Split testing, also called A/B testing, is a common practice to test one version of an app, advertisement, landing page or website versus another version. Let’s talk about why A/B testing can be beneficial for your business.
But why? Is split testing important for your site?
What is A/B or Split Testing?
Every visitor to your website is an opportunity to:
- Acquire a customer
- Build a mailing list
- Sell a product
Often, A/B testing is an effort to make a visitor go from a lurker to a person that enters your conversion funnel. The idea is simple:
- A is one version of a website or web page.
- B is another version of a website or web page.
You’ll be showing two variants of your site to your audience. The goal is to gather data on how the user interacts with each version of your site and what certain changes on your site mean for your conversion rates.
Why is Split Testing Important?
Split testing is important because it can boost conversions. You may not be a business owner or you may not be selling a product directly – may be you’re building a mailing list. In this case, you may want to convert more users from readers to subscribers.
After all, a mailing list can be useful in the future, whether you want to make and promote your own product or sell someone else’s product.
Split testing allows you to boost conversions for everything, including:
- Calls to your office
- Direct e-commerce sales
- Mailing list subscriptions
- Downloading files
If you want a reader to take a specific action, A/B testing can help. Your action may be keeping readers on the page for longer, or sharing your article on social media.
In any case where an action is taken, you can rely on testing to help you better understand users and nudge them to take the actions you desire.
What Can You Learn from Split Testing?
Split testing will teach you a lot about your site’s visitors and their habits on your website or app. What you learn will depend on the goal of your split test, but some of the many things you’ll learn are:
- Which website variant leads to the best conversion rates or results
- What colors work best
- What font size or color encourages user action
- Which placement works best for conversions
- What button size or color works best
- What copy or text works best on a site
You can learn a lot from making minor changes on your site. If your CTA is too far down on the page and visitors do not have a high dwell time, they may never click on your CTA. That means they will not join your mailing list, complete a contact form, or link to add a product to their cart.
Split testing allows you to:
- Better understand your user’s behavior
- Make positive changes to your site
- Increase conversion rates
A common mistake that people make when conducting their split testing is that:
- Version A will be a completely new version of a site
- Version B will be radically different than version A
Split testing needs a control, or a version of the design that doesn’t change. Let’s assume that the control is Version A and that you want to sell more widgets on your landing page.
You may use Version B to test out different button placements to see if users convert at a higher rate when the button is placed on the top right of the screen than on the bottom left.
You may also use split testing to try variations for:
- Any element
- Landing pages
But you want to be able to know which element led to the results you achieved. If you include more than one change between Version A or B, you’ll have to question whether it was the different text or button that led to better results.
Due to uncertainty when making multiple changes, it’s often best to make changes one at a time to determine what works best for your site.
You may update the intro text and find that dwell times rise, and conversions increase by 1%. Afterward, you may change the button placement and find that sales decreased, so you switch back to the previous variant and make another change to test the results.
Small, measurable changes can have a dramatic effect on your website and your goals.
A/B testing allows you to better understand how these changes correlate to user behavior. Examples of split testing and the results that others have achieved through their tests are the best way to understand how to use testing on your website.
A Real-life Example of Split Testing and Results
Real-life examples of split testing and the results that website owners have been able to achieve can be found online. A lot of companies have outlined their tests and the results that they’ve been able to achieve.
ComScore Achieved 69% Increase in Leads
Optimizely outlines the steps that comScore took with their A/B test to achieve a 69% higher conversion rate on product pages. The testing starts with a hypothesis, which outlines what is expected from the test.
The company correctly hypothesized that if they were able to do two things, they would increase leads:
- Improve social proof
- Improve brand validation
What the team did was create three variations of testimonials and tested them against the original control page. The variations tested:
- Different logo placement
- Different design orientation
The changes were minor, as discussed earlier, so that the team could test each variant to find what changes led to a positive increase in leads.
- Variant 1 had the customer logo in the testimonial.
- Variant 2 contained no logo or variant, but instead placed the testimonial below the product description horizontally.
- Variant 3 was similar to Variant 2 but included a logo.
Around 2,500 visitors were included in the test, with the team dividing traffic equally among all variants. The primary conversion goal was to cause visitors to click on the “Demo requested” page.
The conversion rates were:
- Variant 1 – 5.5%
- Variant 2 – 3.8%
- Variant 3 – 4.3%
In comparison, the original version of the page had a 3.3% conversion rate. Variant 1 led to a 69% increase in leads. The team chose to implement the first test variant.
HubSpot’s Site Search Increases Conversion Rate by 3.4%
HubSpot conducts a lot of split testing to determine how small changes on their site can lead to big improvements. The company found that users that engage with search increase blog lead conversion by 163.8%.
The team involved with the test did not provide as much information as our previous example, but the results are really what matter.
Little in the test information points to the team’s hypothesis, but we can determine that the team wanted to increase search usage, as it was proven to increase lead conversion rates. The HubSpot team also chose to have three variants to test on their design:
- Variant 1 placed the search to the left of the log-in button and changed the placeholder text to “search by topic.”
- Variant 2 placed the search in the same prominent position as Variant 1, but the search conducted searches on the blog and not the entire site.
- Variant 3 was the same as Variant 2, but the placeholder text changed to “search the blog.”
As you can see, the changes in these variants are very minor. Small placeholder changes were made to determine if wording and what portion of the site was searched made a difference in user behavior.
Variant C was the test’s winner, with the variant leading to:
- 3.4% increase in conversion rate
- 6.46% in users engaging with the search bar
We do not have any information on the other variant increases, but something as simple as changing the search’s placeholder text to “search the blog” and searching the blog rather than the entire site led to great results for the site.
There is no information on how much traffic was driven to each variant, but HubSpot has more than enough traffic to test multiple variants of a page without an issue of small sample sizes.
When you have lower visitor counts to a site, you need to keep your test variants low to ensure the sample size is large enough to come up with accurate results. If your site only receives 100 visitors per day, you may have to run a three-variant test over the course of a month to have a decent sample size.
Larger sample sizes provide a clearer picture and proof of how each variant impacted your site.
Once you determine which test leads to optimal results, you can implement these changes to a new page and enjoy higher conversions as a result.
Get Started with Split Testing!
If you’re ready to take the dive into split-testing to improve the ROI on your Ad spend, or to simply offer the best user experience possible, we can help. Let’s talk about what Bocain Designs can do for you. Contact us to get started with split-testing!