You are currently viewing How to Split Test Your Advertising

How to Split Test Your Advertising

Knowing how well your Marketing Communications converts is a critical part of any business. Be it promotional material, email communication, ad design and so on. Do you ever wonder if the marketing communication you publish is as effective as it could be? As a business owner, you likely put a large amount of money and effort into your advertising initiatives. You could be losing out on sales and incoming profits if you don’t know what is converting best. That’s where split testing your advertising can come in handy. If you aren’t getting the ROI from your efforts, do you know where to start to look to make changes? The right graphic design, the right image selection, a well crafted eye-catching headline are all critical components required to get your audience to take action, however, what converts is often not always what we think will convert. It is important to get the data – stop guessing and start testing! In the following pages we will discuss the importance of Split Testing, some key definitions along with strategies and tactics you can implement right away to start creating higher converting marketing communications that your target customers will take action on.   1

What is Split Testing?

  Split testing is the process of comparing multiple versions of a piece of advertising (digital banners, flyers etc.) OR marketing communication (landing page, email etc.). A & B testing is the same process but with just A and B. Where split testing can include splitting traffic over multiple variations of the same item tested a/b/c or even d testing. One important distinction is that Split Testing and A/B Testing focuses on a single change or variation. For example, ONLY the ad copy is changed and not ad copy, image and headline. The later would be an example of Multivariate Testing. For the sake of this guide, we will focus on Split Testing and not Multivariate Testing, however, it’s important to understand the distinction between them because jumping into Multivariate testing before you are ready can be a disastrous waste of time.  

Split Testing Versus Multivariate Testing Section

  Split Testing is sometimes confused with multivariate testing, however, hey are actually quite different.  Think of multivariate testing as a more complex version split testing with many more variables involved in the testing process.  This can be a great method for testing overall website design to determine which is the most user-friendly, and which website layout factors could be affecting a website’s conversion optimization.  Multivariate testing is very powerful, but the huge drawback is the amount of traffic that’s necessary to receive results.  To conduct a multivariate test that’s statistically viable there must be enough traffic to test each of the different variations and combinations involved while in a relatively short amount of time.  This can be difficult for even some of the larger websites to do because it takes a great deal of time and energy. It is also essential to have reliable testing tool to accurately gather and display the data collected thru testing.   Multivariate testing, opposes the traditional scientific notion. Multivariate testing is the process of testing more than one component on the web site in a live environment. Essentially, it can be described as running multiple A/B/n tests on the same page, at the same time.”   – Lars Nielson of Sitescore   Split-Test  

Why Split Test?

Simply put – Split Testing allows you to learn WHY visitors are not converting. Using a general average – of every 100 site visitors, 3 will take action, using my grade 3 math that’s at minimum 97% of traffic wasted. Split Testing allows you to broaden the top of this funnel and helps you convert more of these visitors into customers. It is a great and accurate way to determine the best performing marketing communications.

Benefits of A/B Tests:

  • often take significantly less time to create and implement than Multivariate Tests
  • get meaningful results in quicker time
  • usually see larger gains and bigger differences in test results



How to Split Test?

Above we briefly explained the importance of only testing 1 variable at a time so the most accurate data can be gathered. Accurate data = Actionable Data.

Step #1 – Testing Preparation:

Before you start testing it is important to know your current data, observe current user behavior and then build your hypothesis…then you can test your hypothesis.

1a. Know Your Data:

Depending on what piece of marketing or graphic design you want to test – study the current traffic and conversion data. Look at traffic to the webpage or current email open and click rates if testing an email. Find the areas in your conversion funnel with the highest bounce rate and start here.

1b. User Behavior:

If it is not clear what is stopping users from converting, it will he difficult to know what to test. Dig into heatmaps to learn where users click and spend time on our page or email. Conduct surveys or pull data from past surveys. Use all available data to gather insight into what you feel is stopping users or potential customers from taking the next step in the conversion journey.

Step #2 – Create your Hypothesis:

From the data observed above you should now have a hypothesis of what is not working and where exactly the problems are in your conversion funnel. Using an email campaign as an example – perhaps you’ve been able to determine that your email CTA needs to be tested because your email campaign has great delivery rates, good open rates and yet click thru rates from email drop off considerably. This indicates there is something not working with the email’s CTA. Know what you will do with the data you receive from your test. To avoid focusing on smaller elements with little or now impact, focus on the biggest impact items first.

Step #3 – Testing your Hypothesis

Continuing with the above email example – now it’s time to test your CTA. Further narrow down WHAT part of the CTA you want to focus on testing first. Is it the actual CTA offer? If you only have 1 offer, then write a second versions of CTA copy and test it against the current version.

REMINDER: All other components of these emails remain the same (email headline, email copy etc). If you start changing these you will loos insight into WHAT is contributing to the change in results.

Step #4 – Analyze Results

After releasing these different emails to your lists you’ll be able to see how each has performed against the other and determine which Call to Action has generated the most clicks, thus you’ll have your winner. If you’ve hit your target Click Through Rate, then use this CTA and move onto a different variable to test (perhaps email ad graphics or subject line). If your results have not determined a clear winner keep testing other CTA variations such as CTA button colour, size or location.


Split Test Result


Step #5 –  Action Results and Test Again

This information will help to make further marketing decisions in knowing what paying customers respond to in your advertisements.  After you’ve gather statistics on this split test, you’re able to then create another A/B test to continue working towards finding that sweet spot of the best advertisements.

Additional Strategies and Tactics:

We’ve used CTA and email template testing in the above example, however, this concept can be applied to a wide variety of marketing initiatives such as landing page design and layout, social media graphics, ad design, brochure design, sales letters, banner design and ad campaigns….you get it and the list goes on. Below are some tips to take into consideration as you progress along your Split Testing Journey.

Important items to test include:

  • Headline
  • Sub headline
  • Ad copy + main body text
  • Graphic Images
  • Call to Action text
  • Call to Action buttons
  • Keywords
  • Testimonials
  • Social Proof


  • Colour of buttons
  • Types of images
  • Layout and location.

It helps to get into the psyche of your clients and find what appeals to them the most.




Length of Time for a Successful Split Test

Split testing can be a balancing act when it comes to the amount of time that you want to run each individual test.  Too short of a timeframe, and you may find that you haven’t collected enough information for conclusive results and you’ll have no way of knowing whether or not your results would have remained the same.  However, too long of a timeframe for your split test can be equally ineffective because it will prolong your ability draw conclusions and action changes from these results.

TIP: A good rule of thumb when email testing, you’ll want to have a database of aproximatly 1000 valid emails before conducting A/B testing. If your website receives <100k traffic/mo it is best to do A/B testing. If your business is small or just starting out and you can’t wait to hit these numbers – keep split tests to just A/B and strive to hit the above targets before putting significant $ behind any decision you make from these test.

There are a number of factors that will help to determine the length of the time for a successful split test.  The amount of traffic that you receive will help to determine the amount of time to spend on one aspect of your split testing.  Decrease the amount of variables that are being tested to help decrease the amount of time necessary.  Finally, you’ll want to increase the expected improvement by making bigger changes.

The good news is that if this has left you a little lost, there are several websites that you can use to determine the proper length of time to run your split test.  You’ll just need to enter in a variety of information, and each tool may require different information.  You can expect to enter in the rate of conversion on your control page, what you’re expecting your conversion rate to improve to, how many variations in the test, average number of visitors, and the percent of visitors that will be part of the test among other potential questions.

Available Split Testing Tools

There are a number of split testing tools and programs available online that can make split testing a possibility for just about any company.  One of the first tools that you may already be using is Google Analytics.  This is a good starting place for your split testing, and is free, so that makes it perfect for many businesses.  However, it can take a moment to learn how to best use this service for your split testing needs.  There are some other tools available online that are a bit more user-friendly, but they typically cost a little more.  One of the more affordable tools is Crazy Egg which offers a free trial and is only nine dollars a month.  Unbounce is a great tool if you plan to do Landing Page conversion testing, they offer a full 30 day free trial and their software integrates with many popular digital marketing tools. There’s also Visual Website Optimizer, which offers a test length tool that you can use for determining that sweet spot of how long your split test should last.  Other services to check out includes Kissmetrics, Unbounce, and Optimizely.

Unlimited Graphic Design and Split Testing

Advertising can be expensive, and that’s a great reason for you to turn to split testing, but split testing can also run deep into your budget if you’re not careful.  One way that you’re able to deal with keeping costs down for your split testing needs is to utilize an unlimited graphic design services.  This means that you’ll be able to have unlimited variations on your ad design and marketing communication for a predictable flat fee that you’ll know in advance. At Design Productive specialize in helping small businesses create the marketing communications they need to grow. You won’t have to worry about having the budget or resources to create all the design variations you need to meet you’re testing requirements, you’ll have unlimited access to make changes and variations to your artwork with a quick 2 day turn around.  This can be a huge savings for your campaign that will let you focus your budget in other areas. Or use Design Productive as your on demand project design team – outsource your Split Testing variations to us while your in-house design team stays focused on other critical product and marketing communications.

Split Testing Best Practices Section

– Test often rather than only doing one test and stopping.

– Test one thing at a time to help reduce your campaign time and to better pinpoint what’s causing the success of your winner to help you duplicate it for future campaigns.

– Test the right amount of time to not skew your results.

– Data collected is much more important than any gut instincts that you may have on what your clients respond to in an advertisement.

– Remember that you’re not just working on increasing your leads, but the quality of those leads.  It’s only half the battle to get people to visit your site, click your ad, or whatever you’re split testing.  The rest of the battle is getting to qualified leads that will follow your call to action to sign up for your newsletter, purchase your item, contact you for a consultation, or whatever you need your leads to do.

That’s split testing in a nutshell.  Hopefully, this guide will point you in the right direction for reaping all of the benefits you are seeking from finding the best mix of advertising features to get those qualified leads flowing into your business.  Split testing is a very valuable tool that can often be underestimated when it comes to the power that it provides your company.  There’s a lot to be learned when it comes to split testing, but one of the most important lessons to keep in mind is that trial and error is part of the learning process.  Use the different tools and software available to you to keep your scientific experiment of A/B testing moving forward.

Download your free split testing book here. Get over 100 split testing results free.

Leave a Reply