Requestly, a US-based YC company - which helps developers with API testing - used Typo to shorten their PR cycle time and bring teams to process alignment as they scale.

The Challenge

Sachin, CEO at Requestly, noticed that as the team grew quickly, they faced issues like blocked PRs and delays in releasing new features. Even with everyone working hard on new releases, these delays were hard to avoid. To address these challenges, he conducted individual meetings, pinpointing internal process shortcomings. As the team scaled, the regular collection of feedback from the development team posed a growing challenge.

Typo’s unique insight

When Sachin examined the Typo data, it was clear that the review time was excessively high indicating how long PRs remain idle before receiving feedback. Extended periods of unreviewed PRs inevitably postpone subsequent revisions and merges, compounding the overall delay.

Upon further investigation, it was noted that coding time was notably prolonged due to numerous iterations in pull requests (PRs) before submission for review. A DevEx survey conducted by Typo revealed consistent feedback from developers citing unclear requirements from product managers as a common issue.

Actions taken

Sachin reduced PR review time at Typo by implementing an automated review system that fixed non-contextual code issues, allowing reviewers to focus on more critical aspects. He also established SLAs, introduced a “Review Buddy” system, and enhanced PR guidelines, emphasizing smaller PRs. These measures led to a 40% reduction in review time, faster merges, and 30% faster deployments.

Sachin took proactive steps in the company by implementing regular meetings between the product and dev teams to clarify and refine requirements. He strategically broke down complex requirements into clear, actionable user stories, and utilized prototypes and validate these requirements early in the development process. It reduced the coding time by 22% and improved the NPS of the team by 20%.

What next?

Sachin established a workflow where PRs were first scanned using Typo’s automated code review system before being fine-tuned by human reviewers. He prioritized conducting one-on-one sessions based on DevEx survey insights to address teams needing the most support effectively.