Index Analysis
business analytics
slug: topic-map-business-analytics-index-analysis
Vocabulary:
- index: A ratio showing how a segment performs relative to a baseline, typically expressed as a value where 100 = baseline
- baseline: The reference point for comparison, often the grand total or a specific segment
- over-index: When a segment's index > 100, indicating above-average performance
- under-index: When a segment's index < 100, indicating below-average performance
- rate: A normalized measure like mean, percentage, or ratio (not absolute counts)
- penetration: The percentage of a population that exhibits a characteristic
- affinity: How strongly a segment gravitates toward a behavior compared to the total population
- lift: Similar to index, often used in marketing contexts
- composition: What percentage of the total does this segment represent
- incidence: The rate at which something occurs within a segment
Concepts:
- normalization: Index removes the effect of segment size to enable apples-to-apples comparison
- relative_performance: Index focuses on "how efficiently" not "how much"
- level_specific_indexing: Each level in a cube can have its own baseline for indexing
- cascading_indexes: Index at level 2 can be against level 1 parent or against level 0 grand total
- interpretability: Index of 150 means "50% higher than baseline" - intuitive for non-technical audiences
- actionability: Over-indexed segments suggest opportunity for targeting or expansion
- standardization: Index allows comparison across different measures with different scales
Procedures:
- identify_baseline: Determine what to index against (level 0, parent level, specific segment, time period)
- validate_rate_metric: Ensure the measure.metric is a rate (mean, percentage) not a sum or count
- calculate_index: (segment_rate / baseline_rate) × 100
- flag_significant_deviations: Identify indexes that are meaningfully different from 100
- rank_by_index: Sort segments by index to find strongest/weakest performers
- create_index_bands: Group segments into categories like "strong over-index (>150)", "moderate (100-150)", etc.
- compare_indexes_across_dimensions: Which dimension shows the most variation in indexing
- time_series_indexing: Track how index values change over time for each segment
Procedures_advanced:
- weighted_index: When baseline should consider segment sizes differently
- multi_baseline_indexing: Index against multiple baselines simultaneously (vs grand total AND vs parent)
- statistical_significance: Determine if index deviation is meaningful or just noise
- index_decomposition: Break down index into contributing factors
Topics:
- segment_performance_scoring
- market_basket_affinity
- demographic_targeting
- channel_effectiveness
- product_penetration_analysis
- geographic_hotspots
- time_of_day_patterns
- customer_segment_behavior
- conversion_rate_benchmarking
- media_efficiency_analysis
Categories:
- comparative_analytics
- normalization_techniques
- performance_measurement
- opportunity_identification
- targeting_optimization
Themes:
- fair_comparison: Removing size bias to see true performance
- signal_detection: Finding segments that punch above or below their weight
- resource_allocation: Where to invest based on efficiency not just volume
- pattern_recognition: Which combinations naturally go together
Trends:
- real_time_indexing: Live dashboards showing current index vs historical baseline
- predictive_indexing: Using historical index patterns to forecast future performance
- multi_dimensional_indexing: Indexing across 3+ dimensions simultaneously
- dynamic_baselines: Baselines that shift based on context or time period
Use_cases:
- retail: "Premium products over-index in urban stores (index=175) vs suburban (index=85)"
- marketing: "Email campaigns over-index for conversion (index=220) vs social media (index=95)"
- ecommerce: "Mobile users under-index for cart completion (index=67) - optimization opportunity"
- healthcare: "Condition X over-indexes in demographic Y (index=310) - screening campaign target"
- finance: "High-value transactions over-index on weekdays (index=156) vs weekends (index=44)"
- media: "Content category A over-indexes with audience segment B - programming decision"
- hospitality: "Business travelers over-index for minibar purchases (index=245)"
- saas: "Enterprise customers over-index for feature X usage (index=190) - upsell opportunity"