ShopPulse Analytics OS

ShopPulse is a client-grade product I'm building in the open as part of my portfolio: an analytics OS for small e-commerce teams who currently run their business out of storefront admin screens and spreadsheet exports. This page documents the full build specification — the problem it targets, the system design, and the v1 scope — and will grow into a complete case study as the build ships.

Small store owners rarely lack data — it's just scattered. Orders live in the storefront admin, payouts in the payment dashboard, ad numbers in another tab, and stock counts in a spreadsheet that gets updated when someone remembers. Mornings start with an export-and-paste ritual, stockouts get noticed after the sale is already lost, and nobody can answer 'how did last week actually go?' without an hour of assembly.

One dashboard over a clean data model: orders, products, and channels imported into PostgreSQL, with KPIs computed server-side and presented in views a non-analyst can read in thirty seconds. Per-product thresholds drive low-stock alerts instead of manual checks, date-range comparisons replace ad-hoc spreadsheet math, and every table exports back to CSV so accountants and agencies keep their existing workflow.

• Mornings start with numbers, not exports — the daily status ritual drops from an hour to a glance.
• Stockouts become an alert you act on, not a surprise you apologize for.
• One shared source of truth ends the 'which spreadsheet is current?' problem.
• Clean CSV exports keep accountants and agencies happy without handing out store-admin access.

• A complete v1 specification: 14 screens, the PostgreSQL data model, and the alert-rules design.
• KPI definitions written down precisely — what counts as revenue, when an order counts, how refunds land.
• An import mapping designed for messy real-world CSVs: inconsistent SKUs, currencies, and date formats.
• UI direction consistent with my FieldOps build: dense tables, calm charts, no dashboard theater.

• Alert design is a product problem, not a query — naive thresholds train people to ignore alerts.
• Every store defines 'revenue' slightly differently; pinning metric definitions before building charts prevents expensive rework.
• CSV import is the real onboarding feature — if seeding data is painful, the dashboard never gets adopted.

• v1 — core dashboard: KPI overview, orders, products, CSV import/export.
• v1.1 — low-stock alerts with per-product thresholds and a restock queue.
• v2 — channel and campaign attribution, plus multi-store support.
• Later — scheduled email digests and a Shopify importer.