Do you know about A/B testing, which is a perfect conversion rate optimization tool to select the most effective variant of many elements on your eCommerce store? With A/B testing, you can predict visitors’ interests to design your website, drive up your sales and keep the business profitable.


1. What is A/B testing?


As its name implies, A/B tests are trials, during which you try out two different variants of a given element. For example, you experiment a red CTA (call-to-action) button on one of the pages and a blue one on the other, so you can easily know what keeps your visitors stay on the page for a longer time.


A/B testing (also known as split testing or bucket testing) is a method of testing customers’ response to A against B, identifying which one is more effective. A/B tests are based on random approach and statistical hypotheses in the field of statistics. For instance, on your eCommerce website, the purchase funnel is a prime choice for A/B testing because marginal enhancements in drop-off rates imply a considerable growth in engagement and revenue, of course.  


2. What is a perfect A/B test?


Characters of a perfect A/B test:

  • Experiment on only one element at one time – or you risk causing conflicts between the changes resulted from several elements.
  • Run two pages parallel and simultaneously – your A/B tests should last at least for a week as a shorter time will not produce reliable results.
  • Choose a remarkable sample: the number of your testers should be at least 1000 people, which gives reliable statistics for you.
  • Test for big changes: you should not spend time worrying about some micro elements of your website because they are unlikely to influence conversions or to even be noticed by visitors.
  • Execute experiments randomly: you need to allocate the visitors coming to two pages unsystematically instead of the given segments, channels or any other dimensions. Your data will not be misreported.


3. A/B Testing Process


Now, it’s time to discover how to effectively start running an A/B test, consisting of five different steps to follow.


  • Step 1: Gather relevant data


The collected statistics can give you a complete view into where to optimize. You should focus on high-traffic areas of your website or app as they will let you collect data rapidly. Beside, the pages with low conversion rates or high drop-off rates are your potential subjects.


  • Step 2: Identify your goals


The aim of A/B testing is to identify the variables which you suspect that different versions would outperform the original versions. There are a variety of goals to target, from email signups, product basket designs or payment methods to anything you consider.


  • Step 3: Formulate hypotheses


After deciding your goals, you generate A/B testing ideas and hypotheses for why you wonder the variation is likely to be better than the current version. Then you make a list of ideas, give priority to them related to expected effects and complexity of implementation.


  • Step 4: Develop variants


You can take advantage of A/B testing software (e.g. Optimizely) to make different changes to a metric on your website or app experience. This is based on your initial aims, such as alternating the color of a button, changing the arrangements of elements, hiding navigation components, or something entirely custom. Several prevalent A/B testing tools allow you to use a visual editor to run these modifications simply. And remember to control your experiment so that it will work as you expected



Optimizely Web UI


  • Step 5: Run the test


Sequently, commence your A/B test and wait for visitors to join in. Particularly, visitors at your site or app should be randomly distributed to both the original and modified experiences. Next, their communication with each version is evaluated, summarized and compared to identify how each performs.


  • Step 6: Analyze Results


You will carefully analyze the results after your experiment finishes. The data is exhibited by your A/B testing software, which demonstrates the dissimilarity between two versions’ performance, and whether there is a statistically considerable dissimilarity.

If your new version outperforms than the current one, you succeed. Let’s apply your experience from this A/B testing on other components of your website to enhance your visitors’ engagement. However, if your test gains a negative result or no result, don't worry. You should consider the experiment as a learning experience and generate a new hypothesis to attempt again.


4. Why should you apply A/B testing?


Nowadays, in e-commerce industry, it is essential to provide customers with the better and better services on your website or app. If the website doesn’t appeal, it’s more likely that your visitors will leave your e-store for another.  

A/B testing gives you an analysis into possible outcomes in users experience from the collected data. You can easily generate hypotheses to test your original and new versions of any elements. Being able to predict changes is always an invaluable edge.

Moreover, Hypothesis testing doesn’t cost you a dime, in contrast to actual testing. A real-life experiment can cause immediate consequences should it fail. With hypothesizing, you are not vulnerable to losses.

Finally, A/B testing can be consistently and effectively used to better customer experience and boost conversion rates and sales over time.


5. What to test on an eCommerce website?


Not every bit of details requires modification as elements in a website do not have the same level of importance. Some of them are more impactful than the rest, keep in mind the initial goals you have chosen and translate them into strategic schemes. For an eCommerce website, you should evaluate:

  • Content: headlines, texts, product descriptions, testimonials, etc. There is a reason why the Internet has the mantra “Content is King”. Words are powerful, especially in business and marketing. Choosing the right message, with the correct words, for the right audiences helps to improve conversions immediately.
  • Images and video: Media are eye-catching, how many times have you scrolled through 5 inches of texts without glancing just to stop at an ad banner for 30 seconds? With that in mind, high-quality media with rich contents can help elevate any site.
  • Design: fonts, colors, position of elements on a page, etc. Design is a key criterion to attract your customers with the brand’s unified identity. So matching colors with a deep understanding the meaning of them can leave a huge impact on visitors’ impression.
  • Other elements: purchase funnel, call-to-action buttons, e-mail signups, social sharing buttons, homepage promotions, etc.




A/B testing is an invaluable resource to help you make game-changing decisions in the eCommerce scene. This entry may be not cover everything you want, but it is a promising start for you. If you believe in the power of A/B testing for continued revenue growth, that’s a fantastic place to be. So, let’s make an A/B testing instantly!