Foundational research

Zooming in

In order to get a clear picture of the current experience and to test our hypothesis. We conducted twelve user interviews, here are some pain points we've gathered. They can be mainly categorized into three buckets :

User personas

Key user segments -- who are we designing for

Then we created two general user personas that represent the key user segments we've identified: users who use Split as an experimentation tool, and others who use Split as a feature flagging tool.

Journey mapping

Current state VS new potential

After mapping out the user journey for both A and B personas, I realized that we are currently compelling them to use the same interface. Although the goals of these two personas overlap, they are not identical. Therefore, the current flow isn't efficient for either persona.

Task analysis

Efficiency through step by step

Then I took a closer look at the desired outcomes for each persona, examining the underlying goals driving these outcomes. To achieve these goals, users must complete multiple steps (individual tasks). Through task analysis, I identified numerous steps that can be simplified or replaced through automation.

In the image below, the chart on the left represents the current steps users need to go through. In comparison, the 'new-potential' chart contains fewer steps and requires less user input.

Measuring success

What success look like

Knowing where the problems are, we started to draft a three-step milestones plan.

Milestone One

Low-hanging fruits

As our first step to enhance the discoverability of the alerting feature, we added entry points to the creation flow to prevent 'dead-end' scenarios, significantly improving the click rate on the alert policy page.

Outcome

Post-launch feedback and next steps

After working closely with engineering team to bring designs to life, we launched Milestone One. The automation of feature flag alerts set up has significantly increased the click rate and the percentage of unique users who engaged with our alerts tab.

Meanwhile we have clear plans for improving and expanding the alerting features in the future. The next milestones include supporting intelligent alerting threshold setup, shortcuts, and metric dashboard view. We are currently working on implementing these suggestions to further improve the user experience.

Learnings

Looking forward

By listening to our users and making changes that provided real value to them, we created a dynamic alert creation process with improved capabilities. I grew as a designer and walked away with a few key takeaways:

• Identifying and pursuing low-hanging fruit can be fruitful.

By prioritizing these accessible opportunities, the team can demonstrate tangible progress. This will not only results in immediate wins but also sets a positive tone for ongoing design efforts, encouraging a mindset of continuous improvement and innovation.

• Thinking and designing with a systems mindset is a great way to achieve flexibility.

A systemic approach fosters modularity and interoperability, allowing for seamless adjustments and updates. This not only streamlines the design process but also empowers the product to evolve alongside changing requirements.

• When the initial approach encounters challenges, it doesn't signify the end of the world.

Instead, it opens the door to valuable learning and adaptation. In the face of setbacks, the key is to view them as stepping stones toward a more refined and innovative design.