Recalling The Scientific Method: A/B Testing & CRO Crash Course

Starting Off With The One We've All Heard Of

Today, I want to share my most recent thoughts and insights on two important concepts for digital marketers to be very comfortable with: A/B Testing and CRO Optimization. Let's get started!

So, on a basic level, A/B testing is essentially a controlled scientific manipulation - where there is a "control" and a "manipulated variable". Ever hear the term "treatment" in a research context, especially in psychology-focused work?

In marketing, the idea is that a firm will produce two versions of something (a video, a website page, etc.) to see which one lands best with customers and, ideally, creates the most conversions. There can be as many manipulations as desired, like the image below depicts, but for this context, I will discuss it with the parameters being there is an "A" version of something and a "B" version.

A favorite professor of mine coined the phrase "nailing the process". In terms of how to do just that, here is my take.

1. Write down the hypotheses. (no, don't verbalize them, actually write them down). Is this testable? Does it improve marketing efforts? Is a conversion problem being solved?
2. Run the tests! Create both manipulations (say, two different versions of landing pages on a website designed to capture customer attention) and have traffic direct to both, randomly, so both receive equal visitors.
3. Analyze the results. Decide where to analyze (SPSS, Excel, Google Sheets, wherever!) based on outcome goals, and search for variations. Or key differences. Or flags. Or any other differences between both, and examine for experimental errors. Could these differences have been unintentionally manufactured in?
4. Once you are confident the differences between groups are likely organic, decide what it means! Evaluate the insights and decide on next steps strategically based on your customers' response.

The takeaway here is relatively straightforward - A/B testing is incredibly useful. The insights it provides allow for optimization of things that mean the most to customers - how easy a site is to use, how appealing titles are, how content is received. Improving that user experience (UX) is how conversions begin to roll in.

Constant improvements to the experience can be provided, as A/B testing isn't a one-time deal. It is an iterative process that takes into account the expressed wants and needs from customers in as real of time as a company wants.

CRO - We Know We Need To Optimize, But How?

Hubspot does a better job defining the concept of conversion rate optimization than I: the process of increasing the percentage of users or website visitors who complete a specific action to increase the number of leads you generate. Conversions mean everything to digital marketers - these are the actions we as a company want our customers to take. Booking a trip, ordering a product, purchasing a subscription.

When thinking about conversion rates, the higher they are, the better. Say you have a conversion rate of 6.5% for products for sale on your "merchandise" page. This means 6.5% of the visits to this page led to conversions (i.e. sales). If you are unhappy with your conversion rates, that is when CRO comes into play.

So, where would you start when it comes time to optimize? I again refer back to Hubspot's take, where they offered four great suggestions on sections of a website that most benefit from CRO.

• Homepage (example: that digital "assistant" who pops up with a chatbox to answer questions)
• Pricing Page (example: an easy, seamless, quick option to get a quote)
• Blogs (example: the call-to-action (CTA) designed to stimulate..well, action)
• Landing Pages (example: a page containing a sample or snippet of some media, this entices users to download and/or pay for the remainder)

This brings us back to the idea of A/B testing. I mentioned that the underlying goal there is to increase conversions. Simultaneously, CRO is designed to, again, increase conversions. Both are tools that aid in generating leads and creating sales.

A/B testing is the system used to carry out conversion rate optimization, and to find out whether or not what we are doing is working. Whether it's adding new CTA's to blogs, changing or modifying the location of a "Buy Now" button, or any other development to try and optimize conversions, A/B testing those changes against a control is how marketers gain insight into what is happening. The use of these tools together is how conversions are created.

When deciding how to best optimize, it starts with the numbers. Decide which key metrics mean the most to you and your team, and figure out where those come from. These are the numbers that need to improve: figure out what ideally needs to happen on the customers end for the outcome to be favorable. And then, set your focus to your people, your users.

Understanding the "why" behind why users want what they want is harder. To optimize most effectively, having quantitative data is good, but qualitative to supplement is even better. Landing pages with surveys, mailer lists, and QA checks all will help you understand what the next steps are in driving more traffic and creating more conversions from both existing and new customers.

This homepage below illustrates CRO beautifully, as it should. being a firm specializing in improving conversions for other firms. Immediately I was hit with a tidbit of information, a download to follow, a digital assistant, and a user form to fill in things about me. Well done, Heap!