Sales & Revenue
- How is revenue changing over time?
- Which products and categories generate the most revenue?
- How do discounts, refunds, and returns affect realized revenue?
- Which channels contribute most to growth?
Northstar Data Platform
Business Analytics
Northstar is being designed to produce business-ready analytical models rather than stopping at technically successful data movement. This page defines the questions, metrics, and modeling principles the platform will support as the underlying business data and cloud pipeline are implemented.
Analytical Goal
Turn operational activity into useful business questions
The technical platform exists to support analysis. Northstar's source systems and analytical models are therefore being designed together so the generated data can answer realistic questions across revenue, customers, products, inventory, fulfillment, and marketing.
Current Status
Business model defined, analytical outputs forthcoming
The synthetic source environment is still being expanded, so Northstar does not yet present production-like KPI results or dashboards as completed evidence.
The analytical requirements are being defined now so source entities, relationships, history, and downstream models are built to support meaningful analysis rather than generating arbitrary data and deciding what to measure afterward.
Business Questions
What should the platform help explain?
The business model is intentionally broad enough to create analytical relationships across departments instead of producing isolated metrics from unrelated source tables.
Customers
- How many customers are new versus returning?
- Which customer segments generate the most value?
- How do purchasing patterns change over time?
- What behaviors distinguish higher-value customers?
Products
- Which products and categories perform best?
- How do price, demand, and seasonality interact?
- Which products frequently experience stockouts?
- Where do refunds or returns disproportionately occur?
Inventory
- Where is inventory concentrated?
- Which products turn quickly or slowly?
- How often does demand exceed available stock?
- Which warehouses or products create fulfillment constraints?
Fulfillment
- How long does it take orders to move from purchase to shipment?
- Which orders or warehouses experience delays?
- How do fulfillment times vary by channel or product?
- Where are operational bottlenecks occurring?
Marketing
- Which campaigns or channels contribute to customer acquisition?
- How does marketing activity connect to downstream purchases?
- Which campaigns generate valuable customers rather than only traffic?
- How does campaign performance vary over time?
Target Metrics
Metrics will have explicit definitions
These are representative metrics Northstar is designed to support. Final calculations will be tied to the implemented source model and documented analytical grain rather than inferred from generic retail definitions.
Gross Revenue
Total merchandise value before refunds, returns, and other downstream adjustments.
Net Revenue
Revenue remaining after refunds and other modeled adjustments are applied.
Orders
Completed or qualifying customer orders within the selected analytical period.
Average Order Value
Qualifying revenue divided by the number of qualifying orders.
Units Sold
Total product quantity represented by qualifying order-line activity.
Refund Rate
Refunded value or qualifying refunded activity relative to completed sales.
Repeat Customer Rate
The proportion of customers with more than one qualifying purchase.
Inventory Turnover
A measure of how quickly modeled inventory is sold and replenished over time.
Stockout Rate
The frequency with which customer demand encounters insufficient available inventory.
Fulfillment Time
Elapsed time between defined order and shipment or delivery milestones.
On-Time Shipment Rate
The percentage of qualifying shipments that meet the modeled fulfillment expectation.
Customer Lifetime Value
A derived measure of cumulative modeled customer value across purchasing history.
Analytical Model
Business logic belongs in layers
The target analytical architecture separates raw operational activity from curated entities, reusable analytical models, governed metrics, and final presentation.
Source Entities
Customers, products, orders, order lines, inventory, payments, refunds, shipments, web events, and marketing activity.
Curated Data
Validated and standardized business entities stored in analytical formats after quality processing.
Core Models
Reusable customer, product, order, inventory, fulfillment, and calendar structures.
Business Models
Revenue, customer behavior, product performance, inventory health, fulfillment, and marketing models.
Metrics
Governed calculations with explicit definitions, grain, filters, assumptions, and business meaning.
Portfolio Outputs
Dashboards, trend analysis, segmentation, operational indicators, and examples of analytical investigation.
Modeling Principles
Reliable analytics require more than writing a query
Define the grain first
Every analytical model should make clear what one row represents before metrics or joins are added.
Separate facts from dimensions
Transactional events and descriptive business entities should have clear analytical roles and relationships.
Reuse business logic
Important calculations should be defined once in governed models rather than repeatedly rewritten in individual reports.
Reconcile to source activity
Revenue, counts, quantities, and other important measures should be traceable back to underlying source records.
Preserve time correctly
Historical analysis should account for when business events occurred and how descriptive attributes may change over time.
Make definitions visible
A metric is more useful when its meaning, filters, grain, assumptions, and limitations are documented.
Example Flow
From transaction to business metric
Orders / Order Lines / Payments / Refunds
|
v
Curated Business Data
|
v
Core Models
/ | \
/ | \
Customer Product Order
Model Model Model
\ | /
\ | /
Business Models
|
v
Governed Metrics
/ | \
/ | \
Net Revenue AOV Repeat Rate
\ | /
\ | /
Analytics ViewsAccurate
Reconcile to the source
Important analytical totals should be reconcilable to underlying orders, payments, inventory activity, and other operational records.
Reusable
Centralize business logic
Shared analytical models reduce the risk of multiple reports calculating the same business concept differently.
Explainable
Document what metrics mean
Definitions, filters, grain, assumptions, and lineage should make analytical results understandable rather than treating dashboard numbers as self-explanatory.
What comes next
As the source model, ingestion layer, and analytical models are implemented, this page will replace design-only examples with actual Northstar KPIs, trends, segmentation, dimensional models, reconciliations, and business analysis generated by the running platform.