September 18, 2015 · Alyssa Fate

Improving Conversion Rates with A/B Testing

Your business is booming (or at least we hope it is) and you’ve built an impressive new website to help elevate your success. You briefly sit back, kick your feet up and suddenly realize that consumers are constantly changing, so shouldn’t your website be too?  

The recipe to understanding consumers’ actions is a complex one, filled with a long list of fresh ingredients. From new technology to seasonality, B2C businesses are tasked with following the constant ebb and flow of a consumer-driven market. This means that B2C businesses must continue to adapt their websites based on their consumers’ ever-changing wants and needs.

A stagnant website will result in a stagnant conversion rate.

This begs the question: How do B2C businesses determine what the consumer wants and needs?

Aside from the all-important value that you should already be providing to your consumer, there are a number of components (among many others) that influence conversion rate. Performing A/B tests can ensure your website is always adapting to fit consumer wants and needs as well as working to achieve the best possible conversion rate.

A/B testing (sometimes called split testing) is comparing two versions of a web page to see which one performs better.

More than half of all organizations are likely testing. 

As marketers with increased knowledge of the inner workings of a website, we cannot assume to know how a consumer will navigate.  A/B testing can improve the user-friendliness of a site and can help streamline the conversion process. Whether your goal is to grow subscriptions, capture email addresses, gain survey responses, multiply video views or increase revenue, it has become increasingly imperative to utilize A/B testing to determine what drives the consumer’s decision-making process.

We’ve skimmed the surface of a sea of possibilities to provide the less experienced some good starting points among the immense depths of A/B testing.

1. Device

  • Consider your personal use of devices. Do you make use of your large desktop monitor the same way that you use your mobile device? Compare the screen size of your 6s iPhone to Apple’s new iPad Pro. Do the elements of your website look the same?
  • Consumers are now spread across an extensive variety of device types, sizes and even brands. Take this into account when setting up your test. Consumers on a tablet may have different intentions than those on a desktop device. Does your test make sense serving all device types? Does it display neatly across all browser sizes? The same A/B test may provide different winning variants on separate device types.

  1. Time
  • If you browse the internet you’ve most likely encountered a first-time visitor pop-up on a website. This tried and true method is deemed irritating by many, but is a marketer’s go-to in achieving a variety of goals.
  • One way to employ A/B testing of time is to serve immediate and delayed versions of the same pop-up. The delayed version gives consumers a chance to check out what your website is all about and determine if the value you can provide is worth committing to. Be sure that your value is clearly stated in the areas a consumer can search within the controlled time delay.
  • A good way to determine the delay would be to review the average time spent on a page. Try a 50/50 split where variant A receives the pop-up immediately and variant B receives the pop-up 20 seconds after being on the site.

  1. Verbiage
  • Your beautifully buttoned-up website has been live for some time now, but your conversion rates have been mediocre at best. Why? (Remember what happens when you assume!)
  • Testing website copy is a simple and effective method of improving conversion rates. Here you can test headlines, calls to action, length of copy and much more. Try switching it up and invoking a sense of urgency in your calls to action with words like HURRY and EXCLUSIVE.

  1. Placement
  • I’ve come to your website and i’m ready to commit. I want all that your weekly e-special has to offer. There’s only one problem – I can’t find the e-special sign-up box that you’ve unfortunately tucked between the weather and some other microscopic links at the very top of your homepage. I’m gone, forever. (You’ll miss me!)
  • Of course, sign-up forms are not the only thing you can test the placement of. (Seriously, you can test anything.) We recently tested the placement of photos within a homepage grid and increased conversion rate by eight percent! (Read more here.)
  • Take a good look at your data and base your placement tests on the most active areas of your web pages. Testing the placement of sign-up forms, buttons, links and more can freshen up your website and improve your conversion rates.

  1. Color
  • Isn’t it a fact that the color red makes us hungry? (We’ll leave it up to you to determine if that is true or not.) Either way, the colors on your website could certainly affect your conversion rate. Consumers may have a hard time reading the light yellow font describing a BOGO deal on your white background. (Silly, but it happens.)
  • Review your goals and test colors based on what you want to achieve. Are you targeting families at the beach? Try testing bright colors that compliment each other and grab the consumer’s attention to make a lasting impression.
  1. Creative
  • We’ve already chronicled the importance of good website photography here and here, but if you don’t believe us test it out! Testing a creative aspect of your website can include images, widgets, logos, buttons and much more.
  • Test complex versus simple designs or look to market trends to generate possible hypotheses. A recent trend that has peaked my interest is wielding UGC, or user-generated content, to improve conversion rates. Try testing traditional images against UGC that incorporate your product or service to help humanize your brand.

Now, combine any of the above elements, sprinkle in some magical developer dust and voilà! Multivariate Testing.   

Multivariate testing is a technique for testing a hypothesis in which multiple variables are modified. The goal of multivariate testing is to determine which combination of variations performs the best out of all of the possible combinations.

An example of multivariate testing would be testing the color and placement of a new button.

  • Variant A would receive a blue button on the left.
  • Variant B would receive a blue button on the right.
  • Variant C would receive a green button on the left.
  • Variant D would receive a green button on the right.

Whether you’re testing two or more variants, testing can get tricky. Remember to always form your hypotheses around data and test in a high-traffic season to see the best possible conversion rates.

Source of Graphs: State of Online Testing Survey 2015, WhichTestWon & Marketo

Share via
Copy link
Powered by Social Snap