Industry — Manufacturing
ERP for Manufacturing
Manufacturing is where generic ERP most reliably fails. Not because the software is bad, but because the shop floor runs on information that finance-first systems were never designed to capture.
Is your data ready for ERP?
30 minutes, no pitch. We reply within one working day.
First, the question that decides everything else
Before budget, headcount or platform preference, one distinction narrows the field more than any other: are you a discrete manufacturer or a process manufacturer? Most mid-market ERP handles discrete well and process poorly. Buying the wrong side of this line is the single most expensive mistake available in manufacturing ERP, because you spend the entire implementation budget customising your way back to where you should have started.
What manufacturing needs that generic ERP does not have
Every ERP sells finance, sales and inventory. Manufacturing needs five things beyond that, and their absence is what turns a promising demo into a stalled rollout.
- Multi-level bills of materials, with versions
- A sub-assembly that is itself made from sub-assemblies. When engineering changes a component, the system must know which parent assemblies and open work orders are affected, and which version was used on last month's build.
- MRP that schedules against real capacity
- Many systems plan as though capacity were infinite, then hand you a schedule the factory cannot physically run. Finite-capacity scheduling is a genuine dividing line between platforms, and it is rarely obvious in a sales demo.
- Work-in-progress visibility
- Material that has left the store but is not yet finished goods still has value and still has to be accounted for. WIP is where the shop floor and the ledger disagree most.
- Routings and work centres
- Which operations happen in what order, on which machine, taking how long. Without accurate time standards, MRP produces confident schedules built on fiction.
- Lot and batch traceability
- For food, pharma and chemicals this is a regulatory obligation, not a feature. You must be able to trace any finished unit back to its input lots, and forward from any input lot to every unit it touched, quickly enough to matter during a recall.
Platform fit, stated honestly
This describes what each product can actually do for manufacturing. It is not a ranking of which we would rather sell — we hold no reseller agreements, so we have no commercial reason to prefer any row.
| Platform | Discrete | Process | The honest note |
|---|---|---|---|
| SAP Business One | ●●● | ●●○ | Strongest statutory and costing discipline. Rigid by design. |
| Odoo | ●●● | ●○○ | Full MRP module. Process/batch depth is the weak point. |
| NetSuite | ●●○ | ●●○ | Advanced Manufacturing is a separate paid module — budget for it. |
| Dynamics 365 BC | ●●○ | ●○○ | Sensible if you are already standardised on Microsoft. |
| ERPNext | ●●○ | ●○○ | Real BOM and work orders. Smaller ecosystem. |
| Zoho | ○○○ | ○○○ | Not a manufacturing system. Inventory only, no MRP engine. |
| Salesforce / HubSpot | ○○○ | ○○○ | CRM. No manufacturing capability whatsoever. |
The two bottom rows matter most. Zoho and Salesforce are both excellent at what they do, and both are routinely mis-sold into manufacturing businesses that then discover there is no MRP engine underneath. If a consultant has proposed either as your manufacturing system, ask them specifically how it plans production against capacity.
Where manufacturing implementations actually break
Configuration is rarely the problem. Four things account for most of the overrun we scope against, and all four are data or adoption problems rather than software problems.
1. Your bills of materials are wrong
Not through negligence — through drift. Substitutions made on the floor and never recorded, scrap allowances that exist only in a supervisor's head, engineering changes captured in a spreadsheet nobody migrated. Every factory believes its BOMs are broadly accurate, and the first MRP run is where that belief meets evidence. We scope BOM verification as its own phase for exactly this reason.
2. Nobody can reconcile opening WIP
To go live you need an accurate value of everything part-finished on the floor at cutover. Most businesses have never had to produce that number to an auditable standard, and producing it for the first time under a go-live deadline is where schedules slip.
3. Routing times have never been measured
MRP schedules using the time standards you give it. If those standards are estimates copied from a decade-old costing sheet, the system will produce a plan that looks authoritative and cannot be run. The factory then quietly reverts to the whiteboard, and the ERP becomes a data-entry chore rather than a planning tool.
4. The shop floor will not use something slower than paper
An operator with a job to finish will not fight a fourteen-field form. If recording a completed operation takes longer than the paper docket it replaced, adoption fails and every downstream report becomes fiction. Shop-floor screens have to be designed for gloved hands and thirty-second interactions, and that design work belongs in the plan rather than in the fortnight after go-live.
Common questions
Which ERP is best for manufacturing?
There is no single answer, and any firm that gives you one without asking whether you are discrete or process manufacturing is selling, not advising. Discrete manufacturers assembling countable units have far more options than process manufacturers working to formulas and batch yields. That distinction narrows the field before any other requirement is considered.
Can we use Zoho or Salesforce for manufacturing?
Not as your manufacturing system. Zoho Inventory manages stock levels but is not an MRP engine — it does not plan production against capacity or handle multi-level bills of materials. Salesforce is a CRM and has no manufacturing execution capability at all. Either can sit alongside a manufacturing ERP for sales and service; neither can replace one.
How long does a manufacturing ERP implementation take?
Longer than the same ERP in a services business, because the data is harder. Bills of materials, routings and opening work-in-progress balances all have to be correct before go-live, and in most factories they are not correct today. Expect the data preparation to take as long as the configuration, and plan for it explicitly rather than discovering it in week nine.
Do we need MRP or is inventory management enough?
If you assemble or transform materials and need to know what to buy and make, and when, you need MRP. Inventory management tells you what you have. MRP tells you what to order and produce given demand, lead times and existing stock. Businesses that buy inventory software expecting MRP behaviour are the most common mis-purchase we see in this sector.
Platforms worth shortlisting for manufacturing
Ordered by how well the product handles manufacturing, not by what we would rather sell.
Not sure which side of the discrete/process line you sit on, or whether your BOM data is ready? That is the first thing we work out on a scoping call, and it costs nothing.