3 keys for engineering improvement: listen, act, and learn

Reimagine engineering success: measure with purpose, invest in improvements, and build a learning culture.

There's no shortage of commentary on engineering metrics. But too often, we see teams obsessing over metrics without clarity about why they're measuring, or whether they're genuinely improving.

At GitHub, we believe there's a better way: shifting from measurement obsession to meaningful improvements through intentional listening, actionable investments, and a learning mindset.

Here are three critical shifts your engineering team can make to drive improvements across your engineering system.

Shift 1: From measuring everything to (sensibly) measuring what matters

Metrics are powerful, but only if they're aligned to your purpose. Too often, teams track everything simply because they can rather than because they should. The result? Confusion, distraction, and lost focus.

At GitHub, we start by clarifying our goals: what business outcomes do we need to achieve, and how does engineering performance directly influence these outcomes? By linking engineering measures explicitly to clear business impacts—such as user satisfaction or revenue growth—we ensure our metrics serve the team's ultimate purpose.

Let’s take the metaphor of a compass vs. stopwatch. A stopwatch tracks every second, but it can't tell you if you're heading in the right direction. On the other hand, a compass keeps you oriented toward your true destination. Our true North is meaningful business outcomes.

Imagine two engineering teams deploying software. Team A meticulously counts deployments per day, fixates on throughput, and pushes for maximum deployment frequency. Team B aligns deployments with clear business impacts and focuses on deployments that measurably improve user experience or reduce customer-reported pain. Rather than sheer deployment volume, Team B tracks customer-reported bugs, user satisfaction scores, or feature adoption to ensure they're progressing toward their business objectives. Team B is more likely to support long-term business outcomes.

It's also critical to balance the cost and benefit of measurement. Not every valuable insight requires complex telemetry or extensive instrumentation. Often, qualitative approaches—such as developer surveys or direct user feedback—provide a cost-effective and powerful complement to quantitative metrics.

DORA insight: The annual benchmark data published in the DORA report is derived from survey responses. The survey is highly reliable because, in addition to its very large sample size, its questions are designed in ways developers can accurately and feasibly answer. If the goal were to measure improvements in deployment time down to the second, telemetry would be more suitable. But for understanding broader capabilities and trends, the DORA survey approach provides a pragmatic and cost-effective solution.

Shift 2: From reporting to funding action

Measurement alone doesn't drive improvement. You also need action. Metrics without funded follow-throughs are like diagnostics without a treatment plan. Too often, teams invest in metrics dashboards but don’t allocate the necessary resources or time to understand or address the data. This means they never come up with targeted interventions to drive improvements and then deliver those initiatives.

At GitHub, we believe improvement requires intentional investment: clearly identifying priorities from metrics and feedback, then explicitly dedicating resources to action. This might mean funding developer experience initiatives, reliability improvements, or tackling technical debt. Action doesn't spontaneously happen. You have to deliberately prioritize and fund it.

Let’s take a situation where an engineering team uses surveys and metrics to identify that developers lose significant productivity due to flaky test suites. Rather than just reporting this issue in dashboards, successful improvement requires engineering time to stabilize and improve testing infrastructure. This funded action directly improves quality, developer morale, and ultimately, business outcomes.

Shift 3: From gaming the metrics to genuine learning

When measurement occurs without corresponding support for meaningful action, teams are incentivized to start gaming the metrics—optimizing for numbers rather than authentic improvement. This leads to superficial fixes that mask deeper issues.

Two common system archetypes relevant to metric gamification are fixes-that-backfire and shifting the burden.

The fixes-that-backfire archetype occurs when short-term solutions intended to quickly improve a metric inadvertently trigger negative side-effects, ultimately making the original issue worse. For example, a team trying to shorten their pull request review times rushing through code reviews or skipping thorough testing. Although the pull request review time metric initially improves, the resulting technical debt and increased bugs eventually slow the team down even more.

The shifting the burden archetype occurs when teams repeatedly rely on temporary workarounds rather than addressing underlying root causes, creating long-term dependencies on superficial fixes. For instance, a team manually restarts flaky servers whenever they fail, instead of investing time into proper infrastructure improvements. Over time, the repeated manual interventions consume more resources, leaving less capacity to address the underlying problem and reinforcing a cycle of firefighting.

At GitHub, we advocate a culture of genuine learning over gaming metrics. We foster psychological safety, encourage experimentation, and celebrate transparency. This includes openly sharing failures and setbacks. Our goal isn't to temporarily move the needle, but to build lasting mastery and sustainable growth. We prioritize sincerity and authentic improvement over short-term metric wins.

Embrace the Listen-Act-Learn mindset

Shifting from measurement obsession to explicitly setting goals and driving meaningful improvements isn't easy, but it is essential. By aligning measurement with clear business purpose, funding action, and prioritizing genuine learning over superficial metric improvements, you'll unlock your team's true potential.

"Listen-Act-Learn" guides us on this journey, reminding us that metrics aren't the goal. They're just one tool in our toolkit for building better systems, stronger teams, and truly impactful engineering.

Want to learn more about the strategic role of AI and other innovations at GitHub? Explore Executive Insights for more thought leadership on the future of technology and business.