There’s no manual or rulebook that explains how to optimize your website’s monetization. As a result, publishers end up either switching between ad layouts to see which one results in the highest revenue or just sticking with what’s working for them. Or worse, they simply go with their intuition. But there’s a problem with these strategies. Without proper testing, publishers never get the empirical data on what actually works (and what doesn’t).
To navigate these scenarios, it is crucial that publishers assess their users’ reaction to their ad setup, especially since a wrong change in your ad display can be detrimental to web traffic. This is where A/B testing comes into play.
What Is A/B Testing?
A/B testing is the comparison of performance between variants. Rather than implementing changes based on bias or intuition, A/B tests provide concrete, granular data that allows publishers to gauge the outcome of any decision they make in terms of ad formats, ad layouts, or even new demand partners.
With this data, publishers can then optimize their site layer by layer without negatively affecting user experience. Perhaps adding another banner ad won’t affect your bounce rate like you originally thought, or maybe adding a video content unit in some of your articles will increase your users’ engagement and time on site. You won’t know unless you run A/B tests.
In ad tech, the scope of A/B testing has expanded far beyond what was possible just a few years ago, meaning there really is no limit on what publishers can and should test. Let’s take a look at some of A/B testing ideas to help publishers get started.
1. Ad Layout Testing
What to test: The ad combination that generates a higher revenue while maintaining a great user experience.
It is important to test different sets of ad layouts on different kinds of users. This helps the publisher determine which ad layout works best for each scenario: mobile users vs. desktop users, local-based users vs. global users, short articles vs. long articles, the list goes on.
Nowadays, there are ad layout testing platforms, like ReigNN, that make it possible to create variations of whole ad layouts with specific combinations of different ad sizes, placements, and formats, allowing the publisher to match the best ad layout for each specific type of user.
2. Content-Type and Structure
Now that you have a better idea about which ad layout works best in terms of user experience and revenue, it is important to optimize other parts of your website. You can use A/B testing to test not only the actual structure of your page, but also different types of content.
Perhaps long-form content isn’t as effective as listicles. Or, maybe, beginning each page with a picture instead of an opening paragraph results in higher engagement. The data collected from these A/B tests enable publishers to make decisions regarding their long-term content strategy and, therefore, their bottom line.
4. Demand Sources
Seeing that your revenue is directly correlated with how well ads perform on your site, it is wise to test which demand source generates the best results for your inventory. Perhaps you’re debating between different ad networks, or you’re unsure if your users will be tolerant of a new format. Instead of making a decision based on bias alone, A/B testing different demand partners provides concrete data about the incremental value of each demand partner.
The Anthem for A/B Testing: “Theory, Test, Result, Repeat”
It’s important to remember that A/B testing is a process. Publishers need to first come up with a working hypothesis and then choose which percentage of users they wish to expose these changes to.
The length of the A/B test depends on several factors. Different demand sources have different traffic levels, while conversion rates will also have an impact. If your site already has high conversion rates, then you need less time to reach statistical confidence.
It also depends on which variant you’re testing. The more drastic the change, the less scientific you need to be. But when you’re dealing with minor changes, such as microcopy, more data will be required to prove its negative or positive impact.
Analyze your results, increase the percentage of users (if the trend is positive), re-analyze, and repeat. Once the change has reached 100 percent of your traffic, it’s time to start the process over with another hypothesis—adding another layer to your site’s monetization.
What to Keep in Mind
A/B testing is an effective tool when optimizing your site’s monetization. However, there are some important points to keep in mind:
- Remain calm and give your testing the time it needs. It would be wrong to expect substantial data within two days.
- Optimize layer by layer. You might have several things you wish to test but testing them all at once will make it harder to determine which parameter had a positive (or negative) impact.
- What worked two months ago may not work today. The data collected from A/B tests are based on how users respond to change. However, traffic volume and user behavior are constantly changing, meaning there is no end in sight for optimization possibilities.
- Apples to apples. Make sure your test is running in the same environment so the comparison will be valid.
