8 Tips To Run an Effective AdSense Experiment

If you’re one of the many publishers and bloggers who relied on display advertising through AdSense as a primary source of revenue, there’s a pathway to higher earnings by determining the optimal AdSense settings for your site. As we’ve covered previously, there are countless opportunities to tinker with AdSense settings and layouts that will result in different payouts. If done correctly, this process should ultimately help you generate more revenue from each pageview on your site.

Here are our tips for getting the most out of your AdSense experiments.

Tip #1: Get a Baseline

Before you can start creating variations, you need to set a baseline–the numbers you’ll be trying to beat with each experiment you run. If your site gets enough traffic to make display advertising a viable revenue stream, you probably have this already; it’s just a matter of pulling the historical performance data from AdSense.

Which leads us to…

Tip #2: Keep Meticulous Records

In the haste to improve earnings, keeping track of your tweaks probably seems like a boring administrative task (because it is a boring administrative task). But it’s important to document what you’ve tried and the extent to which it worked or didn’t work. Keep an “AdSense Journal” of every test you run, including details on the exact inputs (i.e., the settings you used) and outputs (the money you made and CPMs you generated).

Tip #3: Know Which Metrics Matter

When running an experiment, be sure you know which metrics you’ll be evaluating afterwards to determine the impact of your changes. It may sound obvious, but total revenue is irrelevant since that depends on the number of impressions served.

Cost per click (CPC) and CTR are also meaningless metrics in isolation, since each tells only half of the story. What you do want to consider is the revenue generated per 1,000 pageviews; this metric is a standardized reflection of what you earn for a fixed amount of traffic.

If you have other sources of revenue such as affiliate links or lead gen widgets, you should also take into account the impact that any changes to AdSense has on these.

Tip #4: Pick a Variable

Optimizing your AdSense account will require multiple rounds of testing–which takes time. Don’t try to get it all done at once.

Here’s the wrong way to approach this:

• Variable: 300×600 ad unit, text and images, blue text, white background, Arial font

Suppose the control delivers a CPM of $5.00 compared to a $6.00 CPM for the variation. You know that the variation worked better, but have no idea which variable really drove the chance. It’s possible some of the changes made hurt your earnings, but were offset by the positive impact of others.

The goal of an experiment is to determine the impact of changing a single variable–not the impact of making wholesale changes. Change one variable at a time, or else you’ll end up with no idea which one actually drives the changes.

Tip #5: Make Sure Your Conclusions Are Significant

In the rush to arrive at the optimal AdSense settings, there can be a temptation to draw conclusions from experiments before they’ve had a chance to deliver meaningful results. Specifically, monetization experiments need to accumulate a large enough sample size in order to be confident that the results you’re seeing are truly indicative of what can be expected going forward.

Make sure the conclusions you reach are statistically significant, or else you risk embracing the wrong optimizations and hurting yourself in the long run.

Tip #6: No Change Is Too Small

Don’t feel the need to make wholesale changes to your ad units to see an impact. Sometimes the subtle changes–using a different font or moving from a dark blue to a light blue for links–are all it takes to see a material change in earnings. There’s nothing wrong with small, precise steps; as a result, you’ll know the exact impact of changing a relatively minor variable.

Tip #7: Go Against Your Gut

Inevitably, you’ll come into the experimenting process with preconceptions about which variations you think will “win” and which you think will “lose.” There’s often a hesitation to run experiments that include layouts or creative combinations that we feel will be ineffective (often because they’re ugly).

There’s an old football maxim, “That’s why they play the game.” That applies to display ad monetization as well: there’s a reason you run the experiment. Sometimes, the underdog will win and surprise everyone. There’s no shame in having your gut proven wrong–especially if it makes you more money in the process.

Tip #8: Don’t Be Discouraged by “Losses”

When you start testing AdSense layouts, you’ll probably be expecting immediate wins and constant improvement. But that’s not how if works. In fact, there’s a very good chance that your first AdSense test will lower your earnings.

And that’s completely fine. Don’t get discouraged when consecutive experiments result in big CPM declines. That shows that the process is working; the changes you’re making are having a meaningful impact on earnings–even if it is in the wrong direction initially. Take it as confirmation that the tests you’re running can produce meaningful results, and not as an indication that you should quit.

READ MORE: monetizepros.com