methodatlas
Method guideEngineering

Goal Question Metric

A practical method guide: clarify prerequisites and materials, understand the complete process, and look up outputs, pitfalls, and stop criteria.

ComplexityMedium
Time90-180 min
Participants3-6
FormatWorkshop
MaturityCanonical
Plan Goal Question Metric
01

Prerequisite

What needs to be finished first

Complete firstBounded subject

The scope and purpose of Goal Question Metric are stated as a concrete question.

Without scope, GQM plan becomes arbitrary.
Complete firstAccess to relevant perspectives

Measurement object and purpose, Quality focus and viewpoint, Context and available data and domain decision makers are available.

Without these perspectives, central claims in GQM plan remain unsupported.
02

Preparation

What needs to be ready before start

Materials

Measurement object and purpose, Quality focus and viewpoint, Context and available data, shared workspace, source links, and decision log.

People / roles

Goal owner · domain expert · measurement/analytics · data owner

Pre-read

Collect Measurement object and purpose, Quality focus and viewpoint, Context and available data in advance and mark open assumptions.

Time needed

60–90 minutes

Setup

Prepare Goal → Questions → Metrics with an explicit interpretation model visibly and calibrate assessment terms before starting.

03

Core question

The one question this method answers

Which questions must be answered to assess the defined goal from the chosen viewpoint, and which metrics provide those answers?

04

Flow

Phases with goal, actions, output and pitfall

GQM starts with a measurement goal defined by object, purpose, quality focus, viewpoint, and context. Questions operationalise the goal; metrics come afterwards. A data source without a question is not a measurement programme.

GQM plan

Fictional worked example

Goal → Questions → Metrics with an explicit interpretation model as a precise working structure with a fictional example.

Goal

Object · purpose · quality focus · viewpoint · context

Checkout service · release decision · reliability · Operations · EU platform

Questions

Which answers assess the goal?

Q1 failure frequency? Q2 detection time? Q3 segment outliers?

Metrics

Formula · unit · source · segment · quality

Failure rate; median/P95 detection time; payment type.

Interpretation model

How do values yield a domain decision?

Release below 0.8%, P95 under 5 min, no segment outlier.

The example demonstrates the domain logic of Goal Question Metric; worksheets remain blank.

NASA Software Engineering Handbook: The Goal Question Metric Approach

  1. 1

    GQM goal

    GQM goal

    Goal

    Produce GQM goal so the core question “Which questions must be answered to assess the defined goal from the chosen viewpoint, and which metrics provide those answers?” can be answered with stronger support.

    Inputs for this step

    • Measurement object and purpose
    • Quality focus and viewpoint
    • Context and available data

    Actions

    • State object, purpose, quality focus, viewpoint, and context precisely.
    • Record source, rationale, and open uncertainty directly in GQM goal.

    Output

    GQM goal

    GQM goal · source · uncertainty · next decision

    Quality check

    Are all five goal elements unambiguous?

    Pitfall

    An easily available metric is chosen before the goal and questions are precise. Return to the measurement goal, map every metric to a question, and remove unused data.

    Timing: 10–20 minutes

  2. 2

    Questions

    Questions

    Goal

    Produce question tree so the core question “Which questions must be answered to assess the defined goal from the chosen viewpoint, and which metrics provide those answers?” can be answered with stronger support.

    Inputs for this step

    • GQM goal
    • Quality focus and viewpoint
    • Context and available data

    Actions

    • Derive questions whose answers jointly assess the goal.
    • Record source, rationale, and open uncertainty directly in question tree.

    Output

    question tree

    question tree · source · uncertainty · next decision

    Quality check

    Does every question cover a relevant aspect of the goal?

    Pitfall

    An easily available metric is chosen before the goal and questions are precise. Return to the measurement goal, map every metric to a question, and remove unused data.

    Timing: 10–20 minutes

  3. 3

    Metrics

    Metrics

    Goal

    Produce metric set so the core question “Which questions must be answered to assess the defined goal from the chosen viewpoint, and which metrics provide those answers?” can be answered with stronger support.

    Inputs for this step

    • question tree
    • Quality focus and viewpoint
    • Context and available data

    Actions

    • Define objective and subjective metrics with formula, source, segment, and data quality.
    • Record source, rationale, and open uncertainty directly in metric set.

    Output

    metric set

    metric set · source · uncertainty · next decision

    Quality check

    Can every metric answer at least one question?

    Pitfall

    An easily available metric is chosen before the goal and questions are precise. Return to the measurement goal, map every metric to a question, and remove unused data.

    Timing: 10–20 minutes

  4. 4

    Interpretation

    Interpretation

    Goal

    Produce interpretation model so the core question “Which questions must be answered to assess the defined goal from the chosen viewpoint, and which metrics provide those answers?” can be answered with stronger support.

    Inputs for this step

    • metric set
    • Quality focus and viewpoint
    • Context and available data

    Actions

    • Before viewing data, define how metrics answer questions and trigger decisions.
    • Record source, rationale, and open uncertainty directly in interpretation model.

    Output

    interpretation model

    interpretation model · source · uncertainty · next decision

    Quality check

    Is the combination of metrics into a decision traceable?

    Pitfall

    An easily available metric is chosen before the goal and questions are precise. Return to the measurement goal, map every metric to a question, and remove unused data.

    Timing: 10–20 minutes

05

Artifact

What comes out at the end

Sketch: what the artifact looks like

Goal Question Metric · Goal, Question, MetricDerive concrete metrics from goals through checking questions so no data is collected without purpose
Goal Question MetricThe visual shows the GQM derivation path from goal through checking questions to concrete metrics and data sources.Derive metrics from goals and checking questionsGQM prevents stockpiling data: every metric must answer a concrete question about a goal.Goalimprove platform reliabilityQuestionsHow often do deployments fail?checks stabilityHow fast do we recover?checks recoveryWhere do wait times appear?checks flowMetricsFailure Ratedata source: CI/CDMTTRdata source: incidentsLead Timedata source: workflowgoal first,then questionsno metric withoutchecking questionmeasurement answers the question
Form

Versioned GQM plan with date, sources, owners, and next review.

Versioning / ownership

Retain date, source, assumption state, and decision for every version of GQM plan.

Tool alternatives
  • GQM plan as board or document
  • Source and evidence register
  • Decision and action log

goal-question-metric-working-template.md

Plan Goal Question Metric around its domain steps and the concrete decision question.

Goal Question Metric: Worksheet

Question

To be confirmed

Desired outcome

To be confirmed

Scope

Full workshop Time: 60–90 minutes

Complete this worksheet on paper or in your own document during the work.

Preparation for this scope

Method setup

  • Materials: Measurement object and purpose, Quality focus and viewpoint, Context and available data, shared workspace, source links, and decision log.
  • Roles: Goal owner · domain expert · measurement/analytics · data owner
  • Advance information: Collect Measurement object and purpose, Quality focus and viewpoint, Context and available data in advance and mark open assumptions.
  • Overall time needed: 60–90 minutes
  • Setup: Prepare Goal → Questions → Metrics with an explicit interpretation model visibly and calibrate assessment terms before starting.

Prepare for the work steps

GQM goal: GQM goal

  • Measurement object and purpose
  • Quality focus and viewpoint
  • Context and available data

Questions: Questions

  • GQM goal
  • Quality focus and viewpoint
  • Context and available data

Metrics: Metrics

  • question tree
  • Quality focus and viewpoint
  • Context and available data

Interpretation: Interpretation

  • metric set
  • Quality focus and viewpoint
  • Context and available data

GQM plan

Goal

Object · purpose · quality focus · viewpoint · context

...

Questions

Which answers assess the goal?

...

Metrics

Formula · unit · source · segment · quality

...

Interpretation model

How do values yield a domain decision?

...

Enter only real input, sources, and decisions.

NASA Software Engineering Handbook: The Goal Question Metric Approach

GQM goal: GQM goal

Expected artifact: GQM goal

GQM goal · source · uncertainty · next decision

Entry:

...

  • Are all five goal elements unambiguous?

Questions: Questions

Expected artifact: question tree

question tree · source · uncertainty · next decision

Entry:

...

  • Does every question cover a relevant aspect of the goal?

Metrics: Metrics

Expected artifact: metric set

metric set · source · uncertainty · next decision

Entry:

...

  • Can every metric answer at least one question?

Interpretation: Interpretation

Expected artifact: interpretation model

interpretation model · source · uncertainty · next decision

Entry:

...

  • Is the combination of metrics into a decision traceable?

Open questions and next steps

...

Method guide: https://methodatlas.meierhoff-systems.de/en/methods/goal-question-metric/run-sheet

06

Example output

Concrete filled scenario, fictional example

Fictional example for orientation. Counts and timings describe this case and are not universal requirements.

goal-question-metric-beispiel.md

Concrete filled scenario, fictional example

Fictional goal: analyse the checkout service for the purpose of release decision with respect to operational reliability from the Operations viewpoint in the EU platform context. Q1: how often do real checkout attempts fail? M1 failure rate = failed / started attempts by payment type. Q2: how quickly is failure detected? M2 median and P95 detection time. Interpretation: release only below 0.8% failure and P95 under five minutes with no segment outlier.

07

Pitfalls

Recognize symptoms and steer against them

Quick overview, the full text is in the matching phase in the flow

  1. GQM goalAn easily available metric is chosen before the goal and questions are precise.View in phase
  2. QuestionsAn easily available metric is chosen before the goal and questions are precise.View in phase
  3. MetricsAn easily available metric is chosen before the goal and questions are precise.View in phase
  4. InterpretationAn easily available metric is chosen before the goal and questions are precise.View in phase
08

Stop criteria

When to pause, resolve prerequisites, or choose another approach

GQM plan answers the core question with traceable evidence or explicitly marked uncertainty.
Owners, next actions, and review dates are named for open points.

Where to go next?

Paper illustration of KPI Tree with its method-specific working model.
Product Strategy
KPI Tree

Condenses the derived metrics into a driver tree for ongoing reviews.

A paper-based illustration representing OKRs with its core stages and visible working result.
Product Strategy
OKRs

Adopts suitable metrics as key results for a concrete goal.

View in playbookSystematically Improve Engineering QualityMove from a vague quality goal through visible flow and tested hypotheses to a durably anchored standard path.

Finished the method guide?

Go to the profile for purpose, similar methods, and sources or continue to the next method in the catalog.

Go to profileMore methods