Landing Page Testing: Testing for Impact, Not Variations

• Tweet
• Buffer

Last column, I shared the following test, which I found "in the wild," and asked readers to identify the variables. Hoping that instead of testing all the variables, you could narrow it down to the most meaningful variables so that you minimize the time and resources needed to complete the test.

Depending on how you want to define your variables, these two pages have around a dozen changes.

Here is my list:

1. Layout orientation. As I mentioned last time, one is laid out horizontally (A), the other vertically (B).
2. Different taglines (UVP). A) Make every visitor count. B) A marketplace of landing page designers.
3. Call to action headline. A) Improve your conversion rates! B) Get started here!
4. Call to action header background color. A) Darker shade of blue than B.
5. Call to action graphic. A) No arrow pointing down. B) Has an arrow pointing down.
6. Call to action button copy. A) Invite me. B) Create account.
7. Call to action button color. A) Blue to match header. B) Orange for higher contrast.
8. Call to action button size. A) Full size of form. B) About two-thirds the width of the form.
9. Call to action form field. A) Four fields to complete. B) Five fields to complete with added company as optional field.
10. Required text for call to action. A) Has none. B) Has it at top of form.
11. Point of action assurance. A) Privacy guarantee above call to action button. B) Below button.
12. Testimonial. A) To the left of the form on top. B) Across the bottom of the page.
13. Copy. A) Three paragraphs of text. B) Adds fourth paragraph about "A fresh pair of eyes." Also, the order of the paragraphs is changed.

Most of the time, when I see such a list of variations, it would seem that the business is testing for variations first and impact second. Testing has to start with the "why" first. Why will this matter to my customer? Why will changing the layout from horizontal to vertical change their feeling about buying from us?

This goes back to what we all learned in elementary school. You start a test with a hypothesis first. Instead, most companies today throw variables at the wall, sometimes not even aware that they are changing variables and all they look for is what version gave them a lift. This is a great strategy when you have endless resources and time to wait for tests to complete, or when you have a page that is converting at around 70 percent-plus and you want to start fine-tuning it.

So what are the three variables I would test from these two landing pages?

1. Different taglines (UVP). I would want to test at least three or more very different versions to find what resonates most.
2. Call to action form field. Whenever we can reduce a form field by 20 percent from five fields to four fields, that should have an impact.
3. Call to action headline. We want to make sure our offer is clear and this should summarize it quickly.

Did you identify these as the big ideas to test first? Have you been testing intelligently or have you been guilty of slice-and-dice optimization? Testing these three variables first enables us to complete our test in 18 days versus 108, based on the number of variations that need to be tested.

What about testing the call to action button and other variables?

Sure, I would eventually want to test that, but the first goal should be to make sure our offer is viewed as relevant and valuable enough to our visitors to want to complete the call to action. Have you ever stopped yourself from clicking on a button because it was blue and not orange if you found the offer relevant and compelling?