15 Oct 2019 by Tim Clayton
Super-charging E-commerce with Constant A/B Testing
Experimentation is now a must for any online business.
- A/B Testing is an ongoing process of improving small details
- Measurable insight allows you to see the value of every design decision
- Saleor Cloud gives you A/B Testing of e-commerce out of the box
Saleor is designed with a GraphQL API that gives your developers total front-end freedom to build the storefront that works for your business in whichever technology you prefer. You create a brilliant design but, even though it looks good, how do you know that it is leading customer journeys in the most effective way and driving up conversion rates and average order values?
Changing minor elements of your online store - such as the color of a single button or the size of product images - can have a major effect on a customer's decision-making process. Seemingly trivial details can be the difference-maker for the success of your business.
A/B testing allows you to offer ever-improving user experiences and get a better return on investment from your online store. Here's how it works and what it can do to elevate your e-commerce business.
After testing, Yuppiechef increased conversion rates 100% by removing a single distracting element from the page. (1)
How does A/B testing work?
Saleor's extensible API allows you to create different versions of your store, which you can then measure against one another to work out which design works best and then make informed business decisions.
Let's take a simple example. You have a grey 'buy now' button under the products on your online store; A/B testing would allow you to create multiple variants of your store with orange, green, and blue versions of the same button. You then have four different storefronts; which one a new visitor sees is equally and randomly assigned, meaning each version is seen by 25% of customers. By running this test for a week, you may get a result that looks something like this:
It's clear that the green or blue color is better for business than the other options. In this case, you might choose to implement a green button as your new standard (baseline) for your storefront. However, it makes sense to run another set of tests with only two versions of the button - green or blue - to validate your results and make sure that green is the best option.
A typical test would run for perhaps a week, or a number of user visits. You can then implement the results into your page design and move onto the next set of tests. It is important to take this incremental approach, as changing a number of elements at one time will make it impossible to clearly identify which factor is influencing results.
A/B testing is an iterative and proven way to test and improve almost any part of your page design, such as:
Product photos: The type of photography, size and number of images per product, whether zoomable or popup images are better.
Placement: Every button, element, advert, text box, and image can be tested in different places around to page to see how it affects conversion rates and user experience.
Copy: Test different slogans and headlines, font types and sizes, lengths of text, and levels of description. For example, Booking.com is almost continuously running A/B tests (2) and tries out alternate texts against one another to find the perfect description for different types of accommodation - because what works for a luxury hotel may not work for a cheap and cheerful B&B.
Interpret and take action
A/B testing will often give you results that work against your intuition. But do you argue with clear data-driven results or try to interpret them to understand your users better?
For example, in Saleor, you can build a modern single-page store checkout or you can use the API to create a more traditional 3-step checkout (with separate pages for billing, shipping, and payment).
Your intuition tells you: Your young user base will prefer a modern, single-page checkout. A/B testing results show you: A traditional three-step checkout leads to 50% less abandoned carts.
Your interpretation of the result: Your young shoppers use mobile devices. A traditional three-step checkout makes for cleaner information on smaller screens while the single-step checkout looks cluttered.
Your next step: Test different setups of a three-step checkout in subsequent experiments to find the perfect version for users.
You can continually run tests of different variants to enhance and evolve your storefront. As you remove pain points, you make it easier for customers to navigate your products, which reduces bounce rates as you have less frustrated shoppers abandoning your store when the UI does not work according to user expectations. You also increase the value of your design work as conversion increases.
A vital tool for modern e-commerce businesses
A/B testing also makes a lot of sense from a business point of view. The cost of running tests with Saleor Cloud is minimal as the platform comes with out-of-the-box A/B testing, unlike most competing platforms which rely upon third-party extensions. The costs incurred are fractional charges connected to the additional consumption of AWS services. However, the savings to business are huge. When one UX decision can either increase or cut conversion rates by half, the risk of making a mistake is nullified when you can scientifically measure the value of every design decision.
There are plenty of elements that make Saleor Cloud the right e-commerce platform for your business. We've spent a number of years building a headless commerce solution that offers ultra-fast, dynamic, beautiful anywhere experiences. Saleor's A/B testing is a validated application, built as a seamless integration to the platform, with a dedicated section of the dashboard for experimentation in your store, as well as an SDK for developers to easily implement A/B testing on your storefront.
By finding the easiest way to bring data-driven tools into your business, you are taking the shortest route to future success.
Read our follow-up article, coming soon, that shows how to avoid some of the pitfalls when running experiments and outlines the best way of getting results you can immediately interpret into informed decisions for your business.
And find out more about the work of our Data Science team in these popular articles about adding Recommendation Systems to your e-commerce.
Contact us at firstname.lastname@example.org to discuss how Saleor Cloud can bring data-driven insight to your e-commerce and super-charge growth.