Avishag Sahar
Engineering Team Leader


Title: The Undiagnosed Pull Request Problem

Synopsis: Your product team wants a big feature delivered, involving multiple system components, scenarios, new backend technology, and an overall big scope. For many devs, this task can be daunting, as building big features often comes with existing challenges, including, scope creep, planning inaccuracies, code changes, and maybe the most difficult challenge of all: big pull requests (PRs). But why are large PRs such an obstacle? Simply put: Big PRs can be intimidating for reviewers, who often default to “I don’t have enough time,” leading to bottlenecks in the review process. We’ve also seen that big PRs are often the result of sub-optimal coding design, which can lead to bugs, shallow reviews, and a longer lead time to change. Breaking work down into micro tasks allows for better self-management, helps you create higher-quality code, and makes it easier to get quick and thorough reviews.

In this talk, I’m sharing some key insights from LinearB’s analysis of 788k pull requests. We’ll explore:

  • Why breaking work down into micro tasks makes your PRs more appealing
  • How to better communicate the status of your PRs
  • Simple methods to breakdown your tasks and make them more manageable