If there’s anything I learned in my marketing research class last quarter, it’s that marketing contains a lot more science and experimental aspects than one would think. Thankfully for this particular marketing student (who is not very good at science, *cough* physics *cough*), the focus is more on the analytical side of things, with an emphasis on hypothesis testing. Although my memory of seventh grade science is pretty minimal, I do recall the scientific method: ask a question, construct a hypothesis, test the hypothesis, analyze the data, and draw a conclusion based upon these results. This is where A/B testing, or split testing, comes in. A/B testing “takes the guesswork out of website optimization and enables data-backed decisions that shift business conversations from “we think” to “we know.'” It allows for simple, easy, and inexpensive decision making backed with hard facts and data. Who doesn’t want that in our fast-paced, digital world?
Today, a massive amount of marketing (especially digital) relies on testing various hypotheses and optimizing various aspects of the customer experience based upon these results. What kind of content is more interesting to my readers? What landing page makes potential customers stay on my website longer, and explore more? Should I include social media links on my webpage? These are all things that A/B testing can find out, and companies like Optimizely (who collected a giant list of all the different things to potentially A/B test) specialize in helping businesses and websites discover what works best for them in the digital world. It’s your classic seventh grade science experiment, except with controlled and treatment variables that feel a whole lot more relevant and can majorly effect website traffic, revenue, and more. A/B testing is widely used and hugely successful: ComScore used A/B testing on their site and saw a 69% increase in their conversion rate from the original web format (I think it says a lot that even marketing analytics companies are all about A/B testing, that’s how you know something’s good).
Few things are more frustrating for customers, like myself, than not being able to navigate a website to find what we want (we can just go somewhere else, after all), or finding out that something you want is way, way out of your price range (which is essentially my daily online shopping experience). Guess what can help fix these problems? A/B testing. It allows you to tailor your website to your customer, so they remain loyal to you, and it gives businesses valuable feedback on things like pricing and how to better promote a sale. Remember when Nordstrom changed their half-yearly sale to a “clearance sale” last November? A lot of people didn’t know! Even though the company had used customer feedback and decided to change it (gained through various customer-service tools like social media and reviews, which make retail way more exciting), they forgot to communicate properly. They had to send out an email the second day of the sale clarifying they had just changed the name, as shown in this screenshot.
A/B testing could have helped (i.e. using a landing page with clarification and one without, or emails, etc.), and maybe brought in more revenue for the first few days of the sale.
It’s truly the scientific method in digital format: ask a question (Why is my bounce rate so high?), form a hypothesis (Maybe if I did X to my landing page, I could better funnel visitors through my website), test the hypothesis (A/B testing!), analyze the data (thank you Google, Optimizely, and more), and form a conclusion and optimize my site (I’ll use website format B, because it decreased my bounce rate and increased sales). Simple, cheap, and way beneficial for customers and businesses alike.
Also, I would like to point out that the badass Ms. Amelia Showalter also enjoys the ever-hilarious and inspiring Leslie Knope, which increases my admiration for her more. And it only furthers my opinion that Parks and Recreation is one of the best shows on television.