Building a Culture of Code Quality

In our first post, we looked at how pressure-driven development leads to sloppy code and saw how shifting focus from deadlines to quality leads to happier, more productive teams. In this part, we’ll cover in detail five steps you can take to make that shift, so you have an actionable plan for building a culture of quality in your team.

1. Codify your best practices

Establish a shared and agreed-upon style and standard.

One of the best first steps you can take as a team is to start working on a style guide for the various languages and frameworks you use.

Style guides reduce friction because they take a whole class of issues off the table during code review. They enforce standards by codifying best practices. And best of all, style guides make contribution easier because people will know your project’s contribution requirements up front.

There are lots of great examples online, like GitHub’s Ruby Style Guide or Airbnb’s JavaScript Style Guide. They each represent the guidelines of contributing code to existing projects in those organizations, and guide how code is written when new projects are made.

2. Automate, automate, automate

Let your tools work for you.

When it comes to working on your internal process and company culture, it’s important that any improvements you introduce are perceived as being worth any additional work involved. Style guides are one great example of this, as enforcing a style guide is a much longer project than creating one.

To reduce the amount of work for your engineers, project managers, and QA team, seek to automate whatever tools they are using as part of your overall process. Relieve your team of the burden of running tests locally, and instead run them automatically on a centralized CI server. Instead of relying on everyone to run multiple code quality tools locally, run them simultaneously and automatically with Code Climate. Reducing the friction inherent in running these tools will make a big impact in your organization.

3. Get feedback in the right places

Bring the right data to the right people at the right time.

Once you’ve automated your build process and style guide enforcement, it’s time to make sure that the information gets to the right people. Don’t make people go to the data, bring the data to the people.

Tools that can post where your developers work – like in a GitHub Pull Request – have an advantage of those that don’t. Bringing feedback about code that doesn’t pass the style guide, or doesn’t pass tests inline with code review as it happens makes a huge impact.

When people see data, they can act on it. It’s as simple as that.

4. Measure your progress

Quantify what matters to you.

Building a culture of code quality means creating quantitative goals, and having quantitative goals depends on having stuff that you can measure. Tools like Code Climate can help you create guidelines for progress, and offer ways for your team to see the impact that they are having on the code over time.

Once you have metrics like the GPA of your Code Climate repo or the number of issues found by Code Climate’s static analysis engines, you can begin to act on them. To start with, though, there are just two guidelines you need to keep in mind to observe your code quality improving, your communication getting easier, and your results coming faster:

• Leave code better than you found it
• Don’t introduce new files that aren’t up to quality standards

When you focus on the impact of every proposed change to your code, you’ll improve with every commit.

5. Encourage participation and ownership

Share the load to achieve sustainability.

Once your tools are set up and running smoothly, it’s up to your team to enforce standards and welcome feedback evenly. You’ll find that getting these best practices in place involves making many compromises: older projects will have established standards while new projects may be opportunities to explore new ideas.

When every individual on every team is working along with a set of automated tools that provide actionable, quantified data, you’re on your way to a culture of code quality. Remember that your team wants to execute at the highest possible levels, and everyone participating in an open, encouraging atmosphere will activate the potential of the virtuous cycle.

Getting there takes time and energy, but we think the steps outlined here will send you well on your way. You’ll be able to measure your progress, react when necessary, and stay on top of your goals: to ship awesome stuff, and achieve a virtuous cycle that perpetuates healthy code and happy developers.