06/02/2022

A/B Testing for Beginners: Complete Guide

Insights

12 min remaining

The Beginner’s Guide to A/B Testing

It may seem simple to create a website by organizing pages and adding text. You then added links to your contact pages.

That would be easy!

A/B testing is a great tool for this. You don’t need to know the definition of A/B Testing. This comprehensive guide will help you test and improve your website.

Introduction to A/B testing

A/B testing is something you are probably interested in if you’re looking at this guide. The full benefits of A/B testing might surprise you.

Let’s briefly explain what A/B Testing is and how it works.

Watch the video to learn more about A/B testing. Then, continue reading.

How A/B testing works

Perhaps you remember lab sessions from high school. You may have been involved in lab sessions at high school.

A/B testing may also include control or test groups. Your visitors are the control and test groups.

A control group with the label “A” can be used to test an element of your website such as a button, form, or other component. The new version of your website would be displayed to the group marked “A”. “

The test results are compared at end to determine which group reacts best to the element. If the B group had a stronger reaction, the element should be modified in order to match that of the B group.

What’s better than a conversion? Conversions are when people buy, complete forms, or do other actions you encourage.

What do I do if the A/B Test fails?

Websites can be imperfect. It is not possible to predict what your audience will react to.

You can assume that customers and leads respond better to a button with a blue color. A green button is more noticeable. However, it can be argued that the product copy should be included on the page in order to increase conversions. People will click on it even if it’s below the photos.

Without A/B testing, you won’t be able to tell if your conversions are dropping without monitoring and running different tests. This will help you increase your leads, and ultimately your revenue.

What you need to get started

These are the essentials you need to get started in A/B testing:

  1. Full access is available (via FTP, CMS etc.). You can upload to a website.
  2. We’ll cover a few options to test A/B software.
  3. This is a list with elements you’d like to test
  4. Patience is essential because the best tests usually take around 30 days.

Now that you have all the necessary information, it is time to get started with your A/B testing.

Now let’s look at the elements you need to test on your website. 

What should I do?

As we have mentioned, A/B Testing focuses on elements involved in conversions. This is the key component of website improvement that will increase conversions.

You have the option to change the size and color of the “add-to cart” button. We can also modify the appearance of the contact form.

Buttons

You can use these buttons as a test element on any page of your site. To determine if your visitors react differently to your buttons you can alter their size, shape, and color.

While you may not think that a change to the color of your “add-to-cart” button could have an effect on conversions, this HubSpot study showed that converting from green to red increased conversions 21% 

Action is required

It is important to test not only buttons as calls for actions but also text and link call to actions (CTAs). Stronger or more specific language can help you encourage more people to take the action you desire.

To add a hyperlink from one page to another with similar information, you can change the color, font size and length of the link.

It might be worth testing a call-to-action in a popup rather than just placing it on the page. 

Formulas

Contact forms on your website can be powerful tools for collecting leads. You might test their length, appearance and number of fields to see if they are possible to collect additional information.

Don’t forget to ask the right questions. Clear labels and a reduction in fields can help you clearly identify what information is mandatory and optional.

Any form can be tested on your website. You can also test different labels and pre-filling fields for your checkout forms.

Localities

An example of product copy placed side-by-side on a page is shown . This could be an example of location-based testing.

You can test the position of elements like signup forms for your mailing lists or other elements that might be connected to conversions. Your email CTA may have little effect at the top but is effective at the bottom, where visitors are coming from after reading your content.

Visual elements

It is possible to change it. You need to test colors, layouts and spacing.

If you have a dark background, change it to a lighter one. Visitors will be able to tell if the B version is easier for them and if it’s worth taking a little longer to look at.

These other ideas might be of interest to you

Other elements can be tested. 

A/B testing pricing is a possibility if you have an online store. To test whether people are interested in certain plans, you might change the pricing.

Software to A/B Test

Now you know why A/B testing matters and have some ideas about elements that you can test. You might now be wondering where to begin.

You have many options when it comes to A/B testing software. Prices will depend on the features you choose.

Let’s take a look at the most popular A/B testing options, starting with Google’s Content Experiments.

Google Analytics Content Experiments

Google Analytics content explorations can be a great way for you to get started with Google Analytics. Your visitors will see different versions of your pages if you send them random samples to Content Experiments.

Content experiments differ from other A/B testing because you can test multiple A/B groups simultaneously. 

To create an Content Experiment 

  1. Click “Behavior” and then “Experiments” in Google Analytics. “
  2. Click the “Create an experiment” button.
  3. You can set a goal, specify how long the test should last and what percentage site traffic you wish to be tested.
  4. Add URLs for pages you want to compare.
  5. You should place the provided code in the correct spot on your website.
  6. After you’re certain that the code is working, try it out.

Content Experiments has one caveat. You must make multiple copies of the pages you want to test. They cannot be accessed by others and should appear normal.

You can view your experiment results in Google Analytics after it is finished.

This will help you determine who the winner is. This is why you want to get close to 100 %.

Optimizely

Optimizely lets website owners monitor and run tests with no technical knowledge. Optimizely allows you to “overlay” the control page.

Optimizely can be used for free. Optimizely offers a free or custom-quoted Enterprise Plan. This snippet can be added to the head section of your website. This snippet can be used only once.

Once your snippet is placed, Optimizely can be used to create an experiment on any page of your website. After the test is completed, Optimizely can be accessed to view the results.

Optimizely will not do site changes for your site if you fail a test. This tool will give you confidence that you are able to make significant changes to your website.

VWO

VWO is similar in function to Optimizely. You only need to add one line to your website to enable A/B testing.

VWO is more costly than other options, but the basic plan costs only $9 per month. This plan is ideal for those who are just starting out and need to test with less than 2000 users. VWO is also available for larger sites for $49 per month or $499 per year.

Set up a VWO Test by placing your smartcode at your website’s head. Make any changes you want to the B version of the page using the visual editor.

After you’ve determined the winner, you can start implementing it. 

Unbounce

The final tool we have is not an A/B Testing Tool. It does have some A/B options, so it’s worth considering!

Unbounce lets you create landing pages for products, services or promotions.

Unbounce creates landing pages that look amazing and are easy-to-create. View the dashboard to see the results of any experiment.

Unbounce is $49 per month and can be used by 5,000 users. For any campaign or promotion, you can create unlimited landing pages. It might be worthwhile for your business.

How do you run a test

Your choice of software will impact the setup of A/B testing. 

Select what you would like to do

Pay attention to the most important elements when testing. It is possible to test multiple elements on one page. However, it will be difficult to identify which element caused the fluctuations in conversion rate.

Let’s say your “start now” button doesn’t convert well. It will be difficult to determine why the new version isn’t working.

Choose the variant you prefer 

Would you like to see something changed in the test version of? Your purple button will be replaced by the green one.

It is possible to add another variant to a small aspect, like the button color. To see if color has an effect on conversions, compare your current purple with both blue and green.

Remember that statistical significance will be higher for visitors who have created more variations than you have. (more details below).

Start your test

Create your test using the tool that you prefer. Double-check it in multiple browsers.

Wait

Waiting for the data to arrive is the hardest part of an A/B test.    

If you have many visitors to your website, it might not be necessary to wait as long. This is especially true when the test version exceeds the control.

Seek statistical significance

Be sure to check for statistical significance results as they are important.

It is unlikely that this difference will be significant if the conversion rate of page A is 2.17% and page B is 2.19%. This could be one of many leads or a small portion of your revenue.   

However, this difference could be substantial if you have a lot of leads or traffic. 

Some tests give better results when left running for longer periods of time The Duration Calculator from VWO can help you decide how long your test should run

Reviewing the results

You can now evaluate the results of your A/B test.

Here are some things you should be looking out for when reviewing your results.

This preference is clearly

We have already discussed statistical significance and it is worth repeating. Look at your test results to find data that clearly shows preference by your visitors.

Don’t look at only percentages. A 33% increase in conversion rates sounds impressive. However, if there are only two additional leads to the sample, it is unlikely that this is convincing evidence.

Notable sample sizes

You need to have high numbers of people visiting your A/B testing pages.

It is possible to find that 1,000 visitors is an appropriate sampling size depending on the size of your site. While it is not essential, small sample sizes could lead to misleading results.

You should wait at least one month before making any decisions about A/B testing. This can be used to determine how many visitors you’ll need to complete the test.

Be aware of trends and wider conclusions that you can draw from your tests, especially if they are not your first. A single test of a button and form can tell you a lot about visitors to your site and how they interact with it.

It may be that you find that your landing pages have a greater impact on conversions than others. Maybe you don’t include enough calls to action early. 

Actionable data

You have to ensure that the results of testing elements are useful. Let’s just say “This variation won”, and let’s turn all our buttons to green! It can be hard to decide what actions to take if more than one element is being tested at once.

About the author

Kobe Digital is a unified team of performance marketing, design, and video production experts. Our mastery of these disciplines is what makes us effective. Our ability to integrate them seamlessly is what makes us unique.