Retail analytics
Analytics Hub
Historical data by retail location, auto-generated reports, anomalies, lost revenue, and scenario forecasts.
MVP threshold settings
In the current MVP, settings are read from code and ready to be moved to the database or an admin UI.
Anomaly sensitivity: 2.5
Poor dynamics threshold: -8%
Low markup threshold: 18%
High markup threshold: 34%
Minimum history for forecast: 12
Minimum history for correlation: 6
Rolling baseline window: 3
Assumptions MVP
Critical assumptions from the requirements that are explicitly captured in the product.
• A category's revenue per m² is calculated as the category's contribution to the location's revenue per m²: `category_markup_amount / object_area_sq_m`.
• Segmentation uses the median markup % across categories rather than a weighted average, since the MVP has no revenue or cost-of-goods baseline.
• The seasonal forecast is most useful with 12+ months of history; high confidence is reached at 24+ periods.
• If the Supabase env is not configured, the interface automatically switches to demo mode and uses the built-in seed dataset.
• Auth and roles are ready for Supabase; in demo mode the app is accessible without login for local demonstrations.
• Current Supabase connection status: demo mode.