I’ve loved watching people shop ever since I was a kid. Whenever my parents went to the store, I begged to go along. When I got my own car, I spent hours at the mall, but I didn’t buy anything. Instead, I watched other people buy. I watched the displays and how they were set up—with bright lights strategically placed right at eye level.. I watched people pick up products and ponder purchasing, then put them back down and walk away. I watched the tiny shifts in facial expressions when people were impressed with products, and the disgust on their faces when they felt the price was far higher than the quality.
When I got into the world of e-commerce I found I enjoyed exactly the same thing. When I ran my first few successful advertising campaigns, every new user visiting the website, every click, and every sale was exhilarating. I left my phone on vibrate so I could hear the delightful buzz of every last contact form notification—day and night—for days on end.
It was thrilling—except for one thing, which was incredibly frustrating.
There wasn’t as much to watch.
Because you couldn’t see anyone at the other end of the computer screen, you were like a mall vendor setting up displays and then sitting in the breakroom all day, never knowing how people responded to your displays until you counted the cash drawer. You couldn’t watch a user’s facial expressions or see them pick up and study your product. You couldn’t see anything. And so, much of my psychological studying was useless.
But I couldn’t simply accept that every online store had to be run blind. There had to be a way to learn about what the users really wanted and increase sales.
Enter A/B testing.
A/B testing is a way to test what your users like most out of multiple options.
Think of A/B testing as a way to test option A against option B for outcome C.
With A/B testing, you can get data-driven answers to questions like this:
- Out of these two headlines, which one causes more people to buy the product?
- Out of these two home pages, which one gets people to click through to a second page more often?
- Which of these two menu designs get users to their final destination the fastest?
- Will more users fill out the lead form if we put the lead gen pop-up at the start or the end of the quiz?
There are many online softwares that allow you to perform A/B testing, and we typically will either write our own custom code or use Google Optimize.
A/B testing allows you to define an outcome. Outcomes are often measured in Google Optimize as tags. For instance, a tag may be triggered when a user fills out the lead gen form. You can combine multiple desired outcomes to get a weighted average. For instance, which option gets the lowest bounce rate AND the highest conversion rate?
A/B testing also allows you to assign two or more versions that will be assigned to your users at random. For instance, if you are testing two different variations on a headline, you can display headline A to 50% of your users and headline B to 50% of your users randomly.
It is very important that these tests are randomized and run at the exact same time. The reason is that different time periods may result in dramatically different conversions. For instance, if you test the sales of a health product in December, you often won’t get very good results on average, because during the month of December most people are concerned about spending time with their loved ones, and running from holiday party to holiday party eating delicious food. They are not particularly concerned about their health during this time.
However, in January, many people make New Year’s resolutions and return to the gym.
If you were to run headline A in December and headline B in January for your health product, you may mistakenly think that headline B performs far better. But the difference in conversion rates may have nothing to do with your headlines at all, rather the time period.
So, with A/B testing, you can test headline A and B simultaneously, and it doesn’t matter whether it is December or January. You will test the headlines equally split between equal randomized participants.
A/B testing is evolving into the new way internet e-commerce is managed.
In the past, you would simply design the best website you can and hope that it performs well. You could optimize, but you may not be clear whether your optimizations made a positive difference or if you were experiencing the December / January effect.
Now, the most savvy companies use A/B testing on a monthly, weekly, and even daily basis to keep honing in their sales processes.
It’s better for businesses, and it’s better for customers, too. A/B testing is a way to figure out what customers really want and show that to them.
If you’re curious about how to implement A/B testing, reach out to us! We’d love to hear from you. Get in touch with us at our contact page.