Green Agile Metrics

Sustainable product development requires more than good intentions — it requires visibility. In Green Agile, we quantify what matters, because meaningful change is only possible when teams understand where they stand and how their decisions shift impact over time. Metrics help teams recognize progress, identify obstacles, and continuously improve ecological, economical und sozial sustainability.

All Green Agile metrics are intentionally classified along two complementary dimensions: ESG areas and impact types.
This dual classification helps teams understand what kind of sustainability a metric relates to and how its effect unfolds in practice.

Each metric is assigned to one or more ESG areas:

  • Environmental: Focuses on energy consumption, emissions, resource usage, and ecological efficiency.
  • Social: Addresses accessibility, inclusion, fairness, and the well‑being of users or employees.
  • Governance: Supports transparency, accountability, decision‑making, compliance, and long‑term organizational sustainability.

In addition, every metric is assigned an impact type that describes how change is created:

  • Direct: Measures effects that occur immediately when the metric improves, such as reduced energy usage or improved accessibility.
  • Indirect: Captures effects that emerge through changed behavior or decisions by users or organizations.
  • Systemic: Reflects long‑term, structural effects across systems, teams, or processes and typically improves through coordinated, continuous action.

Together, ESG classification and impact type provide essential context for interpreting metrics and for making informed, balanced sustainability decisions.

Why Quantification Matters

Improvement requires measurement. Without a baseline, there is no way to understand whether a feature, architectural decision, or workflow actually reduces environmental impact or increases long‑term sustainability. Metrics make sustainability actionable, transparent, and comparable, allowing teams to make informed decisions instead of assumptions.

At the same time, not everything can or should be measured with perfect precision. In complex sociotechnical systems, some effects are hard to quantify directly — for example, changes in user behavior, emerging architectural complexity, or long‑term operational impacts. This is where proxies become essential.

Proxies for Sustainable Impact

A proxy is a simplified indicator that approximates a sustainability effect when direct measurement would be too complex, costly, or impossible. Proxies allow teams to stay pragmatic while still making sustainability measurable.

Common examples include:

  • Bytes per View as a proxy for energy usage in frontend rendering
  • Build minutes per change as a proxy for CI/CD compute consumption
  • Retention days as a proxy for data minimization and storage footprint
  • Accessibility compliance rates as a proxy for inclusive user experience
  • CPU time per transaction as a proxy for backend efficiency

Proxies can be very simple, as long as they are consistent and meaningful. Even a basic trend — such as “number of architectural decisions documented with ADR‑Light” — provides insight and helps guide better decisions over time.

Sustainability in software is never one‑dimensional. Teams must often balance competing requirements across ecological, economic, and social dimensions — and sometimes even between sustainability and other quality attributes such as security or usability.

Trade‑offs are not a sign of failure; they are a natural part of designing real systems. Metrics help teams make these trade‑offs explicit, discuss them openly, and document them transparently.

Examples of Common Trade‑Offs

Longer device support vs. security risks
Extending software support for older devices is ecologically beneficial because it reduces hardware replacement and electronic waste.
However, older devices may no longer receive critical security updates, creating social and security risks.
Teams must weigh the ecological benefit against user safety and regulatory compliance.

Energy savings vs. measurement overhead
Sometimes the effort to measure a particular impact precisely (e.g., exact GPU energy consumption for small workloads) produces more complexity, cost, or engineering effort than the expected savings justify.
In such cases, teams may choose a simplified proxy metric instead of precise measurement — prioritizing pragmatism and value over perfection.

Heavy caching vs. data freshness
Increasing cache hit rates can reduce backend compute and energy consumption, but may negatively affect data accuracy or user experience when stale data becomes visible.

These examples highlight that sustainability decisions must be made with full context, not in isolation. Green Agile encourages teams to use metrics to guide discussions — not to enforce purity, but to support informed, balanced decision‑making.

Metric Categories

This section represents the current, complete baseline of the Green Agile Metrics framework.
At the same time, the framework is designed as an open‑source and collaborative system: everyone is invited to contribute improvements, refinements, and new metrics over time.

While the existing categories and metrics form a coherent and usable foundation today, the framework will continue to evolve as practices mature, new sustainability challenges emerge, and the community contributes additional insights.

The Green Agile framework organizes its metrics into clear categories to support different aspects of sustainable product development:

Explore each category to understand how quantitative insights can guide more sustainable choices throughout the entire product lifecycle.

Product Management →

Table of contents