The following post discussing Data Driven UX is based off of a slide deck our VP of Product, Amit Levi, presented at an Israeli UX forum earlier this month. The deck can be viewed below.
Whether it is your Facebook wall, or Uber’s smartphone app, a lot of thought goes into making it effective. Developers try to make sure that the page functions well, while designers work on ensuring that it looks good so that everything comes together to improve your user experience. User experience design, often called UX, drives product development decisions.
When approaching UX, one needs to set out goals and not just focus on good looks. For example, a product page on an eCommerce site may not present a good enough user experience, despite having a clean layout, great product descriptions and easy navigation. The goal of eCommerce pages is to get users to buy products, so UX developers and designers will benefit from understanding why a user may, or may not, proceed with checkout.
Among our different customers we see several that use the CoolaData power analysis tools to take data driven product calls. Here is an actual use case of an eCommerce application working with CoolaData in which purchased conversion for new users was increased by 27%. This increase was achieved using our three step process, (defining KPI -> segmentation -> behavioral analysis).
Step 1: Defining a KPI
Practice: While on-boarding a customer to CoolaData’s analytics platform they identified that purchase form completion (ie. the number of users who complete a purchase) is a crucial Key Performance Indicator (KPI). Based on market standards and current behavior, one of our customers decided it wanted to achieve a 75% conversion rate – meaning 75% of the users who start a registration process will complete it.
Results: By analyzing the number of purchases completed, compared to the number of purchased page views using our Power KPI widget, the customer discovered that its current conversion rate was 67%.
Step 2: Segmentation
Practice: With the goal of increasing the conversion rate from 67% to 75%, our customer used CoolaData’s Power Cohort widget to see how different user segments performed in regards to this specific KPI. In doing so, they were hoping to better understand which user segment needed more attention.
Results: This customer realized that while there was an overall conversion rate of 67%, there was a significant difference between how different users groups perform with 41% of new users completing their purchases, versus 83% of returning users.
Step 3: Behavioral Analysis
Practice: Finally, the customer had to figure out what was different about new users and what was causing a higher churn in their segment. By utilizing our Path widget our customer could follow the specific path taken by new users navigating through the purchase flow and see where new users churn after starting the purchase form.
Results: Path analysis revealed that most new users were churning when asked to create a new account, because they did not want to create a dedicated login. In order to reduce this churn, our customer added a social login option that allowed users to login with a Google or Facebook account rather than open a new one. By making this small addition our customer was able to decrease churn, and increase purchase conversion for new users by 27%.
This is just one example of how data can improve UX and in-turn reduce churn, illustrating how data reveals problems that were not evident to begin with. With any business decision, it is important to be properly informed, and without data it is basically impossible to make an informed decision. This lack of info makes us feel like the stubborn hippo (Highest Paid Person’s Opinion), that thinks it is always right even when it lacks the data to make the right decision. The power of the data driven methodology is that it takes the power from the “hippo,” usually the CEO, and gives the power and facts to those who are on the ground making decisions as discussed extensively in Andrew McAfee’s blog. In order to get past this hippo-dominated form of decision making and move along the path of success, one should use as much data as possible! According to McAfee, the more data one has at one’s disposal, the better informed a decision becomes (as illustrated in his graph below).
While the benefits of being data driven in UX design may now seem clear, the obvious next question is how can one become data driven? There are four main steps to becoming data driven: first it is important to set goals and KPIs. This step goes back to our initial example, where the goal of the page was determined as getting users to complete a purchase. Its KPI would likely be a certain percentage of users complete the purchase.
Once goals are set, tools for data collection and analysis must be chosen, followed by the actual collection and analysis of the data. The last step is optimization, based on insights from analyzed data. By following these four steps, our customer was able to optimize and achieve better conversion.