This is the subhead for the blog post

We’ve all heard that A/B testing is the foundation of good landing page optimization, but what does that really mean? How long should you test? What should you test? How can you set up a test to ensure you learn something from it?

I’m glad you asked! Let’s dive in.

Before you begin, remember…

A test is not worth running if you don’t have a concrete Hypothesis. What are you hoping to prove (or disprove) with your test?

A test is also not worth running unless you know you’re going to get actionable data volume from it. Test pages where you see significant campaign volume. Testing small changes may yield flat results, and you’ll never reach statistical significance.

How long should a test run?

On average, we’ve found that a test should run for at least two weeks to allow for statistical significance. If the test results are flat after 2 weeks or so, it’s probably best to kill it. A test that is limping along will probably never yield discernable results. Move onto something new.

Keep in mind that the 2-week recommendation is for a testing a single variation; testing additional variations will slow down results. Before you throw another variation in the mix, make sure there is significant value in doing so.

What should you test?

There are many variables on a page that can affect performance. Here are the top 6:

1. Hero image/banner

2. Nav bar (top or side)

3. CTA

4. Imagery

5. Form (# of fields, size, position, etc.)

6. Content below the fold – should you consider moving it up?

Past tests have taught us to:

Now, we’re not encouraging you to take these as gospel and skip testing, but here are some adjustments we’ve recommended over and over again as a result of our test data:

Eliminate distractions: Eliminate all distractions so user focuses on the goal only. This might mean that you remove nav bars, remove secondary messaging, or remove any additional links that take the user away from the conversion goal.

image001 image003

 

Clarify messaging: The message of your landing page should be crystal-clear: buy the product, start a free trial, sign up for a newsletter, etc.

image005 image007

If your hero copy is more than a sentence, don’t be afraid to use bullet points. And if your product or service is a relatively complex one, show an image of it in use — or give a step-by-step description of how it works so the user isn’t left guessing.

 

Make a strong CTA: Your CTA should be clear, direct, and action-oriented. (Buy, try, sign up, etc.) Keep your copy short and sweet, make sure the placement is front and center, and make sure that if you have more than on CTA on the page, the messaging is consistent.

image009 image011

Establish clear value: Why should the user follow through on your CTA? Because you’ve made the value of your product or service apparent and established how it’s different from its competitors.

But don’t just tell the user that; add trust symbols: reviews, testimonials, etc. Try different placements for these, too — do they work better in the hero image? Right above/below the CTA? Test it out.

 

Forge an emotional connection: Why should the user care about what you’re offering? Let the user know that you understand the problem or issue that the product or service will solve. Make the copy and the imagery personal to the audience. And reflect that, as much as possible, in the CTA.

image013

 

Simplify fill forms: Would it be nice to know a user’s household income? Sure. But that might also throw up a stop sign in the form-fill process. Forms should only include necessary fields: name, email, phone number, possibly some geo-specific info — otherwise you risk springing a big leak in your conversion funnel.

image015

 

Add video: If a picture’s worth a thousand words, think of the value a video can add. Do you have relevant video content? Maybe it’s video of the product in use, or a glowing testimonial about your service. Try adding it to the page where the user is compelled to view it.

image017

Final thoughts

Testing can be frustrating: iterative, lengthy, or even ambiguous. But if you follow the guidance above and keep an open mind before you see the data, you can come out of your tests with awesome insights that can translate into beautiful, clean, conversion-friendly landing pages.

Good luck!