DMAIC Framework¶

DMAIC (Define, Measure, Analyze, Improve, Control) is a data-driven quality improvement methodology used in Six Sigma.

Overview¶

DMAIC provides a structured approach to process improvement:

Phase	Description	Key Activities
Define	Identify the problem	Project charter, scope, goals
Measure	Quantify the problem	Data collection, baseline metrics
Analyze	Find root causes	Statistical analysis, root cause identification
Improve	Implement solutions	Solution design, pilot testing
Control	Sustain improvements	Control charts, monitoring, documentation

Document Structure¶

{
  "$schema": "https://github.com/grokify/structured-goals/schema/dmaic.schema.json",
  "metadata": {
    "id": "DMAIC-2025-001",
    "name": "Manufacturing Quality Metrics",
    "owner": "Quality Manager",
    "team": "Quality Engineering",
    "period": "FY2025-Q1",
    "reviewCadence": "Weekly"
  },
  "categories": [
    {
      "id": "production",
      "name": "Production Quality",
      "metrics": [
        {
          "id": "prod-001",
          "name": "First Pass Yield",
          "baseline": 85.0,
          "current": 94.2,
          "target": 98.0,
          "unit": "%",
          "phase": "Control",
          "trendDirection": "higher_better",
          "controlLimits": {
            "ucl": 99.0,
            "lcl": 90.0,
            "centerLine": 95.0,
            "sigma": 1.5
          }
        }
      ]
    }
  ]
}

Go Usage¶

import "github.com/grokify/structured-goals/dmaic"

// Create a new DMAIC document
doc := dmaic.New("DMAIC-2025-001", "Quality Metrics", "Quality Manager")

// Add a category with metrics
doc.Categories = append(doc.Categories, dmaic.Category{
    ID:   "production",
    Name: "Production Quality",
    Metrics: []dmaic.Metric{
        {
            ID:             "prod-001",
            Name:           "First Pass Yield",
            Baseline:       85.0,
            Current:        94.2,
            Target:         98.0,
            Unit:           "%",
            Phase:          dmaic.PhaseControl,
            TrendDirection: dmaic.TrendHigherBetter,
            ControlLimits: &dmaic.ControlLimits{
                UCL:        99.0,
                LCL:        90.0,
                CenterLine: 95.0,
                Sigma:      1.5,
            },
        },
    },
})

// Calculate overall health
health := doc.CalculateOverallHealth()
fmt.Printf("Overall Health: %.0f%%\n", health*100)

Trend Directions¶

Trend direction determines how thresholds are interpreted:

Direction	Description	Warning	Critical
`higher_better`	Higher values are better	Below warning threshold	Below critical threshold
`lower_better`	Lower values are better	Above warning threshold	Above critical threshold
`target_value`	Specific target is best	Deviation from target	Large deviation from target

Statistical Process Control¶

For metrics in the Control phase, SPC limits define acceptable variation:

type ControlLimits struct {
    UCL        float64 // Upper Control Limit
    LCL        float64 // Lower Control Limit
    CenterLine float64 // Target/mean value
    Sigma      float64 // Standard deviation
}

A metric is "in control" when:

LCL ≤ Current Value ≤ UCL

Process Capability¶

Six Sigma process capability metrics:

type ProcessCapability struct {
    Cp         float64 // Process Capability
    Cpk        float64 // Process Capability Index
    SigmaLevel float64 // Sigma level (1-6)
    DPMO       int     // Defects Per Million Opportunities
}

Sigma Level	DPMO	Yield
1σ	691,462	30.9%
2σ	308,538	69.1%
3σ	66,807	93.3%
4σ	6,210	99.4%
5σ	233	99.98%
6σ	3.4	99.9997%

Root Cause Analysis¶

Track root causes during the Analyze phase:

type RootCause struct {
    ID          string
    Description string
    Category    string  // People, Process, Technology, etc.
    Impact      string  // High, Medium, Low
    Validated   bool
}

Initiatives¶

Link improvement initiatives to metrics:

type Initiative struct {
    ID             string
    Name           string
    Description    string
    Owner          string
    Status         string
    StartDate      string
    EndDate        string
    MetricIDs      []string // Linked metrics
    ExpectedImpact string
    ActualImpact   string
}

Validation Options¶

Three validation presets are available:

Preset	Use Case
`DefaultValidationOptions()`	Permissive, good for drafts
`StrictValidationOptions()`	Production-ready documents
`SixSigmaValidationOptions()`	Full Six Sigma compliance