• 🗂 Ops manager maintaining 7 browser tabs simultaneously — Shopify, Shiprocket, warehouse sheet, returns tracker, WhatsApp group, Zoho CRM, and Excel daily report. One missed tab = missed SLA.
• 📦 No unified fulfilment visibility — orders dispatched from Shiprocket weren't reconciled with Shopify orders until a manual CSV cross-check every evening. Returns that had been shipped back weren't being credited for days.
• 🔔 Return rate spike (from 8% to 14%) went unnoticed for 3 weeks because there was no automated alert system — the ops team only discovered it during a monthly review.
• 🧮 Inventory levels in warehouse were updated by a staff member typing into a shared Google Sheet — any internet disruption or human error created a phantom stock situation that led to orders being accepted for OOS items.
• 📊 CEO review meetings required 2 hours of prep — someone was always manually compiling data from 5 sources into a single PDF presentation the night before.

KLYX's approach was to build a single Next.js dashboard that acted as the single source of truth for all operational data. Rather than replacing existing tools (Shopify stays, Shiprocket stays, Zoho stays), we built a read-optimised data layer that pulled from all of them via APIs and webhooks, normalised into a Supabase Postgres database, and surfaced through a custom React dashboard.

The dashboard has 5 views: Live Orders (real-time), Fulfilment SLA tracker (per courier, per zone), Returns & Refunds (trend + credits outstanding), Inventory (warehouse QoH with reorder alerts), and CEO View (headline KPIs, weekly trend, return rate alert).

1. Mapped every data source and its update frequency — Shopify webhooks (real-time), Shiprocket tracking API (15-min poll), Zoho CRM (hourly sync), warehouse QR scan events (real-time), returns sheet migration to Supabase table.
2. Built the Supabase schema — orders, fulfilments, returns, inventory_snapshots, crm_contacts tables with foreign keys and indexes optimised for dashboard query patterns.
3. Set up Shopify webhook subscriptions for order creation, fulfilment updates, and refund events — all ingested via Next.js API routes, verified with HMAC, upserted idempotently.
4. Integrated Shiprocket API — polling every 15 minutes for tracking status updates, mapping to internal order records, computing SLA compliance (promised delivery date vs. actual scan events).
5. Built returns intelligence module — return initiation, in-transit status, received-at-warehouse, refund-credited all tracked as discrete state transitions with timestamps. Return rate computed per SKU, per week.
6. Built inventory real-time module — warehouse staff scan QR codes on inbound and outbound stock movements, API updates Supabase instantly, dashboard reflects live QoH. Reorder alerts fire on WhatsApp when any SKU drops below threshold.
7. Built the CEO View — 6 headline KPIs (GMV, orders, return rate, fulfilment SLA %, avg delivery days, outstanding refund credits) with 7-day trend arrows. Weekly auto-generated PDF report emailed every Monday at 8 AM.
8. Deployed on Vercel with Supabase Realtime subscriptions — the Live Orders view updates without page refresh as new orders arrive.

• The return rate spike that took 3 weeks to notice manually was caught within 48 hours in the first month of dashboard use — traced to a single SKU with a sizing issue. Fixed before it cost another ₹8L in returns.
• CEO review meetings dropped from 2-hour sessions requiring 2-hour prep to 45-minute sessions with zero prep — the CEO View auto-generates the agenda.
• Outstanding refund credits reduced by 71% in 60 days — the returns credit tracker made the backlog visible for the first time, and the finance team cleared it within weeks.

• Replace tabs, not tools — the instinct to "switch to a better platform" is usually wrong. The data already exists across Shopify, Shiprocket, and Zoho. A unified read layer that surfaces it in one place delivers more value than migrating to a new ERP that costs ₹10L and 6 months.
• Alerts are more valuable than dashboards — the return rate spike that cost the brand 3 weeks to notice would have been caught in 48 hours with a single alert rule. A dashboard you check occasionally is less valuable than an alert that finds you.
• Invest in your CEO View first — the weekly auto-generated PDF that arrives before the MD's morning coffee changed the quality of decisions being made. A single well-designed executive view justifies the entire dashboard build cost.