Importance of Experimentation in eCommerce Site Design

A 2015 Fortune Magazine article explored how Amazon went from an online bookseller to the world’s largest eCommerce vendor. Founder Jeff Bezos identified controlled experimentation as the greatest contributor to innovation. Just as Thomas Edison experimented countless times to find the right metal filament for his light bulb, Bezos and his team have never stopped experimenting with new ways to present products, convert customers, and finalize transactions. Bad ideas are discarded, while good ones become incorporated. 

Testing and adopting positive changes in your eCommerce website is your best opportunity to improve your store’s image and generate more successful conversions.

What is an A/B Test?

In many scientific or pharmaceutical laboratories, tests are devised using a known or control sample and comparing it to a new, prospective version. This process is known as A/B or split testing. After a series of trials, results may endorse the efficacy of the trial sample or confirm that is only equal to or does not perform as well as the control sample.

A/B Testing for eCommerce websites is similar. Using your existing site (A) and comparing the results against a parallel or B-site in a split test, if designed properly, can provide you with a logical basis for selecting one over the other.

The A/B test should be designed to split the eCommerce traffic into 50% for A and 50% for B. The site that delivers the higher percentage of conversions-per-visit during the test period should be the winner. In most cases, you should be able to determine why the difference exists.

Single and Multiple Variables

In many cases, an A/B Test may compare a single change or variable with the original eCommerce website. Alternatively, you might test multiple variables in an A/B/n testing format that requires much more detailed analyses to make a final determination.

Why All eCommerce Sites are Not Equal

Improvements in eCommerce websites that result in improved ad spend-per-conversion might be enhancements to design, content, functionality, or ease of use. A/B testing can determine which variations affect consumer purchasing behavior most significantly. Testing should be an ongoing process that refines meaningful improvements to better address consumer preferences, improve the customer shopping experience, and maintain interest.

Some reasons why one site generates more conversions than another with similar products is that it:

• Is easier to use
• Presents a more prominent search bar
• Places greater emphasis on products with wider general appeal
• Provides eye-catching graphics
• Uses improved product description clarity
• Includes easy-to-activate product videos with auto-play
• Maintains a persistent shopping cart; holds everything in place after the visitor leaves and follows with email reminders that encourage completing the transaction
• Provides clearer FAQs about returns and credits in case the product is not what the shopper expected
• Security of the site is assured
• Offers payment options should include more than credit card payments; i.e. PayPal, Zelle, etc.
• Asks only for information needed to complete the transaction. Often, near conversions are abandoned when a check-out process becomes too demanding.
• Prominently displays positive reviews
• Includes clearer Calls-to-Action

Tips for Running a Successful A/B Test for eCommerce

Satisfied customers at your eCommerce store drive improved click-through rates, higher levels of conversions and loyalty, improved query response, lower advertising cost-per-conversion, and much more.

To ensure that you get the most from your eCommerce A/B test, preparation and analysis are essential. Just changing a few background colors or rearranging the content without consideration to the intended message may deliver a result, but it is unlikely that you will understand why.

1. Develop a Hypothesis Based on Analysis

Before beginning an A/B test of a modified eCommerce site, begin with a clear hypothesis based on in-depth analysis.

Study your existing site to determine how customers tend to behave while visiting your website.

Begin by:

• Determining which current pages are most popular, most time spent, click-through rates (CTR), visitor path analysis, and more.
• Identifying where the visitors are getting hung up on your site, perhaps becoming confused with what to do next, and at which point they seem to be leaving without buying.
• Conducting live user tests, paying attention to where and when they react, perhaps encouraging testers to talk as they navigate through the site. Live testing can generate interesting results.
• Requesting simple survey responses to identify rough spots and ways to further improve.

2. Gradual Changes Allow You to Measure What Works

Overloading eCommerce Site B with too many changes at once can create confusion. Sure, Site B might deliver better results than Site A, but why?

Was it your new call-to-action button or the stunning graphics that drove higher conversion rates?

To understand you’re A/B Test results better, making individual changes allows store owners to understand comparative results better. Multiple changes in the site at once may show increased conversions although some features, had they been tested separately, might be negative. In other words, test each variable.

Gradual and continual experimentation by incorporating specific and measurable variables should be an ongoing process, allowing you to keep adapting your site to become more successful. 

3. Measure Improvements in the Overall Functionality of the Site

If you do nothing else, improve the usability of your store’s eCommerce site.

Dropped searches and abandoned shopping carts are most frequently the result of the poorly designed purchase and check-out processes. While the goal is not simply to create a site that makes shopping easy, the ultimate objective is to entice shoppers to want to return to your store’s site because of the positive experience.

4. Test Clarifying Your Value Proposition

Improving how you deliver value is always essential.

Without clearly defining your message and product value immediately, your limited-attention-span customers may simply move on with a click.

• More than a simple statement, your value offering should be expressed through clear and succinct copy, incorporating enchanting images and design, landing pages, and current propositions and benefits.
• To improve your eCommerce site’s value proposition, test different ways to express the message, then implement what works best.
• Experiment with a variety of placements to determine which elicits the best response.

5. Highlight Your Online Trustworthiness

Changes to your eCommerce website should earn the trust of your visitors.

If yours is a relatively new or unknown eCommerce store, you may consider emphasizing elements like credit card security, customer service accessibility, ease of returns and credits, and that delivery lead times and methods are reliable.

Earning positive reviews and testimonials from satisfied customers is often the most important endorsement of your trustworthiness. Allowing site visitors to clearly see how others have responded should an important element of your site redesign for A/B testing.

Because legitimate eCommerce has occasionally been invaded by bad characters, the fear of having personal information stolen from is real. This is particularly sensitive with newer eCommerce stores and websites. Have a well-known and trusted 3rd Party company like McAfee or others authenticate and certify your information management system. And, they will provide ongoing coverage to ensure your eCommerce site and data are safe from malware and virus.

Effective A/B Testing Can Transform Your eCommerce Store

Properly prepared, managed, measured, and adapted, A/B testing can systematically evolve your store’s eCommerce website into a more productive and effective shopping destination. Poorly considered ideas or wholesale changes to your eCommerce website without proper preparation and analysis can do more harm than good.

Be methodical and do not jump to conclusions without sufficient data.