Split testing (or A/B testing) is creating two versions of digital content and running both simultaneously before audiences to measure which one performs better based on metrics like website visits, sign-ups, clicks, or purchases. It best serves virtual businesses like business coaching, ecommerce, digital marketing, and software companies that want to increase conversions.
If you’re looking for a great all-in-one split-testing tool, check out Freshmarketer. It provides A/B testing as well as complementary tools like form optimization, heatmaps, and session replays so you can see why viewers interact with your website as they do. Freshmarketer is $99 per month for their A/B testing package. Best of all, they offer a free 30-day trial to get you started.
What Is Split Testing & Why Is It Important?
Split testing involves learning what digital elements (on a website, for example) hinder progress on customers’ purchase paths, then creating a different version of one element only (like a form) to see if it performs better than the original version. These versions are typically sent to equal audience sizes to accurately determine which best boosts conversions. To help determine which version performs best, small businesses can use recordings of user behavior, heatmaps that show where users click and scroll on a page, surveys, and other A/B testing tools.
For successful split testing, first research to find out what website elements keep visitors from converting by observing how visitors interact with your website. Heatmaps, recordings of user engagement with a site, and surveys can help you conduct this research.
Here are a few tools to help you research obstacles to conversion on your website:
- Funnel Analysis – Funnel analysis is software that you can use to map out customers’ paths to conversions on your website, then measure which ones most keep customers from advancing toward a purchase.
- Heatmaps – Heatmaps allow you to see a visual representation of where customers do and do not click and scroll on your website.
- Session Recordings – Session recordings are actual recordings of one person’s entire visit on your website, from what pages they used to enter your website to what they did from there and, ultimately, when and where they exited your website.
- Surveys – Surveys allow you to ask customers why they did or did not perform certain actions on your website (like clicking on a link or filling out a form).
For research purposes, funnel analysis tools helps you understand where people drop off their purchase paths. Heatmaps help you understand what elements do and do not garner clicks and engagement at drop-off points. Session recordings and surveys give you deeper insights into why people aren’t engaging. Once you’ve pinpointed problem elements and have developed a hypothesis as to why, you can create versions of those elements that you believe could solve the issue.
A drag-and-drop editor (also known as a what-you-see-is-what-you-get, or WYSIWYG, editor) is a platform with editing tools (like font, color, and layout options) you can use to easily design variations of your problem element, then use A/B testing tools to launch those versions and split your traffic among equally sized audiences to measure which best converts. Dashboards then show you graphs and percentages of how many converted via each version so you can launch the winning one.
In the end, split testing looks at the portions of your website or virtual marketing that keep people from converting (e.g., taking an action), then tweaks them until they produce the most possible conversions. This allows businesses to offer the types of website experiences Google looks for to rank them highly and, most importantly, boosts revenue with more purchases.
Who Split Testing Is Right For
Split testing should be a priority for small businesses that rely on their websites and digital marketing to drive sales. Examples of these types of businesses include ecommerce businesses, travel and hospitality booking sites, digital marketing companies, and software companies.
Here are some types of businesses split testing is right for:
- Virtual Business Coaching Businesses – Virtual coaching businesses rely on their websites to get potential clients to contact them or schedule a consultation via an online form. Split testing allows these businesses to understand what keeps their online customers from reaching out for a purchase or consultation.
- Ecommerce Businesses – Ecommerce businesses rely on their websites to guide potential buyers to the right products for their needs and ultimately purchase them. Split testing helps these businesses ensure there are no hang-ups in this process.
- Digital Marketing Companies – Digital marketing companies often serve local customers as well as virtual ones across world. This means their websites must do a superb job of offering all the information their customers will need to select their business as a provider. Split testing allows these businesses to observe where they aren’t providing sufficient or persuasive information, then tweak it to boost conversions.
- Software Companies – Software companies often serve a few target markets and have unique solutions for each. Split testing allows software companies to observe what might keep each market from downloading products catered to their needs, then tweak each purchase path to produce the most conversions.
Example of Split Testing
Arenaturist.com, a hotel and resort booking website, relied heavily on their booking form for revenue. The form was used by visitors to book their reservations. As a result, the site featured a prominent form asking people to input their destination, booking date and time, number of people traveling, and any promo codes so people could book their next vacation easily. But the form wasn’t performing well.
To boost conversions, they used A/B testing to test one element of the form—its homepage placement. Their original, or control, form was horizontal and featured just above the fold. However, the variant version looked exactly the same in terms of color, fonts, call-to-action button size, and more, but was displayed vertically. With just one element changed, the company’s A/B testing revealed they could boost conversions by 52 percent by displaying their booking form vertically instead of horizontally.
As this example shows, even small tweaks in your digital marketing material can make a big difference. Some tests will reveal greater conversion boost than others. But what really matters in the end is that you are using your A/B testing to learn about your audience’s expectations, then implementing relevant changes to continually boost conversions over time.
How to Set Up Split Testing in 5 Steps
Efficient split testing begins by observing what areas of a website are problem areas that hinder conversions. Funnel analysis tools make this observation possible. Once you’ve pinpointed areas that aren’t converting, session recordings, heatmaps, and surveys reveal what elements lead visitors astray and why. From there, you can edit those elements to create different versions of the original, then test them against each other to find the one that best converts.
Here are the five steps to conduct split testing:
1. Pinpoint Problem Conversion Points
To uncover conversion problem points, use funnel analysis tools to set up a typical path to conversion (or the action you want a user to take), then track what actions visitors take and do not take along the way. Actions visitors take to convert might include purchasing an item, booking an appointment, or downloading an app. You want your marketing materials to lead visitors to follow through with such key actions. Funnel analysis reports tell you at what point in the conversion path visitors most drop off.
2. Identify Elements Needing Changes
Once you understand where visitors drop off their conversion paths, you can use session recordings and heatmaps to observe visitors’ actions at that point in the conversion funnel. Pay special attention to what elements seem to be distracting, frustrating, or tripping up visitors as they move through that point in the conversion path. Develop a list of possible problem elements.
3. Ask Visitors Why a Problem Element Hinders Conversions
Now you can set up polls or surveys using a tool like Freshsales to ask visitors why the problem element keeps them from converting. You may find they report the shipping and handling form to be too long, for example, or that an offer wasn’t as great as they thought it would be once shipping and handling charges are factored in. Once you have received responses about what is wrong with a problem element, you can begin tweaking those elements to resolve the issue.
4. Launch New Versions of the Problem Element
Use a drag-and-drop editor in your chosen A/B testing tool to create one or more versions of your problem element. Remember to only focus on one element at a time so you aren’t confused about what elements boost conversions. Save each version so you can launch them simultaneously. As a small business, it’s best to stick to one or two new versions at a time so you can evenly divide enough traffic between each to extract definitive results over the course of one or two weeks.
5. Interpret Your A/B Testing Results
Most A/B testing tools offer dashboards with clear percentages of how each tested version helped or hurt conversions by comparing them with your original version. In addition, many of these tools will detect how long you need to test to ensure accurate results, then launch the winning version automatically. Just be sure you look at the results yourself to learn what elements don’t perform well. This helps you understand both what works and what doesn’t.
For those tools that don’t automatically launch a winning version, be sure to extract insights from winning, losing, and inconclusive versions alike. A winning version will have a clear percentage conversion increase over your original version. In the above image, for example, the variant shows a clear 40 percent increase in conversion rates above the original version.
Be sure to also pay close attention to those versions that didn’t win. Jot down the characteristics of the losing versions for each test you launch. Over time, you will likely begin to see trends in what works and what doesn’t. For example, green call-to-action buttons may tend to convert better than blue ones. You may use these findings to test this hypothesis elsewhere on your website on similar elements and boost conversions there as well.
Sometimes, your tests results will be inconclusive. In many cases, inconclusive tests don’t show clear results because the tests simply weren’t bold enough. For your next version, consider making bolder changes. Instead of changing a call-to-action button from light blue to dark blue, for example, make it pop by changing it to orange. Continue creating versions and running your test until you see significant conversion boosts from your changes.
For a complete guide on how to conduct successful split testing, read our comprehensive guide on A/B testing.
7 Areas to Use Split Testing in Your Marketing
After using analysis tools to pinpoint problem elements, what many businesses find is that problem elements are often small and that little tweaks can reap big rewards. These elements can include headlines, offers or promotions, call to actions (CTAs), and element colors. Ways to tweak them include adjusting the copy, formatting, offer-period lengths, CTA placements, and color.
Here are a few things you can test to optimize your online presence:
Page headlines are particularly important to test, as the average site visitor only spends a few seconds on your page before deciding to read on or leave. Therefore, it’s crucial you don’t simply follow “best practices,” but instead learn what makes your audience click or read on. Consider testing your headline tone, length, formatting, and fonts.
Offers or Promotions
Offers and promotions have many elements you can change to boost conversions, including product pricing, different free trial policies, money back guarantee policies, pricing tiers, and how to best incorporate shipping and handling costs. You can have two offers that save the user the same amount of money, and find that people are more likely to claim one over the other based on differences in these elements.
Calls to Action (CTAs)
Calls to action are used by your visitors to either purchase a product/service or perform actions that lead them closer to a purchase. CTAs could include asking your visitors to sign up for a newsletter, download a white paper, or schedule a consultation. Simply put, they’re pivotal to ultimately boosting revenue. Some ways to test CTAs include changing their placement, size, color, and text.
There is so much psychology in color, and you’ll find that you might get different results if you change the main colors of your design or elements of your design. For example, green makes people think of health and freshness, while red instills a sense of urgency. Try out different versions to come to the perfect color combination for your design, and reference the image below for guidance.
Be sure your color changes stay on-brand to ensure you are always recognizable. To do so, you may need to create a style guide. View our guide on how to do so here.
Email marketing software, like Constant Contact, offers built-in split-testing tools to help you set up your email testing, then track different versions of emails to determine which boosts conversions. When testing your emails, some common testing elements include subject lines, send times, personalization efforts, length, image inclusion and placement, and calls to action.
Online advertising platforms like Facebook, Instagram, Twitter, and Google offer their own platforms for deriving insights regarding ad performance. You can use platforms like Google Analytics or Instagram Insights to test different ad versions, then allocate larger budgets for those that convert best. For example, you can test ad offers, copy, call-to-action buttons, form structures, images, and more.
Landing Pages & Popups
Unbounce is designed to help you create different versions of landing pages, pop-ups, and sticky bars, then launch them and track performance to determine which best converts. You can integrate analysis capabilities like heatmaps, session recordings, and more by using tools that track user behavior as well. Elements that are often tested on landing pages and pop-ups include call-to-action buttons, layout, images, copy density, and trigger timing.
Split Testing Pro Tips
Professional marketers have learned a trick or two to get the most out of split testing, including how to know when you’ve extracted clear insights, how to set up tests for clear insights, and how to learn what needs testing.
When we asked the pros for their split testing tips and tricks, here’s what they said:
Wait for a Clear Winner
Scott Beckman, Digital Marketing Officer, Devetry
It’s important to let A/B tests reach statistical significance before you make any changes—and they need a lot of visits to get there. If your site doesn’t get enough traffic, A/B testing might not be feasible and you should focus on getting more traffic first. If you do get enough traffic to warrant an A/B test, you might be tempted to end the test early when a winner starts to emerge, but if you haven’t reached statistical significance yet, there’s a good chance it’s a false winner. Just be patient and let the test run its course.
Test One Element at a Time
Brendan Dubbels, Client Experience Director, ONTRAPORT
The most important thing to remember when running a split test is to keep it simple. Too often, we get carried away and test everything at once (aka multivariate testing). The problem is that the more iterations of a test, the larger sample size you’ll need to get any kind of meaningful results. Further, with multivariate testing, it can be difficult to really nail down what made the difference.
Instead, only test one thing in your emails or landing pages at a time. For example, test only the subject line of your emails, or just the call to action on your landing page. It may take a little bit longer, but your data will be much more reliable.
Ask Visitors What to Test
Alison Garrison, Senior Director of Services, Volusion
You may wonder which variables you should use to begin your A/B testing adventure. One way is by asking people who have never visited your website to give you feedback on what they like and don’t like. If there’s one thing everyone has, it’s an opinion, and you’ll be surprised at how often others can easily recognize problems and opportunities that you yourself never notice. You can also employ inexpensive consumer feedback tools that give you a “fresh pair of eyes” on your site. You’ll likely be surprised by what you can learn in a very short period of time.
Top Split-Testing Tools
Split-testing tools help small businesses observe online audience actions to narrow in on exactly what elements they need to change to boost conversions (a call-to-action button, for example) and then extract clear results about what changes will help them increase revenue.
Here are four A/B testing tools small businesses can use to boost website conversions:
- Unbounce – Unbounce offers landing page, sticky bar, and pop-up templates you can easily edit via a WYSIWYG editor to create different element versions (like calls to action, copy length, headline colors, and more) and test them. Packages start at $79 per month.
- Fiverr – If you don’t have the time or expertise to conduct your own split testing, you can hire an expert to do it for you on a project-by-project basis. Simply open a Fiverr account and use the search bar to search “split testing,” and then connect with a five-star expert from there. A/B testing prices start at $5 per project.
- Freshmarketer – Freshmarketer is a complete conversion rate optimization (CRO) suite that allows you to detect problem website elements via funnel analysis, heatmaps, polls, and session recordings, then create different versions of those elements via a drag-and-drop editor. From there, dashboards make it easy to pinpoint winning versions. A/B testing packages start at $99 per month.
- Google Optimize and Google Analytics – Google Analytics and Google Optimize are free tools that, when paired together, offer a toolkit for website testing. You can use Google Optimize to create different element versions via a WYSIWYG editor, then interpret the results using Google Analytics. With a less-intuitive interface, it requires a learning curve but is great for businesses on very tight budgets.
Frequently Asked Questions (FAQ)
How Many Split-Testing Variables Should I Test?
Only test one variable at a time. When split testing, you want a clear answer that you can easily understand and take action on where necessary. So while it might be tempting to test multiple variables at once (multivariate testing), as a small business, it’s tough to extract enough data to know which among those variables really made a difference. Stick to basic tests that only use one variable so you know what changes to make.
What Should I Split Test?
While many articles on this topic will tell you to test anything and everything, including headlines, call-to-action buttons, article length, and more, we recommend researching what needs testing so you don’t waste efforts and resources. Tools like heatmaps, visitor polls and surveys, funnel analysis tools, and session recordings give you clear insights into what web page features trip visitors up and why. From there, you can test different solutions to find the best one.
How Often Should I Run A/B Testing?
Continual testing is key. Start by researching what keeps your visitors from converting using session recordings, funnel analysis, heatmaps, and even surveys. Develop a list of problem elements, then test them one at a time. When you run out of things to test and problems to resolve, loop back to step one to observe website visitors, make a list of problems, then test to resolve them. Always stay ahead of visitors’ expectations and work to exceed them.
What Tools Do I Need to Start Split Testing on My Site?
To begin split testing your website, you’ll need a conversion-rate-optimization (CRO) toolkit. Freshmarketer offers A/B testing (split testing), heatmaps, session replays, funnel analysis, and polls so you can detect problem elements based on user interactions. Then, you can use a drag-and-drop editor to create different versions and test them against the original. Lastly, their dashboard clearly shows you a winning version.
What Is the Difference Between A/B Testing & Split Testing?
A/B testing and split testing are the same things. Using A/B testing tools with easy digital element editors, they both entail creating and comparing two versions of one digital element (like a call-to-action button) to measure which one best leads audiences to complete key conversion actions like clicks. Elements that can be tested include newsletter signup requests, call-to-action-button colors or placement, product pricing, and more.
Split testing uses testing tools to help businesses learn what digital elements (like calls to action, headlines, and more) hinder digital conversions, then launch a new version to resolve the issue. Through continual A/B testing, companies increase revenue and stay up-to-date on audience preferences. A/B testing is best for virtual companies like business coaches, software companies, and e-commerce stores.
For best results, a good split-testing tool helps you first learn what needs testing, then launches focused tests to deliver clear results. For $99 per month, Freshmarketer offers all the necessary tools to determine what elements you should test, and then develop informed tests to ultimately learn what boosts conversions and launch new versions of your site or digital marketing material. Click below to learn more and get started with this turnkey solution.