Insight Synthesis

3. Synthesize Insights

An insight is NOT just a finding. Transform observations into understanding:

Observation: "Users check their work 3-4 times" Insight: "Users lack confidence in the system's accuracy, creating anxiety and inefficiency"

Observation: "Users prefer mobile for field work" Insight: "Context switching between devices disrupts workflow because data doesn't sync reliably"

Good insights answer "why" and have implications for design.

4. Grade Confidence

For each insight:

• High: Multiple sources, consistent pattern, strong evidence
• Medium: Some evidence, but limited sample or mixed signals
• Low: Hypothesis based on thin evidence, needs validation

5. Identify Implications

For each insight, ask:

• What does this mean for our design?
• What opportunities does this create?
• What constraints does this impose?
• What should we prioritize?

6. Document

Create synthesis document in insights/ folder and update currentstate.json:

{
  "id": "i1",
  "title": "Lack of system confidence creates inefficiency",
  "description": "Users check work multiple times due to past system errors, creating anxiety and wasted time",
  "confidence": "high",
  "sources": ["insights/observation_001.md", "insights/interview_003.md"],
  "implications": "Reliability and clear feedback are more important than features"
}

Synthesis Framework

Pattern → Insight → Implication

Pattern: What we observed across multiple sources Insight: Why it happens and what it means Implication: What we should do about it

Example

Pattern: 6 out of 8 field techs use paper notes despite having mobile devices

Insight: Mobile interface requires too many steps and focus for field context where attention is divided and conditions are suboptimal (gloves, sunlight, distractions)

Implication: Design for quick capture with minimal interaction; offline-first; large touch targets; high contrast for outdoor visibility

Tips

• Involve the whole DesignTeam (multiple perspectives)
• Focus on why, not just what
• Look for surprising findings, not just confirmations
• One observation may support multiple insights
• Keep insights user-centered, not feature-centered
• Link every insight to source evidence
• Update insights as you learn more