Goal-question-metric (gqm)
Goal-Question-Metric is a framework for selecting meaningful metrics by working backward from objectives. Instead of starting with what's easy to measure, GQM begins with what you're trying to achieve (Goal), identifies what you'd need to know to assess progress (Questions), and only then determines what to measure (Metrics). This prevents the common trap of tracking vanity metrics that don't connect to actual objectives.
The three levels
Goals define what you're trying to accomplish and why. A goal has a purpose (what you want to achieve), an object (what you're measuring), and a viewpoint (whose perspective matters). For example: "Improve the onboarding experience (purpose) for new users (object) from the customer success team's perspective (viewpoint)."
Questions articulate what you'd need to know to determine if you're achieving the goal. For the onboarding goal: "Are users completing onboarding?" "How long does onboarding take?" "Where do users get stuck?" "Do users who complete onboarding remain active?"
Metrics are the specific measurements that answer each question. For "Are users completing onboarding?": completion rate, drop-off rate by step. For "How long does onboarding take?": median time to complete, time distribution across steps.
Why gqm works
The framework addresses several measurement problems:
Prevents vanity metrics. Without GQM, teams often measure what's easy (page views, registered users) rather than what matters (engaged users, revenue-generating customers). By starting with goals, you're forced to connect metrics to outcomes.
Creates measurement coherence. Different teams often track different metrics that don't align. GQM creates shared understanding of what success means and how it's measured.
Reveals metric limitations. Working through the framework exposes when metrics don't actually answer your questions. "Unique visitors" might not answer "Are people finding value?" even though it's often used as if it does.
Supports decision-making. When metrics are explicitly connected to questions and goals, their implications for decisions become clearer. A metric movement becomes meaningful when you can trace it back to a goal.
Applying gqm
Step 1: Define goals clearly. Vague goals produce vague metrics. "Improve the product" isn't actionable. "Reduce time to first value for new users" is specific enough to guide subsequent steps.
Step 2: Generate questions thoroughly. For each goal, brainstorm all the questions you'd need answered to know if you're succeeding. Don't filter prematurely-capture all relevant questions.
Step 3: Identify potential metrics for each question. Multiple metrics might address each question. List possibilities before selecting.
Step 4: Evaluate metric quality. For each potential metric, assess:
Step 5: Select and implement. Choose the metrics that best balance information value against measurement cost. Fewer, better metrics beat many mediocre ones.
Example: improving customer retention
Goal: Reduce customer churn from the product team's perspective to improve revenue sustainability.
Questions:
Metrics:
| Question | Metrics |
|---|---|
| Current churn rate | Monthly churn %, cohort retention curves |
| Segment churn | Churn rate by plan tier, company size, acquisition channel |
| Predictive behaviors | Login frequency, feature adoption, support tickets before churn |
| Timing of decisions | Days since last activity when cancellation occurs, contract renewal rates |
| Stated reasons | Exit survey responses, support conversation themes |
The framework reveals that churn rate alone is insufficient. Understanding why and when churn happens requires multiple complementary metrics.
Gqm pitfalls
Goals that aren't really goals. "Track daily active users" isn't a goal-it's a metric. The goal might be "Increase user engagement" with DAU as one metric among several.
Questions with no viable metrics. Sometimes important questions can't be measured directly. "Do users love our product?" is important but hard to quantify. Proxy metrics (NPS, retention) help but aren't perfect.
Too many metrics. GQM can generate extensive metric lists. Prioritize ruthlessly. Dashboard overload leads to decision paralysis.
Ignoring metric interactions. Metrics can conflict. Optimizing for one (acquisition volume) might hurt another (lead quality). GQM should surface these trade-offs.
Set and forget. Goals change as business evolves. Questions that mattered at launch may not matter at scale. Revisit GQM periodically to ensure metrics remain relevant.
Gqm in practice
Many teams use GQM implicitly when they ask "what are we trying to learn?" before adding dashboard metrics. Making the framework explicit adds rigor and creates artifacts that help maintain alignment as teams grow.
Tools like Klero support GQM thinking by connecting customer feedback to goals. When you can trace user requests and complaints back to strategic objectives, you can identify which metrics would best capture whether you're addressing the feedback that matters most.
