How to Lay the Foundation for a Data-Driven Business With A/B Testing
Find out how to optimize your website to give your customers experiences that will have the biggest ROI for your business. Register Now »
As we click, scroll, swipe and tap our way across the Web, we leave a trail of data.
Entrepreneurs are rushing to figure out how to profit from this. Indeed, the promise of data, Big Data in particular, is that businesses of all sizes can now make smarter decisions, create more personalized experiences for each customer and grow faster than ever. The data-driven business is the proverbial city on the hill, an amazing place to reside and the aspiration of those watching from afar.
Realizing the promise of data, however, is harder than it looks. For one, making sense of this data requires specialized training and talent that is in short supply. Second, the entrenched processes that drive many businesses get in the way of new data-inspired approaches.
Finally, data is sometimes ignored when it challenges the intuition of executives and experts. All the data in the universe is useless without a practical, repeatable method for extracting insight from it and a culture that trusts and supports these insights.
Fortunately, A/B testing is a readily available approach for using data to immediately improve business performance while building an analytical culture at the same time.
A/B testing allows a business to make sense of customer data in a practical, measurable way. Far more than a comparison of button colors or badge shapes, properly executed web testing develops a better understanding of user psychology and reveals the unique needs of visitors, leading to the growth and profit promised by the data revolution.
Here are four ways to apply A/B testing to your organization.
1. Make data-inspired observations. The process begins with an analysis of existing data. Maybe returning visitors bounce from the homepage more than new visitors or a large percentage of users exit midway through a checkout process. Or maybe your business has a goal to increase average order value or free trial registrations built around an underperforming part of a website.These observations should inspire questions and help focus testing efforts.
2. Formulate a hypothesis about your customer. Once an area of focus has been identified, it’s possible to make a prediction about how it could be improved.
Looking at average order value, for example, may reveal that it hovers around the threshold for earning free shipping. Increasing the free shipping threshold, a reasonable hypothesis might suggest, will increase the average order value. Such a hypothesis is interesting because it is inspired by previous research, impacts a key business goal, and offers insight into the purchase behavior of customers.
3. Test a prediction. A testing platform such as Optimizely or Adobe Target allows for experimentation that tests a hypothesis by temporarily publishing a redesigned element on your website to compare with the original element. To do this, a portion of the website traffic is split randomly between the various test conditions. Behavior within each group is recorded and measured.
4. Tell the story. After running the test, it is critical that results are carefully analyzed. A test might reveal, for example, that increasing the free shipping threshold did indeed increase average order value but deeper analysis might reveal that it decreased overall orders. Changing the free shipping threshold, this suggests, was persuasive but perhaps not the best way to increase average order values and ultimately total revenue.
Even this simple example illustrates how testing is a practical tool that allows businesses to profit from the data they can easily collect. Each new test builds on the insight gained from previous efforts and contributes to a deeper understanding of the psychology of their customers.
The better a website addresses the mentality and needs of its users, the better it will perform. For many businesses, that means an increase in revenue, an undeniable bonus for a process that also creates a data-driven culture focused on experimentation, innovation and sustainable growth.