In my Part 1 post about optimizing customer testimonials, I laid out my CAPP framework for thinking about the various aspects of a testimonial and the variables you might improve on and/or test.

In this post, I want to illustrate how I would apply this framework using a multivariate test to see which optimized combination of factors would increase conversions the most.

To recap: my framework for optimizing the persuasiveness of customer testimonials included Content, Authenticity, Placement, and Presentation. Each one is an axis along which you can think about, and optimize, your testimonials.

For my imaginary multivariate test, let’s say I have a landing page that already converts pretty well. It follows all the industry best practices, and I’ve tested my way to a pretty optimal page layout, call to action, headline, copy, etc. I’ve also proved the hypothesis that having a testimonial is better than not having one.

Now, I want to find the absolute best testimonial, presented in the best possible way! Each category within CAPP will become a set of variables within my test.

Content

Starting with Content, let’s keep it simple and say that I have two testimonials that are similar in word count and general content and style. I think they’re both persuasive, but of course I want to know which is best for conversions. So I’ll have Content 1 and Content 2 as a variable set.

Authenticity

Moving to Authenticity, I’m going to test 3 different attributions:

1. First Name and Last Initial
2. First Name, Last Initial, and City and State
3. First Name, Last Initial, City and State, and Product/Solution purchased

I’ll refer to this variable set as Attribution 1, Attribution 2, and Attribution 3.

Placement

Placement will tell us where on the page the testimonial will be most effective. For this category, I’ll test 3 again:

1. Center of page (interrupting body copy)
2. Bottom of page (near Call To Action)
3. Right column

Presentation

For this category, I’ll get a bit fancier. I want to learn about font type, font size, and color of background, which means that I may need to do a bit of follow-up testing after this multivariate test is complete. Here are my variables:

1. Serif font, 14pt. size, no background color
2. San serif font, 12pt. size, beige background
3. Serif font, 12pt. size, blue background
4. San serif font, 14pt. size, blue background

Final Recipe and Permutations

Our ambitious project to learn about which combinations of Content, Authenticity, Placement, and Presentation work best has quickly racked up 72 variations! That’s the nice thing about imagining this scenario: I don’t have to pay for the necessary traffic 😉

And so on...

Conclusion

The imaginary results are in. My target audience has spoken, and after some pruning of variations, they have chosen variation #39, which was:

• Testimonial content #2
• First Name, Last Initial, and City and State
• Right column
• Serif font, 14pt. size, no background color

Cool! The right column placement was a surprise, as was the absence of background color…I thought that was a “best practice” 😉

My imaginary lift on conversion rate was 114%, and since this was a paid search landing page, I also saw healthy improvements in ROI and ROAS. I could get used to this!

What follow-up tests would you run? Is it time to dig into those nitty-gritty design elements like fonts and font sizes? Or, should I go get new testimonial content to challenge my winner?