ERPNext Africa · operating playbook

Made-to-order & curated Shopify brands on ERPNext

How a design-led brand that sells on Shopify — some products held in stock, some built bespoke to order, much of it sourced from specialist makers — should set ERPNext up so finance, operations and the warehouse all read from one truth.

The shape of the problemThese brands usually run two flows at once: a stocked catalogue made ahead of demand and sold off the shelf, and bespoke work built to order against a deposit — much of it produced by outside makers, not in-house. Cash and stock behave differently across the two. ERPNext keeps them straight only if it is set up around how the product is really made and sold, not around accounting alone. Companion to Operating with Shopify and the Shopify → ERPNext setup.

Two demand models, one business

A stocked line and a bespoke line behave differently at every step. ERPNext has to tell them apart — flag each item's model, and let each order line route down the right path (a single order can contain both).

Dimension	Stocked · make-to-stock	Bespoke · make-to-order
Sold on Shopify as	In stock — ships from the shelf	Made to order — lead time / preorder
Triggers production	Reorder level / min-max / forecast	The customer's Sales Order
Payment	Full at checkout	Deposit up front, balance on delivery
Revenue recognised	At sale / dispatch	On delivery (balance invoice)
Where value sits	Finished-goods inventory (asset)	Work-in-progress until delivered
Shopify inventory	Two-way sync — live available qty pushed back	Not stock-tracked
Key ERPNext levers	`Item` (stock) · `Reorder Level` · `Production Plan`	`Sales Order` · `BOM` · `Work Order`

The website's “in stock” number is an ERPNext numberFor stocked lines, the quantity Shopify shows as available is your ERPNext finished-goods stock. If the two drift you either oversell (and disappoint) or hide sellable stock — so the stocked side lives or dies on inventory accuracy and a reliable two-way sync.

Sourcing from multiple specialist makers

Many of these brands don't make everything themselves — they curate. One catalogue might be filled by a soft-furnishings maker, a lighting maker and a furniture maker, plus a long tail of smaller suppliers. This is a second axis on top of stock-vs-bespoke: who makes it — you, or a supplier.

Worked example — three makers, one catalogue

Maker (Supplier)	Supplier Group	Category (Item Group)	Typical buy mode
Soft-furnishings workshop	Soft furnishings	Shades & Bedding	Stock the fast lines · make-to-order the custom sizes
Lighting maker	Lighting	Lamps & Lights	Stock the bestsellers · dropship or to-order the rest
Furniture maker	Furniture	Tables & Case goods	Mostly made-to-order (lead time) · a few display lines stocked

The setup that makes it work

Suppliers grouped by discipline — one Supplier per maker, in Supplier Groups (Lighting, Soft furnishings, Furniture), each with its lead time and payment terms.
Item Groups by category · default supplier per item — so reporting rolls up by product line and a purchase suggests the right maker automatically.
Buy for stock — Purchase Order → Purchase Receipt into the finished-goods warehouse → available on Shopify; replenish on reorder level.
Buy to order — the customer's Sales Order raises a Purchase Order (or Subcontracting Order) to the maker; receive it, then deliver to the customer.
Dropship — for bulky or slow lines, flag the item so the maker ships direct to the customer and no stock touches your warehouse.
Landed cost — a Landed Cost Voucher adds freight, import and handling to item cost, so margin isn't overstated.

Adding “other suppliers” later is just repetitionA new maker is a new Supplier in the right group; their products are Items in the right category with them as default supplier. The catalogue, buying modes and reporting all extend with no rework — which is the point of setting the structure up deliberately now. The procurement spine stays the same: Purchase Order → Purchase Receipt → Purchase Invoice, and you track each supplier's on-time delivery like your own floor.

Finance — set the ledger up to answer questions, not just file returns

The chart of accounts you design in week one decides which questions you can answer in year one.

Split revenue — and matching cost of sales — by product line. The single biggest lever. If every sale lands in one Product Sales account you can never see margin per line; give each line its own income and COGS account, or use a Cost Center / Accounting Dimension. Decide the lines up front.
Book deposits as a liability; recognise revenue on delivery. Post customer deposits to a Customer Deposits Received liability, not to income. Recognise revenue and its COGS when the product is delivered and the balance is invoiced, so they land in the same month.
Carry finished-goods stock as an asset; recognise at sale. Stocked catalogue items are inventory you own until sold — on the balance sheet at cost, with revenue and COGS landing together at dispatch, no deposit, no deferral.
Treat the undelivered order book as the forward view. Open Sales Orders not yet delivered are future revenue and WIP — report the backlog and how much is already cash-secured by deposits.
Reconcile to the accounts off the GL, not raw invoices. Invoice documents include proformas, credit notes and intercompany lines and read far above recognised revenue; the GL Entry income accounts are the truth.
Load budgets so a forward-looking P&L compares against a real plan.

PitfallOne revenue account plus deposits booked as income — a P&L that swings on cash timing and can't show margin by line. Fix it before the first month closes.

Operations — the order is a promise; production keeps it

Most failures here aren't software — they're a step someone stopped doing in ERPNext. (This covers what you make in-house; sourced products go via the makers above.)

One Sales Order per customer order — the contract. It carries items, price, promised delivery date and deposit; everything downstream hangs off it.
Open and close every Work Order. The discipline that matters most: mark work started on the floor and completed when it comes off. Un-closed work orders make WIP inflate forever and on-time-delivery meaningless.
A costed BOM per product drives material requirements, planned cost and — with actuals — real margin per job.
Build to the shelf as well as to the order. Stocked lines are produced ahead of demand via Production Plan and reorder points — run those replenishment work orders on the same start/close discipline.
Measure on-time delivery against the promised date, per order line — the best pulse of whether ops keeps the storefront's promises.

Warehouse & stock — keep the system equal to the shelf

Every dashboard upstream is only as good as whether system stock matches what's physically there.

Discipline the item master and category every item. Consistent codes; each item tagged finished good / raw material / consumable, and within finished goods, its product line. No category means no product-line reporting anywhere.
Model your real spaces as warehouses / bins — raw store, WIP, finished goods, dispatch — so counts and WIP valuation match reality.
Post movements as they happen, not in weekly batches — issue to production, receive finished goods, dispatch on delivery, in ERPNext, at the moment it happens.
Finished-goods stock is your shop window. For stocked lines the shelf quantity is the quantity the website offers — count often and keep the two-way Shopify sync honest.
Reorder levels on raw materials raise a Material Request before a build stalls on a missing part.
The Delivery Note is the handoff to finance — it moves stock out and triggers cost of sales, keeping dispatch and revenue recognition in step.

Shopify ↔ ERPNext — who owns what

Shopify is the shop window; ERPNext is the system of record. Keep the storefront simple and let ERPNext own production, stock and finance.

Job	Where it lives
Selling, storefront, checkout, marketing	Shopify
Orders & customers → Sales Orders / Invoices	Sync in to ERPNext
Finished-stock levels of stocked lines	Sync out to Shopify (don't oversell)
Production, stock, purchasing, the ledger	ERPNext
SKU ↔ Item code	Mapped 1:1, exactly and consistently

Rule of thumbIf a decision is about selling, it belongs in Shopify. If it's about making, holding or accounting for the product, it belongs in ERPNext. Don't split one responsibility across both — and decide up front how deposits, VAT and non-stocked made-to-order items behave in the sync.

Go-live checklist, by role

Owner	In place before go-live	Why it matters
Finance	Revenue + COGS accounts split by product line	Margin by line becomes possible
Finance	Deposit liability account & recognition-on-delivery rule	P&L reflects earnings, not cash timing
Finance	Reconciliation to management accounts via GL	ERPNext ties to the accountant's number
Operations	Costed BOM per product line	Material needs + real margin per job
Operations	Work-order start/close routine agreed with the floor	Honest WIP & on-time delivery
Warehouse	Item master coded, every item categorised	Product-line reporting end to end
Warehouse	Warehouses/bins model the real spaces	Counts & WIP match reality
Supply	Suppliers grouped; default supplier + lead time per item	Purchasing routes to the right maker
All	Every item flagged stocked vs made-to-order	Right fulfilment path; no overselling
All	Shopify SKU ↔ Item code mapping verified	The integration stays reliable

The most common ways it goes wrong

One revenue account. Everything in Product Sales → no margin by product line, ever.
Deposits booked as revenue. The P&L swings on when cash lands, not when product is earned.
Work orders never closed. WIP inflates indefinitely; on-time-delivery becomes noise.
Item codes drift from Shopify SKUs. The integration mis-maps and stock/orders silently diverge.
Stock posted in weekly batches. System stock stops matching the shelf; every downstream report is suspect.
Reconciling to accounts off raw invoices. Invoice documents read far above recognised revenue — reconcile on the GL.
Two entities summed together. Intercompany value counted twice; group revenue looks several times reality.
Items with no category. No product-line dimension exists, so no product-line reporting can.
Stocked and bespoke not distinguished. The same item sold both ways — oversell shelf stock, or carry phantom made-to-order “inventory”.
No default supplier or landed cost. Purchases route to the wrong maker and freight/duty never hits item cost — COGS understated, margin flattered.