Tara.ai Issue Tracking

Interviewing customers identified competing pain points

We interviewed and surveyed lead engineers and managers to rank their pain points and metrics. Their most significant pain points were:

• Difficulty tracking blockers
• Lack of alignment on the project
• Challenges with a distributed system of records
• Inaccurate estimates

Prioritizing the right pain point to focus on

During interviews, alignment was a significant pain point for engineering teams, but qualitative data showed more needs for solving problems around blockers, progress, quality, and performance. Despite expecting more interest in alignment tools, it was challenging to make them consistently valuable. We had to test to compare which problem felt the most valuable to users.

Finding the right way to visualize blockers

Our design team created mockups for every problem, ruling some out before showing them to customers. We focused on blockers and explored solutions like a calculator, generated table, Slack integration, in-app notifications, email notifications, timeline reports, labels, and filters. Ultimately, email notifications provided the preferred cadence and visibility.

Notifications helped catch blockers

We launched email notifications for stale PRs, failed checks, and delayed reviews/responses, expecting to increase product engagement and retention. These notifications were low effort, low risk, and highly actionable. Using Mailgun metrics, we identified which customers engaged with our notifications and followed up to learn what additional information they needed.

A solution that drives action

Engineering teams value knowing what actions to take to improve their delivery. Identifying when code was blocked and why was straightforward and kept the team moving.
‍
The feature offered high engagement and efficiently retained customers who regularly checked their emails. Although it did not meet business goals for additional revenue, it served as a model for effective engagement, communication pacing, and actionable user behaviors.