In this guide 6 sections
Buy a generic ERP only for finance and purchasing. Adapt a ready product when its workflows match yours, and build custom only for the parts that win customers. Test each option against multi customer BOMs, consigned stock, revision control and line traceability before signing anything.
Start with what makes an EMS shop different
A brand owner builds one product. You build forty for forty customers, each with its own bill of materials, approved vendor list and engineering change process. That single fact breaks many generic systems.
Write the differences down before talking to any vendor. Customer-supplied parts, mixed turnkey and consigned jobs, and revision-locked work orders are the usual ones.
Everything else is ordinary. Payables, payroll and general ledger work the same in any factory.
- Separate BOM and revision per customer assembly
- Consigned stock that you hold but do not own
- Turnkey, partial turnkey and kitted jobs
- Customer-specific labels, reports and certificates
- Lot and reel traceability to the placed board
Where a generic ERP fits and where it strains
A mainstream ERP handles purchasing, invoicing, inventory valuation and costing well. Its weak points are usually consigned stock held in your name, alternate part approvals by customer and a work order tied to one frozen revision.
Customisation can fill gaps, but each patch makes upgrades slower. Count the patches in the first quote, because they decide your cost of ownership.
Have the vendor import one real customer BOM with alternates, then run a work order from release to close. A live demo of that beats any slide.
What a configured ready product gives you
A ready product built for contract manufacturers already knows about SMT lines, reel and feeder records, ECO handling and customer portals. You configure fields and approvals rather than design them.
The risk is fit at the edges. Check how it treats a customer who sends a revised Gerber pack mid order, or a shortage that needs a broker buy.
Time to start is its main strength. A working system in months beats a perfect one in years.
When writing your own software pays off
Build where your process is the product. A quoting engine that reads a customer BOM, prices it against distributor stock and flags long lead parts can win jobs that rivals quote slowly.
Keep the custom part small and let it talk to the ERP through an API. Replace the whole ERP only if the first two options truly fail on your core flow.
Own the data model either way. If you cannot export your BOM history, you are renting your own knowledge.
- Quote tool that reads customer BOM files
- Shop floor screens for line changeover checks
- Customer portal showing order status and inspection reports
- Rework and scrap capture by defect code
- Dashboards of yield by customer and assembly
Score the options before you sign
Draw a scorecard on a single sheet. Your five hardest scenarios become the rows, and the three options become the columns. Score each by running the scenario, not by asking a salesperson.
Include the people cost too. Who maintains the patches, who trains a new planner and who answers when the line stops at night?
Run the scorecard this week with your operations, quality and purchasing leads in one room, and cut any option that fails two rows.
EMS software scorecard
A single-sheet grid of five scenario rows against three option columns. Rows include loading a customer BOM with alternates, receiving consigned reels, issuing an ECO mid order, tracing a reel to a board and closing a turnkey job. Each cell holds a pass, a workaround or a fail, plus an owner.
Open a blank worksheet to printQuestions owners ask
Is a generic ERP enough for a small EMS provider?
It can be, when you run few customers and little consigned stock. The strain appears when BOMs multiply and each customer wants its own revision rules and reports.
How long does an adapted ready product take to start?
It depends on data volume and the number of lines. Plan for BOM import, stock opening balances and a trial on one customer before moving everyone.
Should we build our own MES?
Rarely at first. Build the small pieces that give an edge and connect them to a bought core, then judge again once real use shows the gaps.
Who should be in the decision?
Operations, purchasing, quality and finance, plus the planner who lives in the system daily. Their scenarios become the scorecard rows.