Download our FREE Testing Toolkit for A/B testing ideas, planning worksheets, presentation templates, and more! Download Now

Optimizely Blog

Tips & Tricks for Building Your Experimentation Program

X

Download our FREE Testing Toolkit for A/B testing ideas, planning worksheets, presentation templates, and more!

Get It Now




For companies with limited time and resources for A/B testing, it can be tempting to rely solely on “CRO best practices”. However, those practices may not apply exactly to every organization. For example, these variations are often expected to be guaranteed, effective changes, but will not always result in a win for every business:

  • Orange CTA button > Blue CTA button
  • Not having a carousel > having a carousel
  • 10 > $10.00

Therefore, it’s not about copying another company’s CTA button, real estate layout, or variation exactly; it’s about the experiment itself. A business should look at the areas to test and apply it to their own unique challenges, generating quality variation ideas with a higher probability of winning.

This philosophy pushes a company to test every part of the customer journey, from the home page to the checkout, to optimize the conversion funnel and collect behavioral data. That data can be used repeatedly to test for compound gains, due largely to the quality of ideas at the start of a testing.

Here are 7 successful experiments that an e-commerce business should use to optimize sales and growth:


1. A/B Test Your Product Images and Models

Adore Me, founded in 2010, has earned $5.6M in revenue by 2013, growing 5,506% in just 3 years. (Source: @rzgreenfield)

CEO Hermand-Waiche confirms that – for each brassiere – Adore Me tests multiple versions of images to run on its website.

According to internal research, a slight change in the model’s position, such as shifting from a hand on the hip to a hand on the head, can double sales!

Adore-Me-Product-Images

Adore Me tested multiple versions of product images, including variations using different positions of models.


2.  A/B Test the Recommendation of Related Products to Customers

59% of retailers simply don't know their conversion rates when shoppers preview up-sells or cross-sells on a product page.

59% of retailers simply don’t know their conversion rates when shoppers preview up-sells or cross-sells on a product page.

Per e-tailing group’s 16th Annual Merchant Survey, 59% of retailers don’t know the conversion rates achieved when shoppers preview up-sells or cross-sells on product pages.

So, when over half retail companies aren’t testing cross-sells or up-sells with recommended products, how does one know what’s working?

By testing and tracking results.

Test a variety of pages: a version with no up-sells and cross-sells, a version showing related products on product detail pages, and a version only showing related products after an item has been added to the cart. The more variations a company tests, the more actionable data they will collect.

A leading online men’s retailer, Bonobos, keeps their product page very simple and focuses on one goal – getting the customer to add the product to the cart.

Their page does not have distractions, such as related products, up-sells, cross-sells, or social sharing buttons.

Bonobos keeps their product page design very simple so that customers are not distracted.

Bonobos keeps their product page design very simple so that customers are not distracted.

Amazon, alternatively, makes a big push for related products; recommendations appear after a user has added a product to the cart. In this instance I added a hammock to my cart and Amazon is recommending a hammock stand or a hammock pillow to go along with it.

Amazon recommends products after a user initially adds an item to their cart.

Amazon recommends products after a user initially adds an item to their cart.

When a customer continues and adds one of their “frequently bought with” products, Amazon will generate a new product list for the customer:

Amazon refreshes a new related product list after a user adds a recommended product.

Amazon refreshes a new related product list after a user adds a recommended product.


 3. Test a Right Column, Left Column Swap

BabyAge, a top 500 e-tailer, saw a 22% increase in conversions when swapping the right column with the left:

The original version (the control, image on the left) was tested against the test idea of a real estate column swap (the variation, image on the right) to find a win, resulting in a 22% conversion rate improvement.

The original version (the control, image on the left) was tested against the test idea of a real estate column swap (the variation, image on the right) to find a win, resulting in a 22% conversion rate improvement.

(Source: BabyAge Case Study)


4. Test Irregular Shapes for CTA buttons

Marketing expert, Bryan Eisenberg (@TheGrok), reveals his test idea secret to Larry Chase (@LarryChaseWDFM).

Button Shapes: What pops?

LC: Do rectangles or ovals make better CTA button shapes?

BE: One of the secrets we learned about having great call-to-action buttons, and this is something you might not even notice, is using irregular shapes, neither rectangular nor oval.

Amazon’s button uses a combination of an oval and a rectangle. The left side of the button is rounded like an oval and the right side has corners, like a rectangle.

Linda Bustos even keeps collections of irregular icons and unusual “add to cart” buttons that may help widen your CTA test ideas.

Examples of irregular buttons or icons that can be tested or inspire new ideas to test.

Examples of irregular buttons or icons that can be tested or inspire new ideas to test.

Bryan Eisenberg also mentions that one client tested a button saying “Buy Now” versus “Add to Cart.”

“Buy Now” increased conversions, but “Add to Cart” increased average order values because shoppers don’t feel pressured to make a commitment.

This small variation reinforces how important it is to have goals in mind when testing, and ultimately, that means a company should optimize for lifetime revenue and happy customers. The two go hand in hand.


5. Test Social Login

Mailchimp tested social login and found that only 3.4% of users clicked them, spurring them to author a written piece about how social login isn’t worth the effort.

eConsultantcy also goes over the pros and cons of a Facebook login on e-commerce sites. However, it is best for a company to test and collect conclusive evidence about social login when paired with their own e-commerce efforts.

Windy City Novelties offers customers a chance to login with their Facebook account.

Windy City Novelties offers customers a chance to login with their Facebook account.

Pacsun offers four social login options to its customers.

Pacsun offers four social login options to its customers.

(Source: @Econsultancy @DavidMoth)


6. Test a More Prominent Search Bar

Talia Wolf From Conversioner increased e-commerce revenue by 550% with an enhanced search bar after analyzing heat maps and customer behavior:

This is the heat map of Wall Monkeys’ original version (control). Analyzing customer behavior and recognizing the importance of the search bar helped them create a variation idea to enhance it.

This is the heat map of Wall Monkeys’ original version (control). Analyzing customer behavior and recognizing the importance of the search bar helped them create a variation idea to enhance it.

Conversioner tested a variation that featured the search bar more prominently, improving user experience and increasing conversion rate, showing the importance of behavioral data.

Conversioner tested a variation that featured the search bar more prominently, improving user experience and increasing conversion rate, showing the importance of behavioral data.

Armed with the approach to “always be testing”, they increased conversions by 550% in two testing rounds.

(Source: Conversioner)


7. Test Your Pricing

A study published in the Journal of Consumer Psychology discovered that including commas and cents in pricing display can change how price is perceived by consumers.

Consumers subconsciously sense that there is a positive relationship between a syllabic length and perceived price.

For example, shoppers think “$1,600.00” is more expensive than “$1,600”, which is more expensive than “1600”.

If an organization is an international business that transacts with multiple currencies, it may need to keep the $ sign in there, but this too can be tested.

Take a look at Thieves Boutique:

Thieves Boutique has eliminated money signs and decimal points, minimizing the price length and improving pricing perception.

Thieves Boutique has eliminated money signs and decimal points, minimizing the price length and improving pricing perception.

After a company has tested the absence of commas and decimals, it can also try a classic idea – ending pricing with the number 9.

Research from the Journal of Quantitative Marketing and Economics proves that prices ending in the number 9 were so effective that they were able to outsell lower prices for the exact same product line.

The study took items at comparable price points, such as a pair of jeans, testing options like $44 vs. $49, only to find that the higher $49 price shockingly outsold the cheaper price point by 24%.

(Source: Gregory Ciotti)

Studies also show that shoppers perceive numbers with dollar signs and decimal points to be larger than their counterparts.

Therefore, consumers do not consciously associate a positive relationship between verbal length and numerical value.


In Conclusion:

These are just 7 optimization ideas that companies can start testing today that will push any testing program to start making better decisions using reliable data.

Remember, correlation is not causation, and there is no magic button or silver bullet A/B test to solve every challenge instantly.

Look for and identify trends that will ensure better experiences for customers. This ultimately results in increased conversions and revenue growth:

GA

By testing every touch point a consumer has with a brand, e-commerce companies can facilitate customers through the conversion funnel. Use these ideas as a starting point to continuously test and evolve pages for improved e-commerce experiences.

Hopefully these concepts sparked some ideas for you and your team to A/B test. We’ve looked at over hundreds of our eCommerce tests to find out which pages have the highest win rate, by consistently A/B testing your eCommerce site you are guaranteed to increase your conversion rates.

Ecommerce A/B Testing

Optimizely X
comments powered by Disqus