Key Takeaways
- Aggregates (totals or averages) can hide important differences between groups that may significantly impact your analysis.
- Segmentation reveals which groups are performing differently than others, allowing for targeted interventions and tailored strategies.
- Use grouped charts, small multiples, and faceted visualizations to show both the whole picture and its component parts.
- Always check if performance varies across important dimensions (teams, demographics, time periods) before drawing conclusions from aggregated data.
Real-world Example
Training Completion Rates
A company reports an overall 85% completion rate for their new compliance training. However, when segmented by department:
- Marketing: 96%
- Sales: 93%
- Support: 89%
- Operations: 55%
The segmented view reveals a significant issue in the Operations department that was hidden in the overall average.
How to Apply This Principle
1. Identify Key Segments
Before analyzing data, identify meaningful segments that could influence your metrics:
- Teams or departments
- Geographic regions
- User demographics
- Product categories
- Time periods
2. Choose Appropriate Visuals
Select visualization types that effectively show both overall and segmented data:
- Grouped bar charts
- Small multiples
- Faceted dashboards
- Stacked area charts
- Interactive filters
3. Test Your Insights
Ask these questions to verify your segmentation is meaningful:
- Does performance vary significantly across segments?
- Are any segments underperforming?
- Would different actions be needed for different segments?
- Do the segments reveal patterns over time?
"The average never tells the whole story. Only by looking at the distribution and segmentation of your data can you truly understand what's happening and why."— Edward Tufte, Data Visualization Expert
Previous Principle
Next Principle: Smooth the Noise