A/B Testing: We Share Some of Our Most Frequently Asked Questions

Corte Swearingen | May 02, 2016 Comments
FAQ
As I work to educate customers about how effective split testing can be for improving conversion rates and lead generation, I find myself answering very common questions. In this blog post, I’ll share these questions along with my answers.

  1. What is, for you, the basis of conversion optimization?
    I don’t really care for the term ‘conversion optimization.’ I prefer the term ‘Marketing Optimization.’ It’s a better descriptor as to what we are trying to achieve. For me, I really don’t care what a person’s goal is when they come to a site, I just want to ensure that whatever that goal happens to be, they are able to achieve it before they leave. Their goal could be to fill out a form, check pricing, purchase something or to share a photo of their favorite product with a friend. With a practice like split testing, we can optimize for all these various actions on a single website. For me, this is the basis of marketing optimization, to understand all the important macro and micro-conversions of a site, and use split testing to improve everything.

  2. What have been the biggest changes on conversion optimization over the last three years?
    The testing platforms have evolved and become much more sophisticated. We can now better segment our test results and even create specific audiences conditions for our tests. We can test one value proposition message to a visitor from New York, and a completely different one for visitors in California. This allows us to better understand visitor intent/behaviour and to fine-tune our marketing messages and ultimately improve business value.

  3. What is, at the moment, the biggest challenge in the field of conversion optimization?
    What I’m starting to see out there is that a lot of people are calling themselves ‘experts’ when it comes to onsite optimization. Running a deep and broad A/B testing program requires a lot of people working together to formulate a proper testing roadmap and then to execute on that roadmap in a way that maximizes results. This is the same thing that happened in the SEO world many years ago – so many people considered themselves SEO experts, but very few of them really had the depth of knowledge to do it properly. It’s crazy that there are still people out there trying to trick the search engines. It’s such a short term strategy.

  4. There are a lot of tools to use for conversion optimization, such as A/B testing tools, heatmaps, and scrollmaps. For you, what are the most important tools besides Google Analytics in regards to conversion optimization?

    I think it is important to divide out these tools into two basic categories; quantitative tools and qualitative tools. Quantitative tools like heatmaps and google analytics answer the ‘What’. What are people doing or what are people clicking on? They are important tools, but not very useful by themselves. For example, we can use Google Analytics to measure the bounce rate of a page, but knowing bounce rate doesn’t provide any information about how to improve the page. Qualitative tools, on the other hand, help answer the ‘why.’ Why is someone clicking on this link or why are they making choice A over Choice B? By using both quantitative and qualitative tools, you’ll start to get a much better picture of where all the friction points are on your site. That being said, my two favorite qualitative tools at the moment are Qualaroo and www.usertesting.com. I don’t have any affiliation with these tools, I just use them on a regular basis and like them. Qualaroo allows me to run exit surveys on a site. When I set up an exit survey, I ask three simple questions:

  1. What is the purpose of your visit today?
  2. Where you able to complete the goal of your visit?
  3. If not? Why?
I think you can see how getting answers to the above questions can be a rich source of actionable insight. I also like running user testing using a tool like usertesting.com. This allows me to define a few very specific tasks I want to watch people perform on a site. I then get videos with audio that I can study to better understand what to improve.

  1. Through A/B testing you can test emails and webpages. What are you testing further with A/B testing?

    I think it’s important to realize that creating a testing culture should go beyond the specific platform. When this is done, anything can be split tested, from your direct mail pieces to email subject lines to your PPC landing pages. In fact, the concept of split testing was first done with direct mail many years before the Internet became available.

  2.  Are you entirely data-driven when we talk about conversion optimization?

    You might be surprised to hear this, but no, I’m not 100% data-driven. I do feel like intuition can play an important role in conversion optimization, as long as those intuitive ideas are rigorously tested. Numbers and metrics can sometimes fool you – you have to be careful about relying 100% on metrics. That is why I mentioned using more qualitative sources of data, so you can better balance all the numbers coming out of your web analytics program. As human beings, we tend to see correlations in the numbers that aren’t always there. Take a look at the below link for a humorous example of how correlation does not imply causation!

    http://www.tylervigen.com/spurious-correlations

  3. A lot of websites don't have a traffic problem, but a conversion problem. Yes or no?
    A lot of websites have a traffic problem AND a conversion problem. Many sites spend gobs of money driving large amounts of traffic that aren’t relevant to the company’s value proposition. A low-traffic website can still make millions of dollars as long as the visitors coming in are highly focused. Of course, with a typical ecommerce conversion rate of 2-3%, there are always opportunities for improving onsite optimization.

  4. Has the execution of A/B testing become increasingly important?
    Yes. As tests become more sophisticated, it now can take a team of front-end developers, programmers, marketers and data scientists to truly create a strong testing program. While test execution is extremely important, so is having strong test hypotheses and creating a proper foundation of gathering data and understanding the site’s potential improvement areas.

  5.  Is the use of A/B testing tools & heatmaps nowadays a must?

    Yes. Otherwise, you are just guessing at what to improve, and will have no quantitative basis to see which of your ideas are truly providing incremental business value.

  6. Which two online marketing items are on your agenda for next year?
    Continuing to speak and educate companies about the power and importance of allocating some of their marketing budget to online optimization efforts and to continue helping companies learn the art and science of split testing. It’s a smart and repeatable process that I’ve seen work time and again across every industry and market type.

Write a review

Authors

  • Ahmed Okour
  • Paul-Ross-Blogger-Bio-Pic
  • Rachel-B-Web
  • Courtney-V4
  • scottstiles
  • stu-3
  • Joseph Gustafson
  • Vince Scarlata
  • Tim Ahlenius
  • Staff Blogs
  • Shawn Griffin
  • Americaneagle Partner
  • Nick Goodrum
  • Missy Hildebrand
  • Mike Avello
  • Corte Swearingen
  • Adrian Krzeszkiewicz
  • Emily Stark