I hear this question being asked all the time — even from seasoned marketers who know how to run an A/B testing program. I see people blogging about what to test, which can be helpful, but it overlooks the most important aspect of testing: learning about the test subject (your visitor, prospect, or customer!).

Your Web site is a series of design and messaging elements comprised of pixels and code. Your site is powered by Web servers, and delivered over wired and wireless connections. There is a lot of technology that goes into making it all work, but ultimately its purpose is to help people accomplish their goal (and I cannot think of a single exception to this statement).

When you ask the question, “What should we test?”, your focus is on the Web page element and not on the person who will generate your test data.

Headlines, CTAs, form fields, checkout flows, images, product copy, security assurances, trust builders, value propositions, sub-heads, cross-heads, navigation, social widgets, sign-up forms, colors, special offers, and subject lines are all things you should be testing. But if you develop your list of test ideas without explicitly calling out the human part of the equation, you’re not going to achieve the same results or return on your time/money as the person who focuses their test ideas on what they want to learn about the visitor.

So don’t ask, “What should we test?”.

Instead ask, “What do we want to learn about our visitors?“.

This seemingly benign change in wording will transform your entire testing program. Here’s how:

1. A focus on learning will spark a lot more test ideas than jumping straight into creating a tactical list of ideas, because there are so many ways to learn about your visitors. In your business, you don’t jump into tactics without first having a strategy in place, do you? Strategy drives tactics; desired learning drives test ideas.

2. A focus on learning will ensure that you learn something on every test. Testing without a customer hypothesis puts the focus on getting a winner — which we all know only happens a fraction of the time. A flat test result is still a success if you set out to learn something about what does or does not influence your visitors.

3. When you learn something about your visitors, you’ll want to learn more (follow-up tests!) — which fuels your testing engine and builds your understanding of the people coming to your site.

4. When you learn something about your visitors, you can immediately apply that learning from your Web site to your other marketing initiatives and even to product development. You can leverage new learning across your entire business.

5. Once you develop a solid understanding of your visitors/customers, you’ll be able to predict their behavior — which is every marketer’s dream. Being able to predict behavior makes selling much, much easier (just ask any successful salesperson!).

Start by asking, “What do we want to learn about our visitors?“.

It’s a simple change to the way you think about testing, but the results will amaze you.