You have the option with Google AdSense (and other contextual ad services) to tweak your text ad units’ background colors and border colors, in an effort to boost their visibility to users and encourage clicks. Are you absolutely sure you know what color combinations work best, and which work worst?

We ran a test on HubPages, and looked at the sidebar ad zone installed to the right of the main body of content. We set YieldBuild to try a large number of background and border color permutations. We expected a white background with white border (i.e. blended into the background, since our page background is white) to perform best.

It didn’t do bad at all. At 3.58%, it was far better than the worst combination, which fared a miserable 0.89%.  However, it was bested by a light-yellow background with a dark-gold border.

The yellow & gold combination came in at a little over 4%, or 13% better than the white-white combination.

What else is worth noting is that there doesn’t seem to be a clear, simple pattern. Looking at the results of this test, you can’t say that light and subtle, or bright and garish, work better. Who would have thought that a light-yellow background with dark-gold border would perform so well, while a medium-yellow background and light-yellow border would perform so poorly?

A very light beige background also performed substantially more poorly (2.46% CTR) than a white background.

The lesson here is that it’s often difficult to rely on rules of thumb if you want to tweak your ads’ background and border colors using intuition alone. The only way to arrive at the optimum combination is to test, test, and test. 

A study conducted by Starcom USA, Tacoda, and comScore suggests only 6% of the online population are responsible for 50% of the clicks on display (image) ads. What’s more, the typical profile of such a “Natural Born Clicker”, 25-44 years old and with an income of under $40,000, do not have the spending power of the average Internet user, much less in proportion to their click activity. They also tend to visit auction, gambling and career services sites more than the average Web surfer.

So a click is not necessarily a click. PPC advertisers need to calculate a campaign’s backend conversion (the rate at which leads convert to a sale) in order to complete the picture of campaign ROI. This study doesn’t call into question the value of PPC advertising as much as it reminds advertisers that determining the success of an online PPC campaign requires looking beyond just one step in the customer conversion process.

You might imagine that choosing whether to put a skyscraper in your left or right rail, or your 728 x 90 banner at the top or in your footer, matters when it comes to their performance. You’d be right. AdSense gives you some clues with its heat map—specifically, ads embedded within your content, or slightly veering to the left-hand side of the page, will perform better than ads relegated to the bottom or right.

But does the precise position of an ad unit matter? Does shifting a banner or skyscraper a few pixels to the right or left, or up or down, make a difference?

Yes, it does.

In the same test that showed in our last entry that 3 ads per page does not necessarily outperform 2 ads per page (and, occasionally, it actually underperforms), we looked at the alignment of specifically three zones on HubPages:

• the title banner (a zone below our header and above the main body of content)
• the skyscraper (usually vertically-oriented ads, typically placed in the left- or right-rail)
• the bottom rectangle (a zone below the main body of content)

(Note, YieldBuild zones are not precise designations for specific ad units—they are just areas that allow YieldBuild to optimize a wide variety of ad unit types and sizes.)

The results were interesting. In each case, a specific alignment of ads did, in fact, make a difference and outperformed other alignments:

In the specific case of HubPages, title banners and skyscrapers performed best when aligned to the right of their zones, while centered bottom rectangle units seemed to outperform those shifted to the left or right.

Would the same be the case for your site? It’s impossible to say. What this does say that, ideally, you should test not only the ad units and layouts on your page, but the ad units’ exact position as well, if you truly want to maximize your ads’ performance.