re using YieldBuild, places YieldBuild zones wherever you think an ad could do well. YieldBuild will determine which zones actually perform best (we do the experimentation for you), and will collapse any zone that isn’t filled with ads.
   If not using YieldBuild, I’d add that you should take a look at placing ads embedded in content, near navigational elements or near pictures where the eye is drawn towards (but not so close as to mislead a site visitor into thinking the ad must be clicked on for navigating the site). More pointers here.
2. Number of ads
   Darren says… add ads, but be careful of putting in too many. Longer pages can carry more ads without bothering yoru visitors.
   YieldBuild adds… true about the number of ads; more isn’t necessarily better. YieldBuild automatically determines that sweet spot (it often serves up fewer than the maximum ads, just because fewer performs better). If you aren’t using YieldBuild, I suggest starting conservatively, with a few ad spots on the page, and then slowly ramping up the number of ads, carefully monitoring CTR, revenue, and traffic feedback. Like Darren said, each page has its own “tipping point,” but you won’t know what that is until you test it.
3. Ad Design
   Darren says… try blended or subtle backgrounds, and use the AdSense ad rotator to fight “ad blindness”.
   YieldBuild adds… If you have to go with a rule-of-thumb, go with blended (the same background color for your ad as the page you’re embedding it in) and/or slight variations of the background color. This is what our own testing has demonstrated. However, as that test shows, sometimes a completely different variant can significantly outperform the blended or subtle variant.
4. Ad Sizes
   Darren says… experiment, but bigger isn’t necessarily better. The most popular standard ad sizes will generally outperform more esoteric ones.
   YieldBuild adds… Darren is spot on about both the most popular ad sizes (FYI- they are 728×90, 160×600, and 300×250) and bigger isn’t necessarily better. YieldBuild, naturally, experiments automatically for each site and serves those ad sizes that perform best. Without the benefit of YieldBuild, I’d recommend sticking with the 3 most popular sizes, placing the leaderboard (728×90) at the top or bottom, the skyscraper (160×600) in your navigation column, and embedding the medium rectangle (300×250) within your content. That last one is likely to do best.
5. Ad Formats
   Darren says… his experience suggests text and image work better than text alone, but you must test.
   YieldBuild adds… you really, really can’t know without testing. YieldBuild tests for the best format automatically, but otherwise, try from among text, link units, image and video formats and see which performs best. Sorry – no rule-of-thumb here. We’ve seen too many examples of sites that run against Darren’s advice, where text ads really do perform best, to agree with him with respect to this.
6. Which Content Converts?
   Darren says… the marriage of AdSense and Google Analytics gives you great insight into which pages are monetizing. Use that to determine your content and promotion strategies, and to help decide from among formats.
   YieldBuild adds… we couldn’t agree more, although YieldBuild will take care of the format guesswork automatically. AdSense actually has a tab for Analytics, and by setting up goals, you can drill down by different campaigns or traffic sources and determine what sort of traffic is earning you the money.

Internet publishing giant IAC sounded a dismal tone in yesterday’s earnings call, when chief Barry Diller shared the following about IAC’s properties’ online ad revenue picture:

• CPC (think AdSense and the like) took a 8-12% hit.
• CPM (display) projected to drop 50% in the month

Both of these figures are considerably less rosy than the general market projections made by eMarketer just a few months ago, but given the size and breadth of IAC’s online holdings, it’s not unreasonable to expect their figures as a bellwether for the entire market. HubPages, which primarily monetizes via CPC, has taken a very similar hit, as well, over the past month or so.

Paul Edmondson, YieldBuild CEO, and Xavier Vespa, from HyveUp.tv, had a chat this past Friday about online ad optimization. Here’s Xavier’s accompanying entry, and here’s the video:

Thanks, Xavier!

ClickForensics, a click fraud auditing company, has released a report that says click fraud rates, the percentage of clicks on CPC ads resulting from fraudulent activity (either from publishers earning money from the ads, or from competitors trying to drain a company’s ad budget), has inched upward: 17.1% in Q4 2008, from 16.6% for the same quarter in 2007.

But wait—doesn’t Google already filter out bad clicks?  Google already has a three-pronged approach to identify and weed out clicks from dubious sources (click farms, botnets, competitors), and reflect the corrected traffic quality in the form of Google’s “smart pricing”. However, ClickForensics claims in its FAQ that Google only filters out 2-3% of clicks (substantially lower than what ClickForensics is claiming is fraudulent, but Google disputes their numbers), and Google itself integrated ClickForensics’ FACTr fraud documentation & reporting service in October 2008.

The deal recognizes that even a data giant like Google can’t always stay ahead of the fraudsters, and gathering useful data from auditing services and, likely, its competitors is a good idea. At the same time, the fact that CPC ad networks earn revenue from even fraudulent activity sets up a moral hazard that outside auditing and cooperation with their competitors can head off.

What’s causing the increase? ClickForensics says its the result of more ad dollars flowing into CPC, more competition among CPC advertisers, and the weak economy giving rise to click farms and other money-making scams. The solution? Information sharing among those affected. Most ad networks’ payment cycles are long enough to deter fly-by-night scammers, but since reporting is real-time and advertisers are charged on a much shorter cycle, more aggressive proactive fraud detection is a worthy goal to make.

We have just launched a suite of performance analytics to our YieldBuild publishers, which will give site owners impressive insight into the traffic and monetization patterns that are bringing in the money. Take a look at these 5 sample screenshots  (click on each to see a full-size version launch in a new window):

Although the tides are working against them, vertical ad networks continue to multiply. More sophisticated targeting offered by larger, general media sellers, constricting and consolidating ad spend on branding, and an overcrowded market might dishearten some hopefuls, but there are a couple of entrants I’ve read about this past few weeks that continue to sally forth to seek their fortunes.

• MTV’s Adify-like cluster of vertical ad networks, grouped under the moniker Tribes, launched a vertical ad network, Comedy Tribe, anchored by Viacom-owned ComedyCentral.com, but including the popular JibJab.com site (remember the 2004 election-year flash video?) among its cadre of comedy-oriented properties. Tribes was launched this past September with a parenting-oriented vertical ad network. The Comedy Tribe will compete with the National Lampoon Humor Network and Connected Ventures (CollegeHumor, Today’s Big Thing) to reach the 18-34 male demographic valued by advertisers.
• Also launched yesterday was Inflection Point Media, a vertical ad network targeted towards SMBs, hopes to capture some of the $5.3 billion online B2B advertising projected by eMarketer for 2009. Leveraging its pool of properties that attract 32 million monthly visitors, claims to have behavioral targeting technology that identifies potential prospects and serves relevant ads/offers to those most receptive to them. Because its hub is a collection of 60 business newspapers, IPM will compete against BusinessWeek and Business.com.

Two more are boasting growth and expansion in a purportedly tough market:

• SportSyndicator, an Adify-supported sports vertical ad network, recently opened an office here in San Francisco to support the 10 million US-based sports enthusiasts it reaches across 100+ sites. Climbing.com and BigLines.com are among its publishers, and North Face and Oakley are among its advertisers. It competes with Sportgenic, Yardbarker, CBS College Sports Media and SI.com
• Niche Gay Ad Network pulled away from its competitors (LogoOnline, PlanetOut/Gay.com and Regent Online) to surpass 1 million monthly uniques on properties exceeding 4 million uniques in December. The network partners with 73 smaller gay-oriented sites and blogs.

The success of vertical ad networks going forward will be their ability to fend off the big guys, who’ve gotten better at allowing advertisers to nail down a specific audience profile through which they can target spending. One of the chief challenges is building scale; most leverage one large publisher or a cluster of related publishers to deliver consistent volume. It remains to be seen, though, if ad agencies’ entries into the ad network space threatens to squeeze them out of play. It’s not hard to imagine agencies buying or even building their own vertical ad networks in order to retain the margin, while relying on something like AdSense to make up for any inventory shortfalls.

I was recently told that you are either an ad agency with an ad network, or an ad network with a creative-like agency.  It makes sense: the agencies want a piece of the spend they control, and ad networks want to attract more dollars by offering deeper integration.

I think there are a couple of questions to explore here.  First, is it a conflict of interest for agency holding companies like WPP to own an ad network like 24/7 Real Media, or is it a more efficient way to move dollars through the system that ultimately passes more money to publishers?

For ad networks, I think the big question is if offering deep creative services scales, or does it raise the cost so high that it squeezes all of the margin from the sale?  My feeling is only a few ad networks will be able to offer the creative services, but it’s more likely sites with sales forces, like Pandora, will make it work as a business. Most other ad networks will be better served by looking for more efficient ways of bringing in ad dollars.

I had a minute to go through Google’s earnings and check out the TAC (Traffic Acquisition Cost).  The TAC decreased one point, which is a significant trend.  However, it may be decreasing because of growth in Google Owned and Operated businesses.

Here are some key stats from Barron’s Online.

• Google sites revenue rose 4% sequentially, and 22% year over year.
• Google network revenue was up 1% sequentially and 4% year over year.
• International revenue was 50% of total revenue, down from 51% in Q3, but up from 48% a year ago.
• Paid clicks increased 18% year over year, and 10% sequentially.
• TAC at 27% was down from 28% in Q3.
• Google had 20,222 employees at December 31, up just 99 from the previous month.
• At year-end, the company had $15.85 billion in cash.

One of the most interesting things is how much Google slowed its hiring—only 99 people for the quarter.  That number has historically grown at 13%.  Another area they seem to be pulling back on hard is capital expenses.  In Q4 of 2007 they spent nearly $700 million.  In Q4 of 2008, they spent $368 million.  They may be reaping the benefits of past investments, or the last two quarters are examples of how quick they can pull in the spending reins and a sign of future growth.

TechCrunch’s Erick Schonfeld reports on Microsoft’s dismal quarterly earnings report, with its Online Services division (including advertising) almost doubling its quarterly loss over Q4 2007 to $471 million. Microsoft says that advertising revenue grew at an anemic 7%, with search advertising higher, but these modest increases were not enough to offset ballooning expenses.

So where were there expenses? The development and early marketing of PubCenter, Microsoft’s answer to AdSense, has probably been running up red ink, but that’s to be expected. I’m thinking more about Microsoft’s bets on advertising on social networks. It inked a 3-year deal with Digg in July 2007, but the bigger bet was its advertising agreement (via a substantial investment) with Facebook. Facebook relies on MSFT as its exclusive advertising partner worldwide through 2011.

So why the loss? It’s possible, if not probable, that the deal guarantees Facebook revenue, and Microsoft is eating the loss. After all, Facebook is a social network, and social network users are unresponsive to online ads. Its own advertising platform turns out poor results; it’s not difficult to imagine that ads served by Microsoft wouldn’t perform any better.

Although Google seems to be the last place the newspaper industry should be looking for inspiration—it just recently shut down its print advertising service—the fact remains that the Internet advertising giant has nailed successfully an entirely new source of advertisers at the same time print is losing theirs. The continued growth of online advertising, even in the face of the recent slump, has drawn away traditional, big-media advertisers from newspapers, but there hasn’t been a concomitant cross-migration of “long-tail” AdWords-style advertisers to print. That is a problem for print, given the measurable performance of online advertising and continued, increasing time spent online; this is looking like a zero-sum game that the newspapers are losing.

But does it have to be? Jeff Jarvis, in a great article at Seeking Alpha, argues that newspapers are still well-equipped to survive the changing times; they simply need to adapt. In addition to looking at commerce (some have managed to be successful at this) and leveraging their unique information and readership into building new services (like real estate agency), there are a couple other intriguing ideas that print should consider, spearheaded by the Big G and other online advertising outfits:

• Form an ad network: Instead of relying on their own ad sales division, newspapers and newspaper networks should outsource ad sales to ad networks that can fill their inventory across all available newspaper inventory.  (Of course, the ad network typically shaves off a comfortable margin for itself, something the newspapers would probably not like to lose. They can form the ad network themselves, like the airline industry did with Orbitz.) This will allow them to sell inventory on metrics meaningful to advertisers—not just geo, but on demographic, vertical/topical bases as well. And it reduces advertiser friction and builds campaign scale.
• Create a new sales force: Maybe Google isn’t the place a small, local advertiser would think of turning to to start a campaign. Jarvis suggests a “citizens sales force,” the sales side of citizen journalism. He might have a point, since a lot of the print long-tail are not going to understand AdWords any better than those online long-tail advertisers that need services like ReachLocal, Yodle and OrangeSoda. He also suggests that local media can probably retool to sell inventory beyond their reach, if there were such a network in place that could serve them.

The Los Angeles Times announced recently that revenue from advertising on its online division covered the paychecks of its entire editorial division—online and offline. This is certainly heralding the value of online traffic, which newspapers have long thought cannibalized their print traffic with dubious financial return. The hidden story is that they’ve made a new advertising model work. Maybe it’s time they apply what they’ve learned online to revive their flagging print division.