Of all websites, ecommerce sites stand the most to gain by using A/B testing. It is the method of choice for anyone trying to get the most out of their site – as evidenced by some of the ecommerce sites that do heavy A/B testing (Amazon, Ebay, and Etsy to name a few). The potential for A/B testing is enormous in large part to the amount of variables you can play with: landing pages, copy, search ads, even search results, pricing, and product images.

Because the conversion goal is not an indirect metric, but a direct sale, it is much easier to measure the impact of the changes on an ecommerce site. If one change leads directly to an increase in revenue per visitor or average order value, you have directly affected the company bottom line.

What is A/B Testing?

Simply put, it is a way to test your audience’s preferences, with the end result hopefully being a conversion. While the testable variables are limitless, the two versions must be presented simultaneously for real world testing.

Say you are testing your “add to cart” button. Page A keeps the same button you have been using; this is called the “control”. B, or the “variable”, will have the new design to test. Divide your live traffic between the versions, and direct them to page A or B. After an appropriate amount of time, you can see which version worked better by measuring which button generated the most conversions.

Multivariate Testing

If you feel you are beyond the split testing level and want to explore deeper, multivariate testing might fit your site. While similar to the A/B set up, it is more complicated, and possibly a longer testing duration, with multiple variables being tested at once to give more specific results.

This sounds great in theory, but you might be wondering, how can I apply it and what results can I expect? Here are four tried and true ideas:

A. Call-to-Action Buttons

The buy button is one of the most important buttons on an ecommerce site, and therefore the most important test. If your button is not working, it may be because it is not noticeable, motivating, or clearly marked. Here are some elements to test:

Button color and size: Size matters. Larger and brighter buttons can increase conversions by 53 percent.

Button placement: Try bringing the button closer to your pricing, as TaylorGifts did, to increase clicks by 10 percent.

Button text: Does your button say “Buy”? Is there a limited time offer you can mention in this space?

B. Pricing, Discount, or Shipping Strategies

There are pricing strategies you can test that help nudge a casual window shopper into a conversion. For example:

End prices with a .97, .98, or a .99: Read more about the psychological testing behind changing prices, and test accordingly on your site.

Design of price: You can change the layout and design elements where you show the pricing. Is it to the left, to the right, close to the buy button? Do you have a link to the return policy or size charts?

You can also test the design of your shipping or discounts.

C. Product Display and Search Results

How do you choose the products on your homepage? When you show search results, should you list them in alphabetical order, or bestsellers first, or lead with new items? What product categories do you show in the menu? These are all questions A/B testing can answer.

Even slightly changing the mix on your product pages can lead to an increase in revenue. Similarly, you can test the default order of search results, or number of results to show visitors. Within these, test whether showing user reviews increases revenue or decreases revenue, by confusing with too many choices.

D. Checkout Page

You get your visitors close to finalizing a purchase, and they abandon their shopping cart at the final page. Things such as requiring a login before they can checkout or not including surprise costs until the very end such as tax or shipping can turn shoppers away. While this is annoying, by running checkout page A/B tests you can minimize shopping cart abandonment. Things to test include:

Including or excluding a seal/badge: Let your tests decide which is best for your site.

Remove unnecessary input fields and provide an overview of the order: See if linking shipping and billing addresses, not requiring a login, or continuing a view of the order helps with conversions.

Along with the suggestions above, there are A/B testing tools available online. Some favorites include Google Analytics, Optimizely, and Unbounce. Your ecommerce site can only improve with testing, so do not be afraid to delve into all sorts of options – with or without additional system support!