Case study — Common Hype × Command Center
soldforaloss.comThe operating problem
Three retail locations and a growing Shopify storefront ran on spreadsheets, vendor emails, and nightly exports. Inventory truth lagged a day behind the floor, profit was a month-end guess, and payroll meant rebuilding the same sheet every week. Nothing showed what actually needed attention today.
What I built
I designed and built Command Center — a private operations platform shaped around how Common Hype actually runs, not a template the business had to bend to. Receiving, counts, purchasing, Shopify orders, profit, scheduling, payroll, and commission moved into one system, with QuickBooks, Shippo, and Discord wired in and an AI layer that answers plain-English questions against live data.
What changed
75%
more efficient daily operations
50%
less manual, repetitive work
Nightly → live
inventory & order reconciliation
8
modules replacing scattered tools
Net position
+12.4% wkNet profit
$48,210
Sell-through
87.3%
On hand
14,902
Trailing 12 weeks
Operations feed
- PO-2148 received — 312 units to shelf2m
- Shopify sync — 38 orders reconciled4m
- Cycle-count variance flagged — Store 211m
- Payroll approved — week 231h
Ask your data
“Which store ran the best margin last week?”
Store 2 — 41.8% on $9,420
How it was built
Built with Next.js · TypeScript · Supabase / Postgres · Playwright E2E · Vercel
- 01
Mapped how money and inventory move
Sat inside the operation — floor, back office, spreadsheets — until the real workflow was on paper.
- 02
Modeled the data, not a demo
Stores, vendors, SKUs, shifts, and margins became one schema that matches how the business reasons about itself.
- 03
Shipped module by module
Receiving, counts, purchasing, orders, profit, and workforce went live on Common Hype's own domain against real data.
- 04
Wired the integrations
Shopify, QuickBooks, Shippo, and Discord connected so the system reconciles continuously instead of overnight.
- 05
Put AI on top of the data
Plain-English questions return answers grounded in the live operation — not a generic model's guess.