Most Singapore manufacturers start on Excel. It is fast to set up, everyone knows how to use it, and it costs nothing. For a 5-person shop doing 20 orders a month, Excel is not a problem. The problem appears gradually, then all at once — usually during a busy period when the cracks are most visible and least convenient to fix.
The question is not whether to eventually replace Excel — it is recognising when the cost of staying on Excel exceeds the cost of a proper system. This is harder to judge than it sounds because Excel's costs are hidden. They show up as staff overtime, missed quotes, stock discrepancies, and production delays, not as a line item on the P&L.
Warning sign 1: Quoting is a bottleneck
If the owner or a senior staff member must touch every quote before it goes out — because the pricing rules are in someone's head, or because the Excel formula breaks when a new material or process is added — quoting becomes the constraint on growth. Customers who receive quotes in 4 hours buy from competitors who quote in 20 minutes.
A custom pricing engine encodes your actual pricing rules — material cost + markup by category, labour rates by process, minimum margins by job type, surcharges for rush or small batch — so that a sales admin can generate an accurate quote without senior review. Quote volume scales without headcount.
Warning sign 2: Multiple versions of the truth
When sales, production, and finance each maintain their own Excel file for the same order — because the master file cannot be edited by multiple people simultaneously — you inevitably get version conflicts. Production schedules a job based on a quantity that sales subsequently revised. Finance invoices the original price after a discount was verbally agreed.
A shared system with a single record per order eliminates version conflicts. Every department sees the same data. Changes are logged with a timestamp and user. The order moves through statuses — quoted, confirmed, in production, shipped, invoiced — with one record throughout.
Warning sign 3: Stock discrepancies you cannot explain
If your physical stock count regularly differs from your Excel count by more than a rounding error — and you spend hours tracing where the gap came from — your inventory system is not working. Common causes: goods received but not updated in Excel, production consumption not deducted, returned goods not restocked, multiple people editing the same file and overwriting each other's changes.
A proper inventory system records every movement — receipt, consumption, transfer, adjustment, return — with a timestamp and user. You can trace any discrepancy to the transaction that caused it. Cycle counts become a validation exercise rather than a damage assessment.
Warning sign 4: Document generation is manual and error-prone
If generating a delivery order, job sheet, or invoice requires copying data from the Excel order file into a Word template — manually, by a staff member who sometimes makes typos or pastes the wrong row — you have a document generation problem. The risk is not just errors; it is the staff time cost and the delay when the document generator is on leave.
A custom system generates documents from the order record automatically. The delivery order pulls from the confirmed order line items. The invoice pulls from the delivery record. No manual transcription, no typos, no dependency on a specific person knowing which template to use.
Warning sign 5: Month-end reconciliation takes days
If your finance team spends two to four days at month-end reconciling Excel order records against the accounting system — because invoices were raised from a different file than the one production used — the cost is real. Staff time, overtime, delayed financial reporting, and the opportunity cost of finance staff doing reconciliation instead of analysis.
Integration between the operational system and the accounting system (Xero, Million, Globe3) eliminates most of this reconciliation. Invoices raised in the operational system sync to accounting automatically. The finance team validates rather than reconstructs.
What the first system should actually include
The most common mistake in a first ERP project is over-scoping. Manufacturers see the capability of enterprise systems and add requirements until the project is too large to deliver. A focused first system covers the core workflow and nothing else:
- Quoting — pricing engine, quote generation, quote-to-order conversion
- Order management — confirmed orders, line items, delivery schedules, status tracking
- Production job tracking — job creation from order, work-in-progress status, completion
- Inventory and BOM — stock levels, material allocation to jobs, goods receipt
- Document generation — job sheets, delivery orders, invoices, packing lists
- InvoiceNow e-invoicing — PEPPOL BIS 3.0 XML if you sell to government or large contractors
- Basic reporting — outstanding orders, WIP, stock on hand, invoices outstanding
Everything else — supplier portal, advanced costing, regulatory traceability, customer portal, mobile shop-floor — should be scoped as phase two once the core workflow is stable and staff are comfortable with the system. A system that covers the core well is more valuable than a system that covers everything poorly.
Realistic costs for a Singapore SMB manufacturer
A focused first system — quoting, order management, job tracking, inventory, document generation, InvoiceNow — typically costs S$15,000–S$25,000 on a fixed-price basis. Discovery (one week, S$1,500–S$3,000) produces the exact scope and price before any build commitment. Build takes 6–10 weeks. Adoption (parallel run and training) takes 2–4 weeks.
The Enterprise Development Grant (EDG) can offset up to 50% of qualifying project costs. The application must be submitted and approved before project work begins. A Discovery engagement produces the project plan and cost estimate required for the EDG application.
How to start without disrupting operations
The parallel-run model — where the new system and Excel both run simultaneously for 2–4 weeks — is the lowest-risk cutover approach. Staff enter data in both systems during the parallel run, which validates that the new system produces the right outputs before Excel is switched off. No hard cutover on day one.
A paid Discovery engagement (one week) maps your current workflow as it actually operates — not as it is supposed to operate — before any system is designed. This produces a written build plan, system architecture, fixed scope, and fixed price. If you do not like the plan, you walk away with the document. The Discovery cost is S$1,500–S$3,000.