Connect with us

Affiliate Tips

Split Testing Best Practices & Pitfalls to Avoid

Let’s face it; we’re all here to make money. Every business no matter what it is has one thing in common which is the fact it needs to monetize traffic. It really doesn’t matter if you’re running a blog which generates its revenue with AdSense or if you’re an affiliate or a product owner, turning a profit is essential to staying in business.

Not everything we do will be a grand slam right from the start; this is where website split testing comes into play. Today we’re going to discuss a few best AB split testing best practices to help you maximize your traffic in any niche or business you happen to be in as well as go over a few common mistakes to avoid so you can skip all those time & money wasters and jump right to the finish line.

What you should do

User experience is important. With so much information flooding our senses at any given moment of the day it’s no wonder the average person’s attention span is just 8 seconds.

The honest truth is that most people just don’t like to think too hard. We’re so used to having what we want when we want it and this includes solutions to any curiosities or problems we may have. The challenge this poses for marketers like us is how to create a user experience which is both easy as well as fulfilling.

It’s up to you to rise above the competition by making your webpages a seamless and enjoyable experience.  Remember that you only have about 8 seconds, in some cases even less to grab someone’s attention and make them want to know more. If you are finding that something is not working quite right then it’s necessary to test different details of your webpages until they start performing better.

This can be anything including headlines, font-type, color schemes, angles, images calls to action and more.

Maybe the placement of a certain button would be better served in a particular area of the page compared to where it was before. I recognize all this can appear overwhelming at first but you should keep in mind that you have the internet at your fingertips and all it takes is a bit of Google savvy and a willingness to research to find the solutions to your problems. Chances are if you’re wondering how it’s done then someone else has already found the solution and all you need to do is find it and duplicate success.


Besides the overall user experience there are a number of other details you should keep in mind:

  • Tracking programs are essential for conversion tracking and split testing. We use CPV Lab to do our website split testing since it is extremely easy to split test both ads and landing pages. It is simple and allows you set a percentage for every new page you want to test. So the first test we will often do is a 50/50 test on 2 landing pages. The winning page out of that test will then be called the “control page”. After that all new pages will be tested at 15% while the control page runs at 85%. Tracking programs such as CPV lab, Prosper 202, or voluum are essential in getting the data you need to make informed decisions. Of course there are many other solutions such as Google Analytics, or even by using the traffic source pixels. All of the places have extensive training inside their platforms on how to do proper setup. Keep tabs on stats such as page views vs. sales to see just how well your offers convert. Facebook is a popular platform which is closely associated with tracking pixels but the fact of the matter is that there are tracking pixels used in just about anything which keeps track of any type of conversion or website stats. All a tracking pixel really is, is just a small snippet of code placed on a page in your website which allows it to track details such as page views, sales, email opt-ins and so on back into your tracking system or traffic source.

Here is a recent split test we did on 2 pages for a gaming type of offer: (click to enlarge)


As you can see we have all the information form this to make an informed decision on which page should become the “control” page. The main indicators are the conversion rate and ROI. But in some cases we may find that the landing page click through rate is also a very important metric to earning more money. For example, while one page may get a better ROI, the other page may be getting more conversions. When making a decision on which page to kill, you need to know what your end goals are.

  • Split test your funnels from the top down. It only makes sense to begin your testing with the page which receives the most traffic first and then work your way down once you’ve optimized that initial page as much as you can. Each page in the sequence will naturally see less & less traffic but by optimizing from the top on down you will gradually see an increase in that traffic all the down your funnel so start from the landing page and work your way down to the final offer. When we find a good “control page” we then simply copy that file and call it version B and make minor changes. It is easier if you change small things because then you know what made the difference. If you change 10 things on a page and run a split test, you wont know which of the 10 things contributed most to that change in conversions.
  • Keep each of your campaigns separate. It might be tempting to try painting over all your campaigns with a broad stroke but the fact of the matter is that each campaign is unique and should be treated as such. What works for one may not work in the same way for any other. For example you should be running separate campaigns for each country, separate campaigns for mobile and desktop, separate campaigns for  different angles, separate campaigns for different traffic sources.
  • Traffic sources matter a lot. It’s safe to say that where you pull your traffic from will have a dramatic effect on how that traffic interacts with your websites and offers. When performing your split-tests it’s vital to keep in mind that you need the right traffic directed to the right offer in order for any of your tests to matter otherwise you’re just spinning your wheels and getting nowhere.
  • Test completely different designs and layouts. While this may sound like a contradiction to the above statement of making small changes, usually in the beginning phases of campaign development you really need to test several drastically different angles and ads to see which one has the most potential to get more conversions. Sometimes it’s tempting to want to stick with one specific look for your webpage, or to stick with one angle. If you’re just not seeing the conversions you’d like then it may be time to scrap that old layout in favor for a totally new and completely different one! Put the old layout up against something completely different and then another and yet another layout and see how each one performs. Take care to make intentional and educated changes to each of your layouts and you just might find your winner.

The list of suggestions could go on and on but these few items listed above are some of the really important ones which will get you pointed in the right direction. It’s time to discuss some of the things you should stay away from when split testing your webpages and offers.

What you should NOT do

We’re all human beings here which means we’re prone to call favorites and get attached to things. Though I can’t blame you for falling in love with that landing page you’ve spent hours upon hours perfecting there is one universal truth which rules above all else and that is the simple notion that the market wants what it wants and it likes what it likes.

If you put a page in front of a steady stream of targeted traffic and your offer is just not converting it’s time to head back to the drawing board my friend- I understand your pain but sometimes these things happen.

Here are some things you should stay away from when split testing any offer:

  • Ending your A/B tests too early. If I had a nickle for every time an affiliate told me “I ran a test and it didn’t work” I’d be a billionaire. No offense but the majority of affiliates especially in CPA have no idea what a real test entails. The biggest mistake they make is in the sample size of their tests. Many people fail to test all times of the day, they fail to send enough traffic to make a decision, and they don’t plan their budget accordingly to get the best results from a real test. The fact of the matter is that there is no magic number to refer back to when testing so it becomes a matter of experience and a willingness to wait it out. Different factors such as ad-budget, sample size (how much traffic you’re getting to your pages) and the number of components you’re split-testing all matters.
  • Not testing for an entire week. Unless you have a very good reason for restricting certain days of the week you would wind up with a more complete picture if you tested A/B splits for a whole week at a time. This ties into the point made above about ending tests too early. If you don’t yet have the budget to run tests for this long it’s understandable to want to end things early and you may certainly get lucky by running a whole bunch of options back to back- but you will be relying mainly on luck and this is no way to conduct your split tests.
  • Split testing with too little traffic. Something that is absolutely necessary when running tests between two options of a single sales page (or whatever) is traffic and plenty enough of it. Think about it this way… if you’re running a campaign which only makes a handful of sales per month and you decide to test a variation of that sales page which converted at just 10% to 15% better, how would you know it wasn’t just by luck? Instead of running tests for weeks or months with something that’s only pulling a handful of conversions per month it may be wiser to just test for a shorter period of time and stick with the winner, no matter how small of a margin that increase is. The idea here after all is to increase our ROI, not decrease it by running “money-wasting split tests.”
  • Running pointless tests is pointless. Yes, an obvious statement right? You’d be surprised how many people run tests for the sake of just running tests. Though you’re free to do that if you wish, it is a complete waste of time and money. I mentioned before that you can test anything from the colors of your site to the font type and this is true. However, any test you perform should be obviously aimed at getting to your goals faster. In our case, most of our campaigns have these goals > increase ROI or increase options. The idea is to make changes that will cause people to perform the desired action intended.
  • Giving up as soon as a test fails. The entire point of split-testing is to find winners and losers. Failures are going to happen and that is a good thing; it lets you know what not to do and helps you learn what should be done. The more you test for a certain market and traffic source the more you’ll learn about those sources which means you’ll eventually be able to develop a certain premonition about them. Everything you do is a learning process and once you hit a certain threshold you’ll discover that you’ll be making educated decisions based on all the data you’ve collected over time. This is not to say you should ever get too cozy in your own experience as the market is always changing so keep on your toes.


Split testing is an extremely important part of any advertising campaign and should be considered when starting on something new. It’s what allows for higher conversions and that in turn grants you access to more traffic through a better ROI (return on investment). In most cases here are the 3 main things we will split test:

  • Offers between networks – did you know some offers convert better on different networks? Its a known fact and has to do with pixels, and throttling. We won’t cover those things in this post, but know that offers convert different know matter what network you are on.
  • Landing page variations – obvious one but we are always testing our control page against another page at 15% new page, 85% control.
  • Ad variations – this can be tricky because some ads work with certain pages. So you have to be mindful that if you are testing a ton of ads and a ton of landing pages at the same time it could get difficult to know which ad works with which landing page. But you must test new ads constantly as well.

Here is a sample of a recent split test of 4 ads we ran on a native ad campaign on Revcontent:


Can you tell which one is the winner?

Please post any comments or questions below.

The information here hardly scratched the surface but it is a good place to start. If you’d like to learn more about split testing as well as see our exact campaigns and strategies we run, check out the member’s area of the Powerhouse Affiliate.

There you’ll learn up to date techniques through our official 12 step FAST START affiliate training program. You’ll be learning from 6-figure affiliate marketers who practice what they preach. Everything is laid out in simple to understand, step-by-step plans that anyone can follow along with.

You’ll have access to our exclusive member’s only forum where any question you may have about affiliate marketing will never go unanswered.

You’ll learn how to:

  • Generate traffic
  • Enter the CPA market and crush it as an affiliate
  • Build that email list you’ve always wanted giving you full control over your very own traffic source
  • Learn how to track campaigns and conversions to optimize your ROI like a pro
  • Increase your conversion rates & so much more

Head over to the Powerhouse Affiliate’s members area and discover what awaits you on the other side.

Click to comment

Leave a Reply

Your email address will not be published.