AdMoolah News and Views

Experimentation is the only way to find out what ads work best on your site. We have talked about different ad colors, positions, and formats. There are some general tips that can be given for each of these topics, but the only way to learn what really works on your site is to try different variations.

Custom Channels

Google’s custom channel feature is the key to experimenting with ads. This allows you to view reports for one particular ad unit. For complete details about setting up custom channels and viewing reports, see Google’s channel support page.

An Example

Let’s say you are curious about which color palette works best on your site. For example, you may want to know if a “blending in” color palette works better than a “standing out” color palette.

This experiment is divided into two phases.

Phase 1: Testing the “blend-in” ad setup.

1. Create a new custom channel called “blend-in”.
2. Create a color palette that blends in with your site.
3. Create the code for the ad layout using the “blend-in” channel and color palette.
4. Place the AdSense code on your site.
5. Run the ad for a chosen amount of time.

Phase 2: Testing the “stand-out” ad setup.

1. Create a new custom channel called “stand-out”.
2. Create a color palette that will stand out on your site.
3. Create the code for the ad layout using the “stand-out” channel and color palette.
4. Replace the AdSense code on your site with this new code.
5. Let this ad run on your site for the same amount of time.

Now you will be ready to see which color palette performed better. You can use AdSense reporting to view the “blend-in” and “stand-out” custom channels side-by-side. You will be able to easily tell which one received a better click-through rate. You can apply this same technique to ad positioning and formats as well.

Length of Experiment

I didn’t specify how long to run each ad setup on your site, because the length of time depends on the level of traffic your site receives. More traffic can shorten the length of the experiment. To determine an ideal length of time to run the ad setup, the click through rate should stabilize at the end of each phase.

Consistency

An important rule is to vary only one thing at a time. Don’t compare a blend-in banner with a stand-out skyscraper. If the blend-in banner had a 50% higher click-through rate, how would you know if it was the ad format or the color palette that made the difference?

It is also important to make sure that the days for the two different phases of the experiment are similar. For example, don’t run the first phase on Saturday and Sunday, and then run the next phase on Monday and Tuesday. Click-through rates typically vary from weekends to weekdays. Also, be wary of holidays and the beginning and end of the month. Selecting consistent time periods ensures accurate results.

Ad Rotation

Instead of running your experiment in two different phases, you can run both ad setups simultaneously by using ad rotation. This avoids the potential problem of choosing inconsistent time periods for each phase of the experiment. Ad rotation is achieved by creating a program that will randomly select which ad setup to show to the user each time the page is viewed. In order to be able to compare the results, be sure to use separate custom channels for each ad setup as explained earlier. The disadvantage of this technique is that it requires some server side programming to implement. If you are comfortable with this type of programming, I would recommend ad rotation.

Here is an example of a PHP script that does simple ad rotation:

function getBannerAd() {
  $adtype = rand(0,1);
  switch( $adtype ) {
    case 0:
      echo '< -- Insert Google code for ad setup 1 here -->';
      break;
    case 1:
      echo '< -- Insert Google code for ad setup 2 here -->';
      break;
  }

Summing-Up

Trying different ad setups is a great way to boost your AdSense revenue. By running well controlled experiments, you can learn how to increase your AdSense earnings. Don’t be afraid to try something different, you never know what might work best!