The Problem-Solving Process: A Modern, Data-Driven Approach for Engineering leaders

Problem

In an engineering organization, when it comes to people and process efficiency, it is subjectivity-driven and not data-driven, which is a significant problem.

• We cannot tell if all our retrospectives and process changes are making our teams better or worse.
• Are we going faster or slower? Who is overloaded, and who has capacity?
• Why do we miss deadlines when enough time has been spent on sprint planning?
• Why can't we predict issues early when we check in with the team every day on the scrum calls?
• When you add a new engineer to the team, is your throughput going up or down?
  

All these happen due to the lack of data-driven nature. As a CTO, I had to make decisions for different parts of the engineering teams, relying on subjectivity, and as we all know, that's not always done right.

Actions taken

To tackle this situation, I reached out to other engineering organizations and leaders to understand how they were dealing with this situation and understood this had been a broader problem than I assumed. But, to become data-driven, the number one challenge was to identify the right metrics that would bring in the desired behaviors.

As part of my research, I came across the DORA metrics, which Google launched. DORA did an extensive analysis with over 33K companies globally and came up with metrics that would help bring data culture to any company. I quickly jumped in to hack some dashboards together and realized that it's a lot of work to get this data from disparate tools like Git and Jira and show them in real-time dashboards. Although I could see the value, I couldn't build it internally as it distracted our core business. That's when I went ahead and launched my startup, Insightly Analytics, to solve precisely this challenge.

Now, we use Insightly to make all team decisions using data, which is the starting point of our conversations in any scrum meeting. We were surprised with many myth busters about our efficiency and how we were investing the efforts in the wrong places due to lack of data.

Today, instead of asking round-robin about the status updates during our standup call, we look at the dashboard, understand the work completed, and finish our calls in half time. With every release, our retrospective meetings clearly show if our previous decisions are helping us get better or worse. Most importantly, we can look backward, identify patterns and celebrate successes on our improvements.

Lessons learned

• Data-driven processes will bring a universal language based on which teams can be aligned and easily communicated across.
• Having real-time dashboards with data can reduce a lot of meeting times.
• When data is available at managers' fingertips, there are fewer slack interruptions asking for status updates and reports.