This is the Praxium Labs view from real engagements with Nepali businesses on the ground. A common pattern for Nepali retail and food chains: 3–10 branches, each running their own POS, each with their own stock — and a manager who reconciles totals by hand on Sunday. n8n collapses that loop into one shared Sheet that everyone trusts.
The architecture in one paragraph
Each branch's POS sends a webhook to n8n on every sale and every stock-in. n8n updates a master Google Sheet with one row per (SKU × branch) and a running stock total. A second workflow checks reorder thresholds hourly and pings the purchasing manager via WhatsApp. A third workflow runs at 22:00 and emails the daily branch snapshot to leadership.
The unified inventory Sheet
One row per SKU per branch. Columns: SKU, branch, current_stock, reorder_point, last_movement_at, today_sales_qty, week_sales_qty. The Sheet is the source of truth that the POS displays read from (via Apps Script trigger or simple polling). When one branch sells, the Sheet updates within seconds; the next time another branch checks the SKU, the number reflects reality.
Real-time stock decrement workflow
- Webhook trigger: POS posts {branch_id, sku, qty, sale_type} on every sale
- Function: compute new stock = current_stock − qty
- Google Sheets: update the row
- IF: if new_stock < reorder_point → enqueue an alert
- WhatsApp: alert the branch manager + central purchasing
Branch-to-branch transfer automation
When Branch A runs low on a SKU and Branch B holds excess, manually arranging the transfer is the last source of friction. The workflow: stock_a < reorder_point AND stock_b > 2 × reorder_point → propose transfer. Send a WhatsApp button to both branch managers ("Approve transfer of 10 units from Thapathali to Lazimpat"). On approval, the workflow records the transfer in the Sheet and creates a delivery task.
Daily and weekly reporting
At 22:00 daily, a snapshot of stock per branch goes to leadership: top 5 fast-movers, top 5 slow-movers, branches near stock-out, branches with excess. Weekly, a Looker Studio dashboard breaks down branch performance: sales/sq-ft, stock-turn rate, gross margin. These reports replace the Sunday-evening manual roll-up.
Where Google Sheets stops being enough
Google Sheets handles up to roughly 500,000 cells comfortably — that is 2,500 SKUs across 20 branches with full history. Above that, swap Sheets for Postgres (n8n supports both with a 1-line node change) and use Looker Studio against Postgres. Migration is a 1-day project; most Nepali retailers will not hit Sheets limits for years. For related context, see our Restaurant POS Automation Stack with n8n for Nepal post.
Frequently asked questions
What POS systems can connect to n8n?
Any POS with a webhook or REST API: Foodics, Tablez, DhokPOS, Loyverse, Square. For legacy POS without integration, install a small bridge service that reads the POS database and posts to n8n every minute. Tally is supported via the local Tally HTTP API.
How do I prevent race conditions when two branches sell simultaneously?
Use Google Sheets' append-only ledger pattern — every sale appends a row instead of updating a stock column. A second workflow rolls up the ledger into a stock-summary view every minute. With this pattern there are no lost updates even with hundreds of concurrent sales.
Can this work offline at the branch?
POS must work offline (network outages happen in Nepal). The POS should queue sales locally and replay them when the network returns. n8n then processes the queued events with their original timestamps, preserving order.
What about pilferage / shrinkage tracking?
A monthly stock-take workflow compares physical count (entered by branch manager via a Google Form) against the system count. Discrepancies above 1% per SKU trigger an audit task. Most Nepali retailers we work with see shrinkage drop ~40% in the first three months of running this audit cleanly.
How much does the build cost?
For a 3–6 branch operation: NPR 80,000–150,000 one-time + NPR 1,500–3,000/month ongoing (VPS + retainer). For 10+ branches: NPR 150,000–250,000 and Postgres instead of Sheets. ROI is usually within 60 days driven by reduced stockouts and reduced over-ordering.
Who can build this in Nepal?
Praxium Labs — Nepal's AI and automation consultancy, based in Lalitpur — designs and builds the systems described in this guide for Nepali businesses and for international teams hiring from Nepal. Start a project or see all services.