Commerce Decisions
Revenue Detective
Answers “why did revenue fall?” with ranked, evidenced causes instead of dashboard archaeology.
The Problem
When revenue drops, analysts spend hours drilling across product, region, payment and returns reports. The question is always the same: what changed, and what did it cost?
How It Works
- Split the date range into two periods and compute recognised revenue (completed orders only).
- Run six independent detectors — product, region, payment failures, returns, stock-outs, cancellations — each with explicit thresholds.
- Estimate the money at stake per finding and rank by ₹ impact, with evidence bullets.
Technically Interesting
Each cause is its own rule with a visible threshold, so every conclusion is traceable. Stock-out loss is modelled from units/day before versus during the outage.
Why It Matters
Turns a reactive reporting task into a prioritised action list — the shape of decision support enterprise commerce teams actually need.
Scope & honesty. Synthetic data (659 orders). Findings overlap — the South-region decline partly contains the stock-out — so impacts are not additive. Loss figures are modelled estimates.