Online customer acquisition is challenging and competitive. The average conversion rate of an eCommerce website in 2020 is ~2.86%. It’s evident that if you can increase your conversion rate by 2%, you can double your revenue. Marketers are yet to figure out a rule of thumb when it comes to CRO (Conversion rate optimisation). But domains like cognitive marketing, AI-powered eCommerce, user experience management, website analytics and website testing have been used to increase conversions. This article will focus on how testing can improve conversions.

What is testing?

Assume that you have a fully functioning website, what does it take for a visitor to search a product, add the product to the cart, trust you with their credit card details, proceed to a checkout and to revisit your site again? All of these decisions are subjective and the slightest change to your website architecture will increase or decrease the likelihood of those decisions. Testing helps you to understand the current state of your website and the effect of a slight change. In 2004, Google did such a test and it became famous. Google wanted to know if the colour blue on their search ads had a significant impact on the click-through rate. They ran a multivariate test to check this hypothesis. They picked 50 different shades of blue and showed them each to 1% of their visitors as their search ad results. They then picked the most engaging blue colour . This slight change of colour resulted in a $200M increase in their ad revenue.

What’s the process when it comes to website testing?

Testing is a 4 step process:

1. Understand the current state and identify bottlenecks.
   It’s important to understand the behaviour of your website visitors and the pain points of their user journey. Tools and techniques like Google analytics, form analytics, heat maps and session recordings can give you valuable insights about when and where the users struggle and exit. This data will help you understand website elements that create friction for your users. You can randomly decide and test any element of your website but a data-backed understanding of what to test is the best practice.
   
   
2. Define hypotheses
   With the site data in hand, you can decide solution(s) to overcome the friction you identified. It could be a change to simplify the number of steps at checkout or changing UI elements. It’s important to understand that hypotheses are just expert opinions and they could be right or wrong. The sole purpose of running a test is to make data-backed decisions by avoiding human intuition. Such a hypothesis could be; “Implying to users that the checkout process is secured can Increase checkouts”
3. Run a test
   Once you have a hypothesis you need to decide what kind of a test you’re going to run. It could be a split test (A/B test) where you test a single variant at a time or a multivariate test where you test multiple variants at the same time. Either way, you need to define a sample size depending on how significant the changes are. In a split test, you need to come up with 2 versions of the element you are testing as the controlled and the variant. It's important that you run a rest for a considerable amount of time before jumping to any conclusions.
   
   For example Magento, PayPal and HiConversion wanted to see how eCommerce shoppers would react to a slight change to the checkout button. All they did was replace the cart icon of the checkout button on some Magento powered eCommerce stores with a lock icon. The results were amazing and they experienced a 21% increase in cart conversion rate.
   
4. Evaluate results and optimise the experience.
   Test data will highlight what type of change you need to make. Change will lead to an improved customer experience, but testing has to be a part of your everyday culture to achieve a significant improvement in your conversion rate. In other words, a static website won't survive the current eCommerce ecosystem.

What elements need more attention?

Out of all the elements that can be tested and tweaked, it's important to understand which elements matter most. Understanding the significance will help you to prioritise your testing calendar. In this section, we can recommend a few variants that could work for each web element, but it’s important to test them in your eCommerce environment.

• Home or landing pages

This is the first impression that users get about your website. You need to make sure that these pages perform well. Testing and optimising the page load speed to meet industry standards (1-3 seconds) is critical. The page should give a clear direction for visitors that pushes them closer to a conversation. Simplified and easy to find navigations, personalised landing pages according to different demographics and geographical locations are key, as is placing your call to actions and promotions on hot spots (according to heat map results) can improve conversations.

Eg - using personalised landing pages based on the state that the user is in vs. a single landing page Australia wide.

• Search and navigation

It must be easier for customers to find what they are looking for. Your search icon needs to have a prominent place on each page. Search results should be presented in a way that customers can easily scroll through and find what they are looking for.

• Product page

Once a visitor lands on a product page, you need to understand what information would convince them to add that product to the cart. There is unlimited info you can offer, but you have limited space and a limited time (users attention span) . Below are examples of the types of strategies that you could run. To determine what’s most effective for your business, you can split test these strategies:

• Cart and Checkout

Cart and checkout should be a high priority because the average eCommerce shopping cart abandonment rate is high at ~71%. You should even consider the possibility of reworking your checkout process according to market demand. Currently customers:

1. Would like to have a simplified payment process.
2. Trust sites when they see security logos.
3. Would like to see multiple payment options
4. Like One-click checkout buttons like PayPal express
5. Multiple checkout buttons would work better than a dropdown menu for payment options
6. Visitors would prefer to see ZipPay and AfterPay over Visa and Master.

All of the above are hypotheses based on current market demand and are worth testing.

• CTA

The slightest change to the shape, colours or font of your call to action button(s) can impact your engagement. There is ample 3rd party research data that explains consumer behaviour and their reactions to different colors, shapes and words. Finding such inspirations and testing them on your website is important. For example, when Facebook found that people are more attracted to circles than squares and rectangles, they changed their entire design to replace as many squares as possible with circles. We all agree that FB is addictive, but it's not just the content and concept. Their designs are tested to drive conversions and form habits.

• Forms and User Accounts

It’s important to find the happy medium between asking too little and too many questions. Most visitors stop submitting forms due to their length and complexity. It’s important to test simplified versions of your forms to improve the submission ratio. You can run tests to figure out if auto-filling information, checklist and dropdowns assists users to complete forms.

Data-driven designs are important to enhance conversions. A test and tweak culture can positively impact any marketing campaign ranging from a word choice in a fart abandonment email to a colour change in your brand logo. Prioritise testing on your design and online marketing strategy must be the way forward. With our experience in eCommerce development, Maytech can help you to plan and execute a strategic roadmap for testing and conversion rate optimisation.