Job costing is where most Singapore manufacturers first feel the pain of Excel. The spreadsheet works well enough when there are 10 to 20 jobs running simultaneously. The formula links are manageable. One person owns the master file. Then the business grows — 40 jobs, 80 jobs, multiple product lines — and the same spreadsheet becomes a liability.
What Excel Does Well
Excel is genuinely good at ad-hoc analysis. When a job director wants to see which product lines drove margin last quarter, a pivot table built in 20 minutes answers the question. When accounting wants to cross-reference job costs against invoices, VLOOKUP gets it done without a developer.
Excel is also forgiving. If the job costing model needs to change — a new overhead allocation method, a new cost category for subcontractor work — a capable person can update the formulas in an afternoon. ERP systems require configuration or development work for the same change.
For manufacturers with fewer than 25 concurrent jobs and a single person responsible for cost tracking, Excel job costing can work reliably. The problems start when the operation grows past that threshold.
Where Excel Breaks
The first failure mode is concurrent access. When three people need to update job costs simultaneously — the production coordinator logging material usage, the project manager updating labour hours, and the estimator checking actuals for a quote — Excel forces serialisation. One person edits, the others wait. In a fast-moving production environment, this creates a 24 to 48 hour lag between activity and the cost record.
The second failure mode is formula fragility. A job costing model with 15 tabs, SUMIF formulas referencing named ranges, and conditional formatting across 300 rows is one accidental paste-as-values away from silent errors. "Silent" is the key word — the spreadsheet continues to produce numbers, but they are wrong. Most manufacturers discover this when a job they believed was profitable turns out to have lost money, and the postmortem reveals a formula break that occurred three months earlier.
The third failure mode is data staleness. Material consumption is typically recorded in Excel at the end of the day or end of the week — not in real time as goods are issued. Labour hours are often entered from memory on Friday afternoon. By the time the job cost report is produced, it reflects the operation as it was two to five days ago. In a tight-margin business, this lag means pricing decisions are made on information that is already outdated.
What ERP Actually Does for Job Costing
A purpose-built ERP does not just replace the spreadsheet — it changes where data entry happens. Instead of a coordinator copying numbers from job sheets into Excel, materials are issued against production orders at the point of use. Labour is clocked against jobs at the machine or workstation. The cost accumulates in the system in real time, and the job cost report pulls live data rather than manually entered data.
This changes what questions you can answer. With Excel job costing, "which jobs are over budget today" requires someone to build the analysis. With ERP job costing, it is a filter on the job list. "What is the actual material variance on job 4821 versus the estimate" is a click, not a lookup. "Which customer accounts have the highest average job profitability this quarter" is a report, not a half-day project.
ERP also closes the loop between quoting and costing. When a new quote is built using actual cost data from completed similar jobs — not averages from a rate card — win rates improve because the prices are right, and margin surprises reduce because the estimate reflects real operating costs.
The Real Cost Comparison
Excel appears to have no cost because the software is already licensed. But the real cost is the labour time to maintain it. A production coordinator spending 90 minutes per day on job cost tracking — a modest estimate for a shop with 60 active jobs — represents approximately 375 hours per year. At S$25 to S$40 per hour fully loaded, that is S$9,000 to S$15,000 per year in labour cost, before accounting for errors.
A custom ERP built by Start Canyon for job costing in a 50 to 150 person Singapore manufacturer typically costs S$15,000 to S$30,000 to build, with no recurring licence fee. The system pays back within 18 to 24 months in labour time savings alone — before counting the margin improvement from better pricing decisions.
The comparison looks different for a manufacturer with 15 active jobs. At that scale, Excel may be the right tool for another two to three years. A Discovery engagement (one week, S$1,500 to S$3,000) produces a breakeven analysis specific to your job volume, margin spread, and current cost of maintaining the Excel system — so the decision is based on your numbers, not a generic benchmark.