How testing can drive your goals (and save your sanity)

 

The digital marketing world moves quickly, as do the expectations of your users. So, when you’re managing content for one of your digital properties, it can sometimes feel like a constantly moving feast. How do you keep up?

 

The answer is to integrate testing into your overall content plan. 

 

•          Is your brand or product department considering a change to design or messaging? Conducting tests can provide leading indicators of what will resonate best. 

•          Are your audiences diverse in their demographics or needs? Testing can help you find common themes that resonate across your audiences. 

•          Is your leadership insisting on “consistency" regardless of the impact on your goals? Running quick low-risk tests can help prove your case for performance-focused marketing and will help you gain their confidence in your approach.

•          Are you reconsidering a website redesign or big media campaign? Testing can be a cost-effective way to test concepts before larger investments are made. 

 

So, where to start? Think about your content as a recipe you are adjusting to bake the perfect cake. You have multiple ingredients that you need to mix together in the right way to get the result you need. 

 

To execute organized, successful tests, follow these steps:

 

1.  Create a "test plan” by:

 

•          Creating clear goals for your digital properties, then setting goals for the content you plan to test within it. Ask yourself: what should this piece of content do to help you achieve your broader digital goals? 

 

•          Considering what you want to learn from your test. Maybe you simply want to drive more sales. Perhaps you are interested in what content different audiences respond to. Maybe you’re simply trying to learn more about who your audiences are. Think about what knowledge will help you be most successful, and then create a test to help you achieve that knowledge.  

 

•          Creating your hypothesis - this could come from your research department, or you may just have a hunch that people will share video A more than video B. Be sure to start out with a hypothesis, so you can prove yourself right or wrong.

 

2. Choose a testing approach - there are typically two core approaches:

 

1. A/B. These are simply two versions of the same page or page elements that are tested at the same time. Traffic is sent to both versions and the response is evaluated.  A/B testing can help organizations become familiar to the idea of testing - they’re low risk yet can reap significant results in short time periods, especially if you have higher site volumes. 

 

(Note: Google’s Content Experiments run an “A/B/N” model allows you to test up to 10 versions of a single page, each delivered to users from a separate URL. It’s an effective way of testing and can generate fast results for businesses focused on lead generation or sales.)

 

2. Multi-variate is popular in organizations that have experience with testing. This tests multiple elements at the same time on the same page. This method can be a fast way of getting to your ideal “recipe” - especially if you’re testing landing pages or emails.

 

3. Select the testing platform (there are a plethora of tools available, ranging from free and basic to expensive and sophisticated - choose the one that fits within your budget and technology parameters but also helps you achieve your testing goals). 

 

4. Decide on the location and time of your test. Start small and build your testing volume and complexity over time.

 

5. Test, evaluate and share your learnings with peers and senior management.

 

6. Repeat.

 

Looking for a good place to start? Focus on where people will take action so you give yourself the best chance of seeing response – your “ingredients” here might be “click here” buttons, form style, placement, words/color/style.

 

 Above all, enjoy it – even the most non-geekiest of marketers begin to get a kick out of seeing a test result prove their theory right....try it, you may like it!