Retention Science lets marketers create and manage personalized email campaigns at scale with artificial intelligence. The Cortex platform delivers predictive product recommendations tailored for each customer and uses self-optimization to send only the highest performing content.
Earlier-years efforts were focused on creating algorithms that worked. But making sense of customer data can be complicated—and we have the help desk tickets to prove it. Our users struggled to understand what Cortex could do (and the role that they, as users, played). Now, our team was challenged with rethinking the way that users moved through Cortex.
Over a 7-month period, we took Cortex apart, designing a new workflow with a refreshed interface to match. In the process, we strayed away from conventional user experience approaches. We didn't want to focus on pushing users through Cortex as quickly (and thoughtlessly) as possible.
Our goal was to help brand new users become empowered and informed decision-makers. To achieve a frictionless solution, we purposefully slowed down our users.
We wanted to take Cortex further than just functional. In this project (affectionately dubbed "Face/Off" in internal documentation by our COO), we sought to transform moments of frustration into moments of lasting positive impact. Part of this meant taking users through a more logical and straightforward journey; another part involved preemptively offering solutions before users asked the questions.
Our process began with restructuring our information to provide better usability. Early versions of Cortex were difficult to navigate; buttons led to unpredictable pages and the information users needed was never where they currently were. By mapping user stories and focusing on the tasks that users performed (in the order they typically performed them), we were able to make Cortex more organized and clear.
Inspired by Microsoft's Clippy, we introduced in-app help documentation—in the form of links, GIFs, videos, and triggers—to help marketers get started. By making users more independent, we could reduce the number of help desk tickets related to nitty-gritty technical questions or issues already covered in previous articles.
Fast users aren't necessarily better users. In our case, when appropriate, we lengthened the pace of the user experience to make sure our users had a strong grasp of Cortex and all of its moving pieces. As an example, we implemented an email wizard that divided email settings into distinct pages. Initially, users could create an email in a single click: one page controlled both email settings and template design.
But some of Cortex's settings are complex (like ones that involve creating filters from customer segments), so we separated the email's form and function into two individual tasks. The new email wizard splits the creation into smaller, simpler pages with fewer fields. This reduces user error, making sure that emails are correctly set up before users start designing them.
This new approach had a significant benefit over the ol' shove-all-the-settings-under-one-tab method that we previously employed. Dynamic image examples that appeared next to corresponding fields gave users visual cues that explained each setting. The email wizard also paved way for a more advanced HTML editor that includes syntax highlighting.