The Scientific Method for Finding the Perfect Landing Page

October 5, 2016

Fiverr, the creative freelance community where every service starts at $5, could see in their website analytics that visitors would come to their site but would bounce from their landing page. They just didn’t know why.

Fiverr’s Growth Specialist Yoav Aziz had a theory: what if users were leaving the landing page because it was too busy? So he created a second landing page that included more white space, highlighting the call to action button at the top of the screen. Aziz ran an A/B test where he showed both the original landing page to half of the site’s visitors and the variation to the other half. The variation received 23.4% more conversions than the original. Success.

But to get to their final landing page, Fiverr had to run over 400 A/B tests. They didn’t survey their customers first about specific pain points and instead experimented with every element blindly until they cracked it. You can focus your testing by asking customers specific questions about how they use and intend to use your product.

The Science of A/B Testing Your Landing Page

A/B testing resembles a lab experiment. It’s a step-by-step process:

  • Formulate your question. Gather information about a phenomenon that needs an explanation (i.e., 35% of users bounced from your landing page) by looking at your site’s analytics. Figure out what question needs to be asked to find that explanation. For example, when the team at Solitaired achieved a 41% improvement in bounce, they first saw in their analytics that certain browsers and mobile devices had a higher bounce rate, which led them to hypothesize and ask questions why this was happening.
  • Determine the hypothesis. Decide what you’re going to test by asking users a specific question about their site usage. Also look at how you’re going to measure success, whether it’s higher conversions, increased sales, or some other factor. Finally, craft your hypothesis statement. Michael Aagaard’s formula for an A/B testing hypothesis is a useful one to follow: “By changing ____ into ____, I can get more prospects to ____ and thus increase ____.”
  • Isolate one variable. If you A/B test both the copy and colors of your landing page at the same time, there’s no way of knowing which one created the increase or decrease in conversions.
  • Create a control version and a variation. When you A/B test, you simultaneously show two versions of your landing page—the original, untreated version to half your visitors, chosen randomly, and the treated version to the rest.
  • Collect results. Since you already chose what factor determines the success of a page for you, look now at what each page yielded. The version that led to more conversions or sales is the winner. It should be your page and will be the starting point for your next A/B test. After Fiverr determined they needed more white space, they took their variation landing page from their first A/B test and used it as the control in their second, which looked at the location of the call to action.

A/B testing is a scientific process driven by data. The hypothesis, though, is sometimes just a guess at what the problem is. The way to fix that is by crafting a hypothesis around what users say about their experiences. This ensures you’re not going by “gut instinct” anymore—you’re going by data.

Recommended Read A comprehensive list of the Best A/B testing tools

From the Survey to A/B Testing

In order to isolate a variable to A/B test your landing page, you need to ask visitors directly what they think through surveys. From there you can craft a specific hypothesis and begin testing your landing page.

 

Highlighting Specific Features

You may be seeing a high bounce rate or even a high average time spent on site because users can’t find what they’re looking for. One of the first questions you should ask your visitors is “Did you find what you were looking for?”

Screen

If visitors say “No” or even “Yes, eventually,” you need to reconfigure your landing page. The hypothesis for your A/B test is:

“By displaying the information for this feature more prominently, I can get more prospects to click on my call to action and thus increase conversions.”

Your A/B test, then, should include a control version, which is your original landing page, and a version that displays the feature in question more prominently. This may mean putting it higher on the page, changing the colors around it, or directing attention to it through visual cues.

Career Point College was seeing a high bounce rate for their site’s visitors and not enough conversions. They hypothesized this was happening because the important stuff was buried on the page. They created a new landing page, one with a more obvious form and no navigation panel, to A/B test against their old one. The new landing page saw a 336% increase in conversions.

Screen

 

Prioritizing Information to Grab Conversions

Screen
Screen

You might also be seeing a high bounce rate because visitors are expecting one thing and receiving another. So ask them directly “Did this page meet your expectations?”

You can then structure your A/B test around what customers respond with. If they came to your landing page expecting to see a certain thing, like the price of your product or testimonials from customers for example, you should make that thing available. It may be that the information they expected to find on the page was buried among useless factoids. In which case the hypothesis for this A/B test is:

“By condensing the information on the landing page, I can get more prospects to click on my call to action and thus increase conversions.”

Screen

Highrise, a CRM app for companies which tracks tasks, contacts, and notes, A/B tested with a more minimal landing page. They compared their original landing page with a shorter one that included the crucial information from the longer version but without most of the graphics. The shorter version increased paid signups by 102.5%.

 

Removing Unnecessary Design Elements

Screen
Screen

Sometimes it’s just a small thing holding you back, like unclear copy or distracting colors. You can ask your visitors what they think with a pointed survey. Asking “Do you find this page easy to understand?” leads you to the next part where visitors can directly fill in what they find issue with on your page.

If visitors respond with “No—there was too much happening on the page,” or “No—I don’t like the layout,” you can craft a hypothesis around that specific pain point.

“By removing X design element on the landing page, I can get more prospects to click on my call to action and thus increase conversions.”

When VVAA, an online association of Dutch physicians, decided to make a small design tweak to their landing page, they ran an A/B test for it. Their hypothesis was that removing the horizontal lifeline from the header image would lead to more users clicking on their call to action. They tested only one variable, and their treatment caused the number of clicks on the call to action to rise by 7.8%.

landing-page
landing-page

The call to action was clearer in the variation because it didn’t have to compete for attention with a similarly colored, large graphic. It was a small change that didn’t require too much reworking, but it led to an immediate increase in clicks.

Landing Pages Influence Both the Top and Bottom of Your Sales Funnel

When you use surveys, you hone in on the intentions of your visitors. You can then build A/B tests for your landing page to change elements and encourage visitors to convert to users.

The ultimate point of A/B testing your landing page is to determine the changes you need to make to cater more effectively to your would-be customers. Doing so will lead to higher conversions on your call to action and should reduce your bounce rate. But the real test of success isn’t at the top of the funnel, it’s at the end.

A more effective landing page will lead to more sales overall.

Do you want a free Survey & Feedback Software?

We have the world’s most versatile user research & feedback survey tool starting at $0

Do you want a free Survey & Feedback Software?

We have the world’s most versatile user research & feedback survey tool starting at $0

Do you want a free Survey & Feedback Software?

We have the world’s most versatile user research & feedback survey tool starting at $0